See discussions, stats, and author profiles for this publication at: [Link]

net/publication/381847426

Scientific Research-An Introduction

Book · June 2024

DOI: 10.9734/bpi/mono/978-81-976007-6-0

CITATION READS

1 1,064

1 author:

J.G.C. da Silva
Universidade Federal de Pelotas (retired).
89 PUBLICATIONS 170 CITATIONS

SEE PROFILE

All content following this page was uploaded by J.G.C. da Silva on 30 June 2024.

The user has requested enhancement of the downloaded file.

Scientific Research: An
Introduction
Scientific Research: An
Introduction

India . United Kingdom

Author(s)

João Gilberto Corrêa da Silva

Federal University of Pelotas, Pelotas, RS, Brazil.

E-mail: jgcs1804@[Link];

FIRST EDITION 2024

ISBN 978-81-976007-6-0 (Print)

ISBN 978-81-976007-1-5 (eBook)
DOI: [Link]

__________________________________________________________
© Copyright (2024): Author(s). The licensee is the publisher (B P International).
Peer-Review Policy: Advanced Open Peer Review policy has been followed for review. This book was
thoroughly checked to prevent plagiarism. As per editorial policy, a minimum of two peer-reviewers
reviewed the manuscript. After review and revision of the manuscript, the Book Editor approved the
manuscript for final publication.

Approved by
(1) Dr. Roxana Plesa, University of Petrosani, Romania.
Reviewers
(1) Mohammed Aboud Kadhim, Middle Technical University, Iraq.
(2) George Manousakis, University of Patras, Greece.
Contents
Preface i

Abstract 1

Chapter 1: Science and Scientific Method 2-28

1.1 INTRODUCTION 2

1.2 SOURCES OF KNOWLEDGE 3

1.3 THE RELATION OF SCIENCE WITH PHILOSOPHY 7

1.4 BRIEF HISTORY OF SCIENCE 8

1.5 SCIENTIFIC METHOD 14

1.6 SCIENCE: STRATEGY AND TACTICS 17

1.7 OBJECTIVES AND SCOPE OF SCIENCE 21

1.8 ROLE OF LOGIC IN SCIENCE 25

1.9 BASES OF SCIENCE 26

Chapter 2: Scientific Research 29-55

2.1 INTRODUCTION 29

2.2 BASIC CONCEPTS 30

2.3 OBSERVATION AND REASONING 35

2.4 SCIENTIFIC RESEARCH PROCESS 44

2.5 OBJECTIVES OF SCIENTIFIC RESEARCH 45

2.6 STAGES OF ANALYTICAL SCIENTIFIC RESEARCH 47

2.7 RESEARCH WORK AND THE ROLE OF THE RESEARCHER IN 51

SCIENCE

2.8 STATISTICS IN SCIENTIFIC RESEARCH 52

Chapter 3: Scientific Research Methods 56-89

3.1 INTRODUCTION 56

3.2 EXPLORATORY RESEARCH METHODS 57

3.3 DESCRIPTIVE RESEARCH METHODS 60

3.4 EXPLANATORY RESEARCH METHODS 70

3.5 USE OF SCIENTIFIC RESEARCH METHODS 86

Conclusion 90

Disclaimer (Artificial Intelligence) 91

Competing Interests 92

References 93-94

Index 95-96

Biography of author(s) 97
 Scientific Research: An Introduction
Preface

PREFACE

This text is intended for use by students and researchers to fill the lack of
accessible literature for the formation of scientific research knowledge
indispensable for higher education professionals.

The literature includes many texts from particular research areas. However,
almost all of them do not address or only make references to the conceptual and
methodological bases that underlie scientific research. This failure is a source of
inefficiency in much research.

The initiative to publish this book arose from the need to fill this gap, providing
the student and researcher with material that is quite sparse and difficult to
access in the literature. Furthermore, many concepts are presented in the
literature vaguely and imprecisely, which does not allow the establishment of a
coherent and complete conceptual and methodological basis.

This book is a compilation of three articles published by the author.

Chapters 1 and 2 introduce the scientific method and scientific research. Chapter
3 provides a description of scientific research methods and their strategic uses in
the sequential phases of each cycle of the application of the scientific method for
generating scientific knowledge: synthesis, analysis, and synthesis. These three
chapters are considered a necessary introduction to the proper understanding of
scientific research and the distinct implications of its methods for the scientific
knowledge generated. The relevant literature considered in the development of
this book is listed in the Bibliography.

The author thanks the students and researchers whose contributions inspired the
creation of this text and were fundamental to better adapting it to its purposes.

Pelotas, March 2024

João Gilberto Corrêa da Silva

___________________________________________________________________________________
© Copyright (2024): Author(s). The licensee is the publisher (B P International).

i
 Print ISBN: 978-81-976007-6-0, eBook ISBN: 978-81-976007-1-5

Scientific Research: An Introduction

João Gilberto Corrêa da Silva a++*
DOI: [Link]
Peer-Review History:
This book was reviewed by following the Advanced Open Peer Review policy. This book was thoroughly checked to prevent
plagiarism. As per editorial policy, a minimum of two peer-reviewers reviewed the manuscript. After review and revision of
the manuscript, the Book Editor approved the manuscript for final publication. Peer review comments, comments of the
editor(s), etc. are available here: [Link]

ABSTRACT
This book aims to review the foundations of the scientific approach to knowledge
generation, the conceptual and methodological basis of scientific research, and
the most common research methods, particularly in biology and agriculture.
Science is characterized and distinguished as an approach to knowledge of nature
by its objects, methods and objectives. Objects are natural phenomena; methods,
disciplined cooperative research and a set of objective, systematic, rational and
critical procedures that have been generically called the "scientific method"; and
objective, the construction of a factual, verifiable and explanatory body of
knowledge. The scientific method is the strategy of science to generate knowledge,
comprising three sequential phases such as synthesis, analysis and synthesis. Its
implementation is carried out by scientific research. Scientific research is
employed in the three phases of the scientific method: in the initial phase of
synthesis, to identify scientific problems, in the analysis phase, to try to solve these
problems, and in the final synthesis phase, to incorporate the new knowledge
generated into the existing body of knowledge. Thus, scientific research comprises
systemic research, which focuses on global systems, in the initial and final phases,
and analytical research, which focuses on subsystems, in the analysis phase. The
function of the research method is to organize research procedures to minimize
error, save effort, and supply an efficient collection of evidence relevant to the
research objectives. The literature presents a wide range of research methods,
with different procedures and names that vary according to the area of science.
Due to their characteristics, each of these methods is most suitable for one of the
three objectives of scientific research: exploratory, descriptive and explanatory, the
latter being the objective of the synthesis phases and the first two of the analysis
phase.

Keywords: Science; scientific method; scientific knowledge.

___________________________________________________________________________________
© Copyright (2024): Author(s). The licensee is the publisher (B P International).

________________________________________________________________________
a
Federal University of Pelotas, Pelotas, RS, Brazil.
++
Retired Professor;
*Corresponding author: E-mail: jgcs1804@[Link];
 Chapter 1
Print ISBN: 978-81-976007-6-0, eBook ISBN: 978-81-976007-1-5

Science and Scientific Method

DOI: [Link]

ABSTRACT

The roots of science go back to the contributions of Greek philosophers some

2,500 years ago. However, science as it is known today and its great power and
remarkable influence over humanity emerged in the 16th century as a
consequence of the Renaissance revolution that radically transformed the
objects, methods and objectives of knowledge of nature. Objects became natural
phenomena; the methods, disciplined cooperative research, and a set of
objective, systematic, rational, and critical procedures that have been generically
called the "scientific method"; and the objectives, the construction of a factual,
verifiable and explanatory body of knowledge. What essentially characterizes
science as a rationally and critically grounded body of knowledge is the method
by which that knowledge is constructed. This chapter reviews the foundations of
the scientific approach to knowledge generation and characterizes the scientific
method.

Keywords: Natural phenomena; sources of knowledge; knowledge generation;

investigation of nature; science strategy.

1.1 INTRODUCTION

The common man can acquire knowledge in several ways. The farmer, for
example, is aware of the plants he cultivates, the time of sowing and planting, the
way to treat the land, and the means of protection against insects and pests. This
knowledge comes from imitation, information transmitted by predecessors, family
members and formal education, and personal experience. This person may also
possess knowledge generated by research carried out by scientific institutions
that is transmitted to him by means of communication and training. Scientific
knowledge can also be acquired more rationally through formal higher education
and enhanced with specialization in postgraduate courses (Silva, 2022a).

The discovery that nature is governed by an intelligible scheme originated in

Greece. Greek theory discovered the universe of ideas and forms, ordered by the
rules of geometry, and the universe of nature, consisting of logically ordered
movements. All Greek theory is dedicated to the description of these two orders,
embodied in Euclid's Geometry, Aristotle's Physics and Plato's Theory of Ideas.
Only at the beginning of the 17th century did modern science originate (Silva,
2022a).

2
 Scientific Research: An Introduction
Science and Scientific Method

Science is a logical process of investigation for solving problems and seeking

answers to questions regarding natural phenomena. Through the scientific
method, scientists attempt to generate a body of knowledge free from personal
beliefs, perceptions, values, attitudes and emotions. This is achieved by the
empirical verification of ideas and beliefs through a procedure open to public
inspection. The reliability of scientific knowledge derives from its evidence base
provided by objective observation (Silva, 2022a).

Scientific knowledge is not absolute and definitive knowledge. On the contrary, it

tends to improve itself and, as a result, lead, for example, to the creation of new
more adequate and convenient methods, techniques and procedures. This
progress is achieved by the permanent activity of inquiry to which researchers
are dedicated. Thus, science is a source of knowledge generation that renews
itself to solve problems, answer questions, and develop more effective
procedures to solve problems and answer questions (Silva, 2022a).

This chapter aims to review the origins, evolution and characteristics of the
scientific approach to knowledge generation. It is based on Silva (2020a) and the
contributions listed in the References.

1.2 SOURCES OF KNOWLEDGE

The search for knowledge of nature was already a characteristic of prehistoric

man. Man's first contact with nature generated eminently sensitive knowledge.
Limited resources allowed the perception of phenomena only through the senses
and superficial explanations. Seeking to unravel the universe, man adheres to
the cult of the forces of nature as a form of knowledge. In the passage from
primitive times to antiquity, man expands the limits of his knowledge, going from
mythological explanations of the universe to explanations of a religious nature.
The next step is interpretation by way of reason. With the rise of philosophy, the
explanation of nature becomes eminently rational. The incessant search for truth
leads to the passage from philosophy to science, that is, to the interaction of
reasoning with the empirical approach to explaining the causal relationships
between phenomena, through rigorous analytical and rational procedures (Silva,
2022a).

Empirical knowledge. The most remote and ordinary way for man to create his
representations and interpretations of reality is through everyday experience and
common sense. The knowledge constituted by these representations arises from
the need to solve immediate problems. This empirical knowledge is a
spontaneous and unsystematic way of representing reality, without an
appropriate method to deepen its foundations. This is the knowledge of the
common person, without training, generated by his relationship with the material
world. It originates from individual and collective experiences and beliefs. This
knowledge is developed, used to predict future events and transmitted from one
generation to the next (Silva, 2022a).

3
 Scientific Research: An Introduction
Science and Scientific Method

Common sense implies a certain degree of abstraction. However, it proceeds to

a simple junction of ideas, notions and concepts and does not reach a level of
elaboration that originates the creation or use of concepts whose meanings
deepen the understanding of reality. This limitation generates knowledge of facts
based on their appearance, without concern for their explanation through
analysis to characterize their origins. The knowledge generated is constituted by
an agglomeration of elements, without unity and coherence. Besides, it is not
subjected to a level of criticism necessary to understand reality beyond what is
experienced. This can lead to fatalistic interpretations of the situations that
present themselves, which attribute to a destiny defined by a superior being
responsible for the events of life (Silva, 2022a).

Mythical knowledge. Without using writing, primitive man sought to explain,

narrate and announce phenomena through symbols and allegories, thus creating
myths. Reality then began to be interpreted based on these myths. Mythical
knowledge is a product of the oral transmission of empirical knowledge from
generation to generation. Myths were rooted in the culture and tradition of
ancient peoples and represented much more than an attempt to explain reality.
Myth constituted the historical foundation of civilizations, explaining the past and
the origin of the present. It was a representation of the real world, recreated from
the subjective elaboration of man's experiences. The origins of inexplicable
phenomena were attributed to immanent powers and forces existing within
objects, animals and people. Ancient peoples worshiped the Sun and the Moon;
the Hindus, the cow; the Mayans, the Incas, the Aztecs and other peoples, the
totems and amulets, monuments and objects that they built themselves. These
peoples believed that these entities and objects had the strength and powers to
do good and evil, and created their interpreters or interlocutors, such as witches,
sorcerers, priests, and shamans (Silva, 2022a).

At a later stage, these mysterious forces are transferred to fictitious beings with
human forms. These beings were invisible, represented by statues to which,
mainly the Greeks, Romans and Egyptians, rendered cults. The interventions of
these superior beings, called gods, would be the origin of all natural phenomena.
It is the phase of polytheism. The cause of a set of phenomena in a particular
sector of nature was attributed to a particular god. Polytheism predominated in
Greece and ancient Rome. Thus, for example, Diana was the goddess of the
hunt, Eolo the god of the wind, Neptune the god of the sea, and Ceres the
goddess of the harvest (Silva, 2022a).

Theological knowledge. Religious experience is as old as civilized man is.

Religious or theological knowledge is aimed at understanding the totality of
reality. Its purpose is the explanation of a unique origin and end concerning the
genesis and existence of the universe. It attributes the cause of all phenomena to
a single superior being, God. In the Judeo-Christian religion, God is the only
creator of all that exists. The creation of the universe and natural phenomena are
attributed to him, in particular the creation of man and animals, their existence,
transformations and purposes.

4
 Scientific Research: An Introduction
Science and Scientific Method

From a theological point of view, the divine existence is evident, and evidence
needs no proof. Based on this principle, it seeks to find explanations for
everything that has happened and happens to human beings and seeks to study
issues relating to knowledge of the deities, their attributes and relationships with
the world and with men. Religion has its foundations in dogmas and rites, which
are accepted by faith and cannot be proved or criticized because it is the only
source of truth. The sources of theological knowledge are the holy books – Bible
for Christians, Talmud for Jews, Koran for Muslims and Veda for Hindus. The
interlocutors between man and God are priests, rabbis, pastors and other
interpreters (Silva, 2022a).

Philosophical knowledge. Philosophical knowledge began with man's first

attempts to understand the world by associating reasoning with observation.
Philosophy developed in Persia, China, India, and elsewhere in the Orient.
Western philosophy originated six centuries before Christ, from the teachings
mainly of Greek philosophers such as Pythagoras (580-500 BC), Socrates (470-
399 BC), Plato (428-348 BC) and Aristotle (384-322 BC), the first ones who
sought to interpret nature by observation and logic, without necessarily
supernatural interpretation.

Philosophy seeks knowledge of first causes or principles. It has no particular

object but assumes the guiding role of science itself in the solution of universal
problems. Thus, philosophy is the expression of the universality of human
knowledge, that is, the source of all areas of human knowledge. In this context,
science not only depends on philosophy but also includes itself (Silva, 2022a).

Philosophical knowledge developed from ideas and theories formulated by great

philosophers, such as the aforementioned Greek philosophers, and Saint
Thomas Aquinas (1224-1274), Francis Bacon (1561-1626), René Descartes
(1596-1650), John Locke (1632-1704), Jean Jacques Rousseau (1712-1778),
Immanuel Kant (1724-1804), Georg Hegel (1770-1831), Karl Marx (1818-1883)
and, in recent times, Bertrand Russell ( 1872-1970), Ernest Nagel (1901-1985),
Karl Popper (1902-1994) and Carl Gustav Hempel (1905-1997).

Philosophy rests on reflection on experience. Reflection provides variations in the

interpretation of impressions, images and opinions. However, philosophy is not
reduced to a search for reflexive and conceptual originality. Above all, philosophy
aims to understand reality and provide reflective and logical content to change
and transform reality. Philosophy has the task of elaborating presuppositions and
guiding principles of human actions. Philosophy is also a critical reflection of
society, politics, law, and education. For this reason, philosophical knowledge
evolves according to the historical context (Silva, 2022a).

Scientific knowledge. Science originated from philosophy. From 5,000 years

before Christ, Babylonians and Egyptians developed important knowledge mainly
in mathematics and astronomy. However, the Greek philosophers were the main
ones who were responsible for the combination of knowledge in these two areas,
which constituted the starting point of science.

5
 Scientific Research: An Introduction
Science and Scientific Method

The task of science is the explanation of natural processes and phenomena. No

system of theoretical ideas, technical terms, and mathematical procedures can
be regarded as scientific unless it relates to these empirical facts at some point
and in some way and helps to make them more intelligible. Scientific knowledge
is a system of methodical knowledge about nature. It differs from other forms of
knowledge in that it requires objective empirical verification of all explanations
regarding phenomena, which allows an understanding of their nature and
causes, free from the observer's influences, desires and prejudices. Scientists
seek knowledge of the relationships between phenomena, that is, of natural laws.
It relies on logical reasoning to deduce new knowledge from general laws or
concepts (Silva, 2022a).

The special character of science can be explained by the circumstances in which

scientists work in their respective fields. These circumstances include the basic
principle of knowing the natural world through demonstrative arguments. Thus, a
representation or interpretation of a phenomenon or process is only scientific
knowledge if it has been verified or empirically demonstrated.

The search for scientific knowledge stems from the perception that the available
body of knowledge is insufficient to understand some phenomenon or natural
event. Part of the available knowledge is common or ordinary knowledge, that is,
non-scientific, and part is scientific knowledge, that is, knowledge derived
through the method of science. This knowledge can be put back to the test,
perfected or surpassed, through this same method. As the scientific process
progresses, portions of the ordinary body of knowledge are corrected or rejected,
and the body of scientific knowledge is increased. Thus, science grows from
common knowledge and surpasses it with its growth. In fact, scientific work
begins at the point where experience and common knowledge fail to provide
solutions to problems related to phenomena of interest, or even to formulate
them. However, scientific knowledge is not a mere refinement and extension of
common knowledge. It is knowledge of a special nature. Science also deals with
unobservable phenomena not considered by the common person, raises
conjectures that go beyond common knowledge, and puts these conjectures to
the test on the basis of its theories and with the help of special techniques. On
the other hand, science is unable to develop knowledge from unnatural
explanations (Silva, 2022a).

Thus, scientific knowledge is radically distinguished from common knowledge in

many respects, particularly in terms of method. However, these two sources of
knowledge have some similarities, at least if consideration is limited to the
ordinary knowledge generated by common sense. In fact, common sense, like
science, aspires to be rational and objective. However, the ideal of rationality,
that is, the coherent systematization of knowledge is only achieved through
theories, which constitute the core of science, while common knowledge is
constituted by the accumulation of unrelated pieces of information. Besides, ideal
objectivity, that is, the construction of impersonal representations of reality,
cannot be achieved without overcoming the narrow limits of everyday life and
personal experience. It demands the formulation of conjectures about the

6
 Scientific Research: An Introduction
Science and Scientific Method

existence of physical objects in addition to precarious sensory impressions, and

the verification of such conjectures through objective experience, planned and
interpreted with the aid of theories. Common sense can achieve only limited
objectivity because it is closely linked to perception and action (Silva, 2022a).

In short, scientific knowledge is rational and objective like that which comes from
common sense, but much more rational and objective than that. The peculiarity
of the scientific approach that distinguishes it from common sense is the way it
operates, that is, the scientific method, and the purpose for which this method is
applied.

1.3 THE RELATION OF SCIENCE WITH PHILOSOPHY

Sometimes the scientific method is contrasted with other approaches to

knowledge generation and the philosophical character and logical structure of
scientific knowledge are questioned. First, it should be noted that science is a
method and not a philosophy. As such, it is not committed to any particular
theory or philosophy of knowledge. Indeed, the scientist's action reveals certain
mental preferences and consistencies of his method that are sometimes related
to the views of philosophical schools known by the designations of rationalism,
empiricism, positivism, pragmatism and determinism.

A brief examination of the philosophical meaning of these designations is useful

to clarify the distinction of the scientist's position in relation to these views, which
imply differences in perspectives concerning knowledge (Silva, 2022a).

Rationalism refers to the philosophical conviction that human reason is the main
instrument and ultimate authority in the search for truth. Rationalism does not
deny the value of sensory experience as a source of knowledge but holds that
only the logical operation of the mind can determine the truth of any experience
or idea. For his adherence to the established rules of logic, the scientist could be
labeled a rationalist. However, this designation would be inappropriate, since he
does not only believe in pure reason as a guide to valid knowledge. The rationale
of the scientific method is grounded in the system of logic employed in scientific
reasoning, but the method of scientific analysis requires much more than just
faith in reason (Silva, 2022a).

Empiricism is based on the conviction that sensory experience must be regarded

as the most reliable source of knowledge. Certainly, science is, in part and in
certain areas, an empirical method as well as a logical one, that is, a rational
method. However, the empirical aspect of science is related to the way data are
perceived and not to faith in the exclusive validity of sensory experiences.
Reasoning about empirical impressions is as important as the sensations
received. Empirical evidence is basic to science, but it only has meaning if
interpreted by particular notions about its attributes, effects, etc. In fact, a large
part of the structure of scientific knowledge is composed of abstractions, that is,
ideas about phenomena and their interrelationships, not empirical evidence.
Science is empirical in the sense that its last resort for establishing the credibility

7
 Scientific Research: An Introduction
Science and Scientific Method

of any particular knowledge is an empirical fact, empirical demonstration, or

empirical prediction. However, to hold that science is only or basically empirical is
to inadequately limit its theoretical framework (Silva, 2022a).

These two philosophical schools have more than three centuries of history. A
more modern school that has been related to science is logical positivism, which
refers to the belief that statements have factual meaning only when they are
confirmed by empirical evidence. In fact, logical positivism is a move by some
philosophers of science towards a unification of the various branches of science
by clarifying ideas and developing methodological precision through logical
analysis. It is a derivation of empiricism that emphasizes the development of
objective communication, especially through symbolic logic and mathematics.
Some scientists have something of a logical positivist in that they constantly seek
a common unity of method, basic principles, and communication. However, even
among today's few “neopositivists” the original restrictive attitude regarding the
credibility of certain types of knowledge has been greatly modified (Silva, 2022a).

The fourth philosophical school that is sometimes related to science is

pragmatism, which is the belief that the ultimate test of an idea's worth is its
usefulness in solving practical problems. Certainly, the scientist is a practical
man who seeks the solution to real problems. However, as a philosophical
position, pragmatism is of little value in modern science. In fact, much scientific
knowledge is purely theoretical, with no practical value in itself. This theoretical
knowledge plays a vital role in the structure of science. The set of abstract
theories that underlie all science, which constitute what is often called "pure
science", is highly unpragmatic (Silva, 2022a).

The last philosophical link to science is determinism, that is, the argument that
nothing has a place in nature without natural causes. Determinism in science is
not a “creed” but a “postulate” that is employed in the analysis of causality.
Science has also been related to materialism, that is, to the philosophical
doctrine that all knowledge can be derived from the study of matter. However,
today science is materialistic, mechanistic or deterministic only to the extent that
it is based on a base of objectively demonstrable facts, with the help of physical
instruments of observation and measurement. Furthermore, determinism should
not be confounded with “fatalism”, that is, the natural inevitability of events.
Science seeks to understand the regularities of phenomena, but such regularities
are not imputed to any unavoidable causal agent. A “postulate of regularity in
nature” is presupposed by the scientist as a principle, with the purpose of
obtaining reliable knowledge. However, this principle is not assumed as a “law of
nature”. This term has no important meaning in modern scientific explanations of
causality (Silva, 2022a).

1.4 BRIEF HISTORY OF SCIENCE

Science in antiquity. Scientific knowledge (or science) has its roots in the
contributions of Greek philosophers. The first attempts to explain natural
phenomena without the mythological foundation of personal agents, but based

8
 Scientific Research: An Introduction
Science and Scientific Method

on reason and evocation of causes, originated with Thales of Miletus (624-546

BC), six centuries before Christ. However, speculative explanations of
phenomena were based on common sense rather than technical arguments
about artificially controlled experiences. This approach emerges with Aristotle
(384-322 BC), in the fourth century BC. Through rigorous observation and
disciplined theorizing, Aristotle studied animal anatomy in detail and created a
biological science. In each study, he defined the area and its problems,
dialogued critically with his predecessors, and then proceeded to develop his
argument through experience and reason. His explanations were grounded in
terms of perceptible qualities and a series of causes (Silva, 2022a).

Post-Renaissance Science. Although the Greeks created a system of thought

similar to the scientific approach some 2,500 years ago, there was little progress
in the following centuries. The great power of science and its marked influence
have recent origins. Science, as it is known today, has roots in the 16th century,
mainly from three sources of influence: a) the discovery of classical antiquity and
the publication of Latin and Greek texts from all fields, including science, in the
Renaissance; b) the invention of the printing press by Johannes Gutenberg
(1390-1468) and its rapid expansion, which made books available and cheap,
before the end of the 16th century, with a marked influence on learning and
culture; c) the discoveries of new lands, by the Spaniards and the Portuguese,
which created new demands for knowledge of astronomy, hydrographic and
mathematical techniques, and introduced new plants, animals, diseases and
civilizations in Europe (Silva, 2022a).

At the end of the 15th century, the first experimenters appeared in Europe.
Speculation is replaced by experimentation. The forerunner of this change in
attitude was Leonardo da Vinci (1452-1510). In studying phenomena in nature,
he sought to carry out experiments, under various conditions and circumstances,
in order to reach a general rule that applied to all experiments carried out. This is
how he established the rule that the weight supported by a column is proportional
to the height and width of the column (Silva, 2022a).

However, in the early 16th century knowledge was still rudimentary and heavily
dependent on confused assimilation from ancient sources. Before the middle of
that century, relevant works appeared, such as the treatise by the Polish
Nicolaus Copernicus (1473-1543), published in 1543, which revolutionized
cosmology with the new heliocentric idea; the new approach to research in
anatomy by the Belgian Andreas Vesalius (1514-1564); and the advances in
algebra of the Italian Gerolano Cardano (1501-1576). Copernicus defended the
idea that mathematics could be applied to research any problem involving
measurable quantities. Although theoretical developments tended to be
speculative, there was considerable progress in many areas. In particular, before
the end of the century, mathematics was usually taught in Europe (Silva, 2022a).

Science in the 17th century. At the turn of the 17th century, a man of traditional
higher education, called "scholastic", still believed that the Earth was the center
of the universe and that the stars and planets revolved around it, moved by some

9
 Scientific Research: An Introduction
Science and Scientific Method

intelligent and divine being, influencing events on Earth according to their

locations and aspects. He believed in a living world, created and guided by God
for the benefit of man. Then, came important scientific discoveries that are still
accepted today, but which were explained on the basis of the culture that was
still present. Thus, in 1600, Englishman William Gilbert (1540-1603), in the
course of demonstrating that the soul of the world was embodied in a magnet,
explained the compass on the grounds that the earth was a very weak gigantic
magnet. Shortly afterward, in 1609, the Polish Johanes Kepler (1571-1630)
discovered the elliptical orbits of the planets around the sun, but his quest for
harmony in the cosmos never ceased. Later, in 1628, Englishman William
Harvey (1578-1657) established blood circulation but explained it as a
microscopic image of the world's circulations rather than a purely mechanical
system (Silva, 2022a).

In the 17th century, there was a radical overhaul of the objects, methods and
functions of knowledge of nature. The new objects became natural phenomena
in a world devoid of human and spiritual properties, and the methods, disciplined
cooperative research, and the functions of combining knowledge with the power
of industry. The great proponents of this revolution in science were the
Englishman Francis Bacon (1561-1626), the Frenchman René Descartes (1596-
1650) and the Italian Galileo Galilei (1564-1642). Bacon contested the exclusive
use of logic and observation, as opposed to Aristotle's rules of logic. He
advocated a new means by which man could establish control of nature, a plan
for the reorganization of science, and proposed a scientific method in his most
celebrated work "Novum organum". Bacon created the method of “exhaustive
induction”, arguing that, ideally, the scientist should provide an exhaustive
enumeration of all examples of empirical phenomena under investigation as a
preliminary to identifying the natural form of which they are a manifestation. He
argued that empirically observed facts are the starting points for all science and
that theory must be considered insofar as it is derived from facts (Silva, 2022a).

In contrast, Descartes focused on the problem of constructing a deductive

system of consistent and coherent theory, by which the argument could proceed
with the formal certainty familiar to Euclidean geometry. He advocated the idea of
a unitary universal science that would bind all possible human knowledge into
one comprehensive wisdom. His renowned work “Discourse on the method”,
published in 1637, marked the definitive consolidation of the scientific method. In
search of certainty, Descartes challenged Aristotle and scholasticism and sought
to compose a philosophy associated with mathematics, where observation and
interpretation are legitimized by empirical demonstration (Silva, 2022a).

Galileo is considered a founder of the experimental method. Galileo was less

radical in his ideals and more comprehensive in practice. Combining
experimentation with mathematics, he contributed considerable advances in
physics and astronomy. He observed that the chandeliers in the tower of Pisa
swayed in the currents of air and, based on the beat of his pulse, he measured
the time taken for the strokes of the largest and smallest arc described by the
swing of the chandelier. Thus, he discovered the property of constancy of

10
 Scientific Research: An Introduction
Science and Scientific Method

pendulum motion. He carried out meticulous experiments on the flight trajectory

of projectiles and the falling of bodies, built a telescope and with this instrument
studied the Moon, the Milky Way, the rings of Saturn and, with detailed
observations, proved Copernicus' heliocentric idea (Silva, 2022a).

Despite their differences in ideas and contributions, Bacon, Descartes, and

Galileo shared a common commitment to the natural world and its study. They
saw nature as devoid of spiritual and human properties. There could be no
dialogue with it, either through mystical illumination or inspired authority. Rather,
it had to be investigated prudently and impersonally, through sensory experience
and reason. Care and self-discipline were needed in both observation and
theorizing, and cooperative work was important for the continued accumulation
and testing of results.

The new ideas of science came to fruition in the 17th century and led to rapid
progress in knowledge in some fields, but slow development in others. Thus,
modern conceptions began in optics with Johann Kepler (1571-1630), and in
electricity and magnetism with William Gilbert (1544-1603). At the end of that
century, Isaac Newton (1642-1727) formulated the law of universal attraction,
and the law of gravity, and brought a new rigor to the methods of experimental
research (Silva, 2022a).

Science in the 18th century. Newton's contributions dominated science in the

18th century. Scientific developments in that century were mainly one of
consolidation.

At the beginning of this period, scientific activity was carried out on a small scale,
mainly by wealthy men and well-trained professionals, such as physicists and
engineers, in their spare time. The mathematical sciences (mathematics,
astronomy, mechanics and optics) were well developed, but physics was still a
set of isolated experiments with qualitative and mostly speculative theories,
chemistry was almost entirely empirical, and biology paid attention mainly to
collecting activities. Before the end of the period, successful examples of well-
organized scientific work already existed, and the foundations for coherent and
efficient theories in almost every area of science had been laid. Force technology
was the first to be influenced by the application of earlier scientific developments.
The steam and vacuum engine, invented by the Englishman Thomas Newcomen
(1663-1729), originated from 17th-century pneumatics (Silva, 2022a).

At the end of the 18th century, the industrial revolution began, transforming
Europe from an agrarian society into an urban society. At the same time, the
French Revolution introduced the modern policies of freedom and democracy.
Scientific activities underwent similar changes: at that time, the social and
institutional foundations for the maturation of science in the 19th century were
established.

The dominant style of science at the time of the French Revolution was
mathematics. At that time, some French mathematicians stood out, such as

11
 Scientific Research: An Introduction
Science and Scientific Method

Pierre-Simon Laplace (1749-1827), Joseph Louis Lagrange (1736-1813), Gaspar

Monge (1746-1818), Jean Baptiste Fourier (1768-1830), Siméon Denis Poisson
(1781-1840) and Augustin Louis Cauchy (1789-1857). The Swiss
mathematicians Jacques Bernoulli (1654-1705), Daniel Bernoulli (1700-1782)
and Leonard Euler (1707-1783) developed the differential and integral calculus,
invented by the German Gottfried Leibnitz (1646-1716). Even in chemistry, the
nomenclature reform achieved by Antoine Laurent Lavoisier (1743-1794) and his
collaborators was mathematical and abstract in style (Silva, 2022a).

Science in the 19th century. The 19th century emerges as a golden age. Science
successfully expands into new fields of inquiry, including a combination of
mathematics and experiments in physics, the application of theory to experiment
in chemistry, and controlled experimentation in biology. This progress was largely
driven by the reform of universities and the establishment of new universities,
where research was encouraged as well as teaching, and by communication
through scientific societies and specialized journals. National and international
meetings were becoming common at the end of this century. The principle of
socially organized research, rather than research by isolated individuals, became
effective. In all fields of knowledge, there was an increasing rigor of methodology
and a deepening of scholarship. During that century, advances in the main
branches of experimental science were so great that in retrospect their former
state seems rudimentary. Physics achieved the close union of precise
experimentation with the abstract mathematical theory that brought
unprecedented depth of knowledge and power to the application of this
knowledge. Different fields were successively unified by the concept of energy.
Thermodynamics united the sciences of heat and work and led to the
development of a theory of chemical changes. Electricity and magnetism were
brought together, initially by experiment and then theoretically, and it was
observed that a fundamental constant of the theory of electromagnetism,
determined by the German Wilhelm Weber (1804-1891), was equal to the speed
of light, determined astronomically by the Englishman James Clerk Maxwell
(1831-1879). In this way, the general properties of matter were successively
understood and made coherent (Silva, 2022a).

In the early 19th century, the English physicist and chemist John Dalton (1766-
1844) formulated the atomic theory of matter, published in his work "The New
System of Chemical Philosophy", published in 1808.

Chemistry started to be built on the theoretical bases of Lavoisier's nomenclature

and Dalton's atomic theory and spent some decades in the task of classifying
substances into elements and compounds. In Germany, Friedrich Kekulé (1829-
1896) discovered the true structure of organic compounds, with the alternative
bonds of the benzene ring. Lothar Meyer (1830-1895) in Germany and Dmitri
Mendeleyev (1834-1907) in Russia dominated the structure of the periodic table
of elements and could predict the properties of unknown elements. From then on,
chemistry began to reach a closer unity with physics and to acquire a growing
potential for industrial applications (Silva, 2022a).

12
 Scientific Research: An Introduction
Science and Scientific Method

The advances in biology have been equally extraordinary. The experimental

approach was initially developed successfully in physiology, mainly by the school
of Johannes Müller (1801-1858) in Germany and by Claude Bernard (1813-1878)
and Louis Pasteur (1822-1895) in France. In England, Charles Darwin (1809-
1882) and Alfred Russell Wallace (1823-1913) conceived of natural selection by
the principle of survival of the fittest and established the theory of evolution
(1859). In Austria, the monk and botanist Gregor Johann Mendell (1822-1884)
established the basic laws of heredity.

Science in the 20th century. Some 19th-century trends became more salient in
the early 20th century. At that time, science became increasingly professional in
its social organization. Almost all research was carried out by highly trained
specialists, employed exclusively or primarily for this work in specialized
institutions. Scientific communities, organized by discipline, enjoyed a high
degree of autonomy in setting research goals and standards, and in certifying,
employing, and awarding their members. Forced by competition, scientists tend
to become highly specialized researchers. The dominant style of this period was
reductionist: research was mainly concentrated on artificially pure, stable and
controllable processes obtained in the laboratory, and the favored theories were
those involving the simplest physical causes, using mathematical arguments.
Scientific developments at that time were modeled after the standards of
theoretical physics (Silva, 2022a).

The positive spirit of this science was shown by its increasing separation from
philosophical reflection. The theories of relativity (1905 and 1916) by Albert
Einstein (1879-1955) and the uncertainty principle in quantum theory (1927) by
the German physicist Werner Heisenberg (1901-1976) raised vigorous
philosophical discussions.

The scientific achievements of the early 20th century are too immense to be
enumerated. However, a common pattern of advancement can be highlighted. In
each of the most important fields, progress was based on the successful
descriptive work of the 19th century. Scientific activity was initially directed
towards a more refined analysis of the constituents and their mechanisms and
then towards syntheses that gave rise to the names of hybrid disciplines such as
biochemistry and biophysics. In physics, classical theories of the main physical
forces (heat, electricity, and magnetism) have been unified by thermodynamics;
and at the beginning of the century, completely new discoveries appeared (X-
rays and radioactivity, for example) and penetration into the structure of matter
(atomic theory and isotopy). These discoveries required a review of the
fundamental laws of physics and some of its metaphysical presuppositions
(relativity and quantum theory). Chemical methods became necessary for much
of this work in physics. On the other hand, the new physical theories were
powerful enough to provide effective explanations for a wide variety of chemical
phenomena. Based on these discoveries, the chemical industry produced a wide
variety of synthetic substances (fibers and plastics, for example) (Silva, 2022a).

13
 Scientific Research: An Introduction
Science and Scientific Method

In the biological sciences, physical and chemical methods have contributed to

discoveries and explanations of ingenious agents (vitamins, hormones) and the
reconstruction of complex cycles of chemical transformations through which
matter lives. Medical science was able to develop bacteriology, and, through the
discovery of specific and general drugs, it markedly reduced the classic epidemic
diseases and the cruel diseases of childhood.

Until the end of that century, scientific advances continued to grow in all areas,
so dizzying that it became impossible to catalog them. Just for illustration: in
transport, culminated with the trip to the Moon and launching of satellites and
probes for space exploration; in health, with the control of many diseases and
increased longevity; in communications, with the large-scale use of the Internet
(Silva, 2022a).

1.5 SCIENTIFIC METHOD

Science is often defined as an accumulation of systematic knowledge. This

definition includes three basic terms for the characterization of science. Yet it is
inadequate, like other definitions that emphasize the content of science rather
than its fundamental characteristic: its method of operation. This is inconvenient
as the content of science is constantly changing, given that knowledge
considered scientific today may become unscientific tomorrow. Furthermore, the
demarcation between science and non-science is not obvious. Indeed, it is not a
sharp line, but a mobile and debated area (Silva, 2022a).

Science aims at understanding the world in which man lives, and knowledge of
reality. Thus, it is fundamentally a method of approaching the empirical world,
that is, the world susceptible to experience by man. The consensus regarding the
essential attributes and processes of the method of science allows for a
functional conceptualization of science through its method, as follows:

Science is an objective, logical and systematic method of analyzing phenomena

created to allow the accumulation of trustworthy knowledge.

For a better understanding of this definition of science, it is convenient to explain

its key terms (Silva, 2022a):

Method: A controversy often arises over the uniqueness of the scientific method.
It can be argued that this is not a conceptual issue, but a mainly semantic
problem that arises from the various meanings attributed to the word “method”.
Indeed, although the various fields of science differ in content and techniques, an
examination of all highly developed science reveals a common basis of
investigation procedures, which constitutes the general method of science. The
implementation of this method in particular areas of science usually requires
specific techniques and procedures, which constitute the particular methods of
science.

14
 Scientific Research: An Introduction
Science and Scientific Method

Objective: Objectivity in science refers to attitudes devoid of personal whim,

bias, and prejudgment, to methods for discovering publicly demonstrable
qualities of a phenomenon, and to the principle that the last resort of a
speculative argument is the objective phenomenon, that is, an observation or
experience that can be publicly verified by trained observers. Evidence in science
is factual, not conjectural, and the “truth” is obtained by empirical demonstration.
Although science is practiced by individuals, the scientific method inspires a
rigorous and impersonal way of proceeding, dictated by the demands of logical
and objective procedures. Scientists constantly seek this way of proceeding
through training and the use of objectification instruments that allow them to look
at their data with as little bias as possible.

Logical: To say that science is a logical method means that the scientist is
constantly guided by acceptable rules of reasoning standardized by reputable
logicians. Competence in science requires competence in logical analysis. Rules
of definition, forms of deductive and inductive inference, probability theory,
calculus, etc., are fundamental to any reputable science. Science is a systematic
arrangement of facts, theories, instruments and processes, interrelated by
reasoning principles. While one can act in applied areas by apprehending and
applying formulas, acting as a scientist requires a thorough grounding in logical
analysis as well as specific factual proficiency and knowledge.

Systematic: Science is a systematic form of analysis. Science proceeds in an

orderly manner, both in the organization of a problem and in the methods of
operation. This is one of the essential features that distinguishes the scientific
approach. The systematic procedure inherent in the scientific approach takes the
form of a sequence of tightly interconnected and logically arranged steps that
allow for few deviations. Verification in science is a systematic process of logical
inference that requires that premises, facts, and conclusions be neatly arranged.

The systematic character of science also implies internal consistency. In a well-

developed science, the various theories and laws are interrelated and
corroborative. They are mutually supportive or at least not contradict each other.
An immature science is characterized by internal disagreements between
theories, laws, propositions, principles and even methods. It should be noted,
however, that complete and final consistency is never achieved, even in the most
advanced sciences. New discoveries suggest new laws, principles and theories,
which, in turn, require the modification of established notions of reality.

Phenomena: The scientific method is applicable to any phenomenon, that is, to

any event or behavior that has objectively demonstrable attributes or
consequences. If an event is presumed to be inherently subjective (for example,
an idea, a feeling, or a dream), then it is not treatable by scientific analysis unless
its presence can be demonstrated through objective attributes or consequences.
Although the phenomena studied by science are publicly verifiable, it should not
be understood that such research objects are the only interest of science. The
scientific method is built on a foundation of ideal abstractions (for example,
notions, ideas, theories, laws, principles, etc.) designed to relate and explain

15
 Scientific Research: An Introduction
Science and Scientific Method

observable objects and events. Much of the content of science consists of

intellectual notions about things and events. However, the object of all such
thoughts is the particular phenomenon under study.

Created: Science is a system made up of diverse factual knowledge synthesized

in an interrelated and logical set created by human ingenuity. In turn, the
scientific method is a creation to serve a particular purpose: the development and
orderly arrangement of knowledge and ideas concerning reality in the form that
seems most fruitful for the ends to be served. It should be noted that man
arranges his thinking concerning the world in which he lives according to various
preferences, and the scientific method is the arrangement that has so far proved
most fruitful for the explanation of objective phenomena.

Accumulation: Science is an accumulative and integrated system, built in an

orderly way, where every fact, law, theory, principle, etc. supports other facts,
laws, theories, etc. However, science is not a mere accumulation of knowledge.
Scientific knowledge is dynamic. Science always looks for additional knowledge,
in the belief that knowledge is never complete. The “truth” in science is always
relative and temporal, never absolute and final. In contrast to many closed
philosophical and ideological systems, science is characterized as an open
system of ideas. Therefore, it is constantly growing, discarding erroneous or
useless notions and replacing them with more correct and useful ones in the light
of new evidence.

The accumulative attribute of science must not mean that it grows by simple
addition. The history of science shows that complex explanations and
designations are constantly replaced by scientifically simpler and more accurate
explanations and terminology. This principle of parsimony of science determines
that the scientist must constantly strive to obtain explanations that involve as few
terms, attributes, concepts and formulas as possible. The accumulative attribute
and the principle of parsimony are closely intertwined. Indeed, science constantly
strives to predict the behavior of phenomena not yet observed on the basis of the
commonly known qualities they possess as members of a class of phenomena.
Together, accumulation, ordering and parsimony allow a large number of specific
predictions to be made from a few basic and general laws.

The term “reductionism” has been used in discussions of this broad principle of
parsimony. Reductionism refers to the general practice of seeking to encompass
as many sub-theories as possible in broader and more inclusive categories of
"grand theories”. Although much of the scientific knowledge at any given time is
temporarily unrelated or uncoordinated, the scientist constantly strives to relate
isolated facts into meaningful sets or models. The history of science shows that
over time such models become integrated into broader systems of facts and
ideas (“theories”) that allow for a greater range of explanations than would be
possible if segmented facts were used in isolation.

Trustworthy Knowledge: In the current context, it is referred to as trustworthy

knowledge for prediction. In this sense, trustworthy knowledge means correct

16
 Scientific Research: An Introduction
Science and Scientific Method

knowledge. Science constantly strives for precision and accuracy. In fact,

science progresses as its measurements and calculations become more refined.
Note, however, that precision and accuracy is not an end in itself. They are only
related to the purposes they serve, that is the promotion of more specific
descriptions that allow for reliable prediction or control.

The function of the scientific method is to understand phenomena in such a way

that the reason and range of accurate predictions can be constantly increased.
Presumably, it is only through a valid and organized system of knowledge, such
as science, that predictions can be effectively extended beyond the limited
experience of a particular and simple group of individuals.

Through the scientific method, scientists attempt to generate a body of reliable

knowledge. This is achieved by empirical verification of ideas and beliefs through
objective observation.

1.6 SCIENCE: STRATEGY AND TACTICS

The scientific method is the general procedure of science applied in the

knowledge acquisition process, regardless of the topic under study. However,
each class of knowledge problems requires the development and application of
special techniques and procedures suitable for the various stages of problem
treatment, from the enunciation of these to the control of the proposed solutions.
Examples of these special procedures or techniques are colorimetric analysis to
determine the physicochemical characteristics of a substance and vigor analysis
to determine the physiological quality of the seed (Silva, 2022a).

1.6.1 Scientific Strategy

The scientific method comprises an ordered set of operations for the

characterization and solution of problems, which is common to all areas of
science. Thus, the scientific method constitutes the strategy of science for the
generation of knowledge. Example 1 illustrates this strategy.

Example 1. Assume the following question: Why is wheat productivity in a

particular region low? A simple answer to this question could be derived from the
empirical observation that the environmental conditions in this region are
unfavorable for the cultivation of wheat. Scientific researchers of this problem
would not be satisfied with simple and generic explanations like this one and
would start by critically examining the problem itself, before trying to find a
solution to it. In fact, that question implies an empirical generalization that can be
refined by breaking it down into less general questions, such as the following
two: Under what environmental circumstances (concerning soil, climate, disease
and pest incidence, etc.) has productivity been low? Under these circumstances,
what are the relevant characteristics of the wheat cultivation techniques (cultivars
used, phytosanitary treatments, fertilization and soil correction, etc.) that may
have implications for productivity? The questions posed in this way are still too
vague and can be further refined by asking more specific questions, such as:

17
 Scientific Research: An Introduction
Science and Scientific Method

Has yield been lower in years of high temperature and relative humidity during
the growing season of wheat? At what stages of plant development are these
conditions most adverse? At what stages of its development is the plant more
susceptible to these climatic conditions? Do these weather conditions favor the
development of fungal wheat diseases? What fungal diseases? Are the cultivars
in use susceptible to these weather conditions? Are they susceptible to these
diseases? (Silva, 2022a)

Thus, an analysis of the overly generic and vague initial problem - low wheat
productivity in the region - leads to a set of more specific problems that have
negative implications for wheat productivity in that region; for example, the
susceptibility of wheat cultivars to high temperature and relative humidity;
incidence of fungal diseases; incidence of rust; susceptibility of cultivars to fungal
diseases. Every simple and precise problem or question that can be solved or
answered with current scientific knowledge and available resources constitutes a
scientific problem or research problem (Silva, 2022a).

Each scientific problem will raise one or more conjectures for a solution or an
answer. Consider, for example, the following problem: damage to wheat
productivity due to the incidence of rust. This problem can raise several
conjectures, such as: a) high temperature and relative humidity favor the
incidence of rust; b) the occurrence of rust can be controlled with the use of
fungicides; and c) the incidence of rust can be avoided with the use of resistant
cultivars. Each of these conjectures that can be verified empirically constitutes a
scientific hypothesis or research hypothesis.

Then, each of these conjectures can be empirically verified through its

consequences. For example, a) if high temperature and relative humidity are
determinants of the incidence of rust and the consequent decrease in
productivity, then wheat crops that differ in those characteristics must present
different degrees of incidence of rust and different levels of productivity; b) if
fungicides control the incidence of rust, then crops with effective fungicides
should be more productive than crops without these fungicides or with ineffective
fungicides; c) if rust susceptibility is an important determinant of low productivity,
then crops that differ in terms of cultivars with different levels of susceptibility (or
resistance) and are similar in terms of other characteristics must have different
levels of productivity (Silva, 2022a).

The verification of each scientific hypothesis may be carried out by scientific

research that will include the observation and collection or gathering of data by
scientific means. For example, in this illustration, through: a) research carried out
on crops, in different locations and over several years, with natural variation in
temperature and relative humidity; b) a survey of available fungicides and a
control (without fungicide); c) a survey of available cultivars with different levels
of susceptibility (or resistance) to rust.

Finally, in each particular research, the merits of the alternatives of the

hypothesis will be evaluated, which may lead to the refutation or non-refutation of

18
 Scientific Research: An Introduction
Science and Scientific Method

that hypothesis. If the observations collected or gathered by the research do not

agree with the consequences derived from the hypothesis, the hypothesis will be
refuted. Otherwise, that is, if these observations are in line with the hypothesis,
the hypothesis will not be refuted. In the latter case, it will be said that the
observations corroborated the hypothesis. Note, however, that a hypothesis is
never proved, as it will always be subject to being refuted by future observation.
Then, the derived scientific knowledge will be incorporated into the previous body
of knowledge (Silva, 2022a).

If a hypothesis has been refuted, it will be necessary to formulate another

hypothesis and restart the procedure; if it has been corroborated, its expansion
or improvement will be desirable.

In general, if research is thoughtful and imaginative, solving the problem that

gave rise to it will raise a new set of problems. The most important and fertile
research are those capable of triggering new questions and not those tending to
stagnate knowledge. In fact, the importance of scientific research is evaluated by
the changes it produces in the body of knowledge and the new problems it
raises.

Example 1 illustrates the general science procedure for acquiring knowledge. In

this process, the following ordered sequence of operations can be distinguished
(Silva, 2022a):

1) Enunciation of well-formulated and fertile questions – scientific problems.

2) Formulation of well-founded conjectures that can be tested through
experience to answer the questions – scientific hypotheses.
3) Derivation of logical consequences from conjectures.
4) Empirical verification of conjectures – scientific research.
5) Analysis and interpretation of the results of the verification of the
conjectures – evaluation of the pretense of the truth of the conjectures.
6) Determining the domains for which the conjectures are valid, incorporating
new scientific knowledge into the available body of knowledge, and
formulating new problems arising from research.

This process of the scientific method is outlined in Fig. 1.1.

Fig. 1.1. Schematic representation of the scientific method process

19
 Scientific Research: An Introduction
Science and Scientific Method

In short, the scientific method starts with examining existing knowledge and
identifying one or more problems of interest. For each of these problems, one or
more hypotheses are formulated. Then, each of these hypotheses is examined
for the logical prediction derivation of consequences that can be objectively
verified. The next step is the objective verification of each of these hypotheses
through new observations. If this objective empirical verification confirms the
prediction concerning a particular hypothesis, evidence accumulates in favor of
that hypothesis, and it is accepted as fact and incorporated into the existing body
of knowledge. Its subsequent life can be short or long, as constantly new
deductions can be drawn and proved, or not, through objective empirical
observation. This circular property of the scientific method is illustrated in Fig. 1.2
(Silva, 2022a).

The fundamental process of the scientific method can be summarized as a

cyclical repetition of stages of synthesis, analysis and synthesis. The scientific
method for solving a generic problem related to a phenomenon starts with a
global view of that phenomenon (synthesis). However, even the most restricted
parts of the universe are too complex to be understood globally and completely
by human effort. It becomes necessary to ignore many of the aspects of the
phenomenon and abstract an idealized version of it, with the expectation that it
will be a useful approximation. Often, certain features of these idealizations are
changed for simplification. This idealization is then broken down into a number of
relatively simple parts into which specific problems for separate treatment can be
identified (analysis). This decomposition aims to identify independent or
interacting parts in a simple way. When the problems relating to these parts are
resolved, the new knowledge is integrated into the existing body of knowledge
(synthesis) (Silva, 2022a).

Fig. 1.2. Diagram illustrating the circular property of the scientific method

20
 Scientific Research: An Introduction
Science and Scientific Method

1.6.2 Scientific Tactics

The scientific method is the common strategy of science. However, the concrete
execution of each of the operations of the scientific method in research requires
a particular tactic that comprises a set of techniques that depend on the topic and
the state of knowledge regarding this topic. These specific techniques change
much more quickly than the general method of science. Furthermore, very often
they are only usable in particular fields of science. Thus, for example, the
determination of nutritional deficiency symptoms in rice plants requires
techniques essentially different from those necessary to obtain plants resistant to
infection with a virus. The effective resolution of the first problem will depend on
the state of the plant nutrition theory, while the second one will depend on the
state of the disease resistance theory (Silva, 2022a).

Scientific techniques can be classified into conceptual and empirical:

- Conceptual techniques are based on definitions, axioms, postulates, laws

and theories. Empirical techniques are related to the observation and
evaluation of characteristics of natural phenomena, through observation
and measurement.

Conceptual techniques make it possible to formulate problems precisely, state

the corresponding conjectures or hypotheses, establish procedures to deduce
consequences from the hypotheses and check whether the proposed hypotheses
solve the corresponding problems. Mathematics offers the richest and most
powerful set of these techniques. These techniques are also powerful in scientific
research into natural phenomena. However, its application requires that scientific
knowledge be sufficiently consolidated to be susceptible to translation and
mathematical treatment. On the other hand, mastery of most empirical
techniques depends only on training. However, talent is needed for its application
to new problems, for the critique of known techniques, and particularly for the
development of new and better techniques (Silva, 2022a).

1.7 OBJECTIVES AND SCOPE OF SCIENCE

Science has two fundamental goals. First, the increase in knowledge - intrinsic,
or cognitive, objective; secondly, the increase of man's well-being and his
dominion over Nature - an extrinsic or derivative objective.

- Science with a purely cognitive aim is called pure science. Applied science
or technology uses the same general method as science and several of its
special methods but applies them for practical purposes.

21
 Scientific Research: An Introduction
Science and Scientific Method

Examples of pure science are physics, chemistry, biology and psychology; of

applied science, electrical engineering, biochemistry, agronomy, human and
veterinary medicine and pedagogy.

This division of science is often challenged, arguing that science is ultimately

aimed at satisfying needs of some kind. However, it is related to the objectives of
the various areas of science and explains the differences in attitude and
motivation between the scientist who seeks to understand better reality and the
scientist who seeks to improve his grip on it (Silva, 2022a).

Science is fundamentally a method of acquiring trustworthy knowledge. In pursuit

of this goal, what it achieves - scientific knowledge - is theoretical knowledge,
that is, an interpretation of reality, not reality itself. This theoretical interpretation
is often expressed in terms of ideal or perfect conditions or forms; for example, a
perfect synthesis of two or more chemical substances, called “solution”, a space
absolutely without matter, called “vacuum”, and an round figure, called “circle”.
These conceivably ideal forms are called models. These models are only
approximations and therefore tentative interpretations of reality. The function of
science is the constant effort to refine and improve such models so that they can
continually approximate reality in terms of growing and more refined empirical
evidence (Silva, 2022a).

Science is based on facts. A scientific fact is a proposition referring to properties

or characteristics of a phenomenon that have been verified empirically and
objectively by the scientific method. However, facts in science are not interpreted
and used in isolation. Rather, they are significantly interrelated by scientific
theories to suggest causal relationships, such as the Mendelian theory of
inheritance and the Newtonian theory of motion. Scientific facts can be used to:
a) suggest new theories; b) suggest revision or rejection of existing theories; and
c) redefine or clarify theories. Thus, scientific facts are the basic elements that
constitute the foundation of reliable knowledge and scientific theory is the
superstructure of that knowledge (Silva, 2022a).

The relationship between theory and fact may not be straightforward. An

interrelated set of facts can constitute an empirical regularity and be formulated
as a scientific law, like the law of inertia and the law of gravity. However, a
scientific theory is a generalized synthetic explanatory statement of the cause of
a phenomenon or the interrelationship between classes of phenomena. In
addition to systematically explaining or taking into account the relationships
between facts and laws, scientific theory also has the function of serving as a
unifying explanation for the possible deduction of hypotheses. Thus, hypotheses
deducible from theory are intermediaries between facts and theories, as are
scientific laws that interrelate verified facts (Fig. 1.3) (Silva, 2022a).

22
 Scientific Research: An Introduction
Science and Scientific Method

Fig. 1.3. Relationships between theory, hypothesis, law and fact

The ambition of science is to develop fruitful theories that cover the widest
possible range of phenomena. The basic objective of science is the creation of
theories that make it possible to explain or predict phenomena. Theory creation
is developed by scientific research with the following functions (Silva, 2022a):

a) Guide the research – it reduces the range of facts to be used and at the
same time determines what kinds of facts are relevant for the research
purposes;
b) Serve as both a conceptualization and a classification system – it allows
the creation of concepts referring to important processes, the classification
of relevant objects (taxonomy) and the creation of structures of concepts;
c) Allow a summary of what is already known about a phenomenon, making
possible an empirical generalization statement or the creation of systems
of relations between propositions (laws, principles, axioms);
d) Suggest the prediction of facts; and
e) Highlight gaps in existing knowledge.

The idea is common that science deals only with “facts” and that its basic
function is to research and reveal the “truth”. This view underestimates the role of
science. The notion of truth has preoccupied epistemologists and philosophers
for many centuries. The difficulty of defining the term “truth” stems from the
assumption that something is inherently or necessarily true, or it is not. However,
the history of human experience has shown very clearly that what is at one time
held to be unquestionably true may later turn out to be unquestionably false (e.g.,
the sun revolves around the earth). Furthermore, at any given time, different
groups can define the same (or apparently, the same) phenomenon very
differently. For example, for some people, it is unquestionably true that criminals
are born bad, or that Orientals are naturally skilled, while for others such notions
have no substantial evidence to support them. This difficulty in defining “truth”,
due to the inherent notion of truth or falsehood, is avoided in science (Silva,
2022a).

A scientific fact is a reliable statement or proposition of truth because it is

supported by objective empirical evidence. Not a certain truth statement or
proposition. Truth in science is never final or absolute. The reliability of a
scientific fact is related to the amount and type of evidence that substantiates it.

23
 Scientific Research: An Introduction
Science and Scientific Method

The reason why all factual knowledge in science is relative rather than absolute
is a consequence of its experiential character. Facts derived from experience
lead to probable truths, never to certain truths because experience is infinite, and
future experience may require a new interpretation of a phenomenon. Thus, the
reliability of a scientific fact depends on the acceptability of the evidence offered.
Some scientific facts are supported by undisputed objective and empirical
evidence, while others are supported by less convincing evidence (Silva, 2022a).

Science seeks to establish conceptual reconstructions of reality through facts. A

scientific law is a conceptual reconstruction of an objective structure; a scientific
theory is a system of such statements. More than that, science aims at a
conceptual reconstruction of the objective structures of phenomena, both current
and possible, which allows for their exact understanding and, therefore, their
technological control. At every step, science performs partial reconstructions,
which are problematic and not demonstrable. With the progress of science, these
partial reconstructions are getting closer to reality (Silva, 2022a).

This process of rebuilding the world through ideas and checking every partial
reconstruction is infinite. proposes a definite and final objective, such as the
complete construction of knowledge without failure. The aim of science is rather
the continuous improvement of its main products - the theories and means - the
techniques.

Thus, scientific knowledge is not a simple accumulation of facts, but a permanent

conceptual revision. Its progress is due to a process of continuous correction.
Scientific activity can be considered as a permanent attempt to decrease the
degree of empiricism and increase the scope of theory. Science is not satisfied
with superficial criteria. It demands a detailed examination of a phenomenon. A
phenomenon is considered known only when it is described and explained with
high accuracy, so that it can be predicted and, if possible, controlled. In
improving knowledge, science pursues four successive objectives: description,
explanation, prediction and control (Silva, 2022a).

The first objective of science is description, that is, the representation of the
phenomenon to identify its essential characteristics. Scientific knowledge begins
with description. Only after acquiring knowledge of the relevant characteristics of
the phenomenon does the explanation of its origins begin. The second objective
is explanation, that is, the characterization of the reason for the existence of the
phenomenon or its cause. Explanation requires identifying the conditions that
result in the phenomenon occurring. The scientist is cautious and conservative.
He recognizes that most phenomena have many causes and that new evidence
may require replacing an older explanation with a more appropriate one. As the
scientific research process evolves, knowledge regarding the causes of
phenomena grows and improves. With this growing knowledge comes the ability
to predict and possibly control phenomena (Silva, 2022a).

The third objective of science is prediction, that is, the anticipation of knowledge
of the phenomenon before its occurrence. The establishment of predictions

24
 Scientific Research: An Introduction
Science and Scientific Method

requires knowledge of the antecedent conditions of the phenomenon. Weather

forecasting, for example, requires knowledge of the weather conditions that
influence it, such as wind direction, temperature, humidity and atmospheric
pressure. The advance in knowledge of the relationship of meteorological events
allows that, nowadays, predictions of meteorological occurrences can be made
with high accuracy. The impossibility of accurately predicting a phenomenon is
indicative of a failure in its understanding. The fourth and most advanced
objective of science is control, that is, the manipulation of the conditions that
produce a phenomenon. Control, in this sense, means knowledge of the
antecedent causes or conditions of the phenomenon. If the antecedent
conditions are known, they can be manipulated to produce the phenomenon as
desired (Silva, 2022a).

1.8 ROLE OF LOGIC IN SCIENCE

Scientific facts are the essential elements that build science. However, they must
be arranged and organized into useful interrelated structures. The most essential
tool of science, along with verified fact, is the system of valid logical reasoning
about scientific facts that allows the derivation of reliable conclusions from them.
These conclusions are propositions about interrelationships of facts that explain
a given phenomenon, which constitutes scientific principles, theories and laws.

At the heart of logical reasoning about facts is a system of rules and prescriptions
that have been established over the course of twenty-five centuries. The correct
use of these rules is fundamental to every scientific effort. The rules of deductive
and inductive reasoning, the correct use of definitions, sampling procedures, etc.
are essential parts of any scientist's intellectual tools (Silva, 2022a).

One of the biggest logical problems involved in scientific reasoning is the

distinction between veracity and validity. This issue is addressed here to illustrate
the very significant interrelationship between scientific facts (that is, statements
of truth) and the logical arrangement between those facts (that is, valid
reasoning) that comprise the theoretical core of the scientific framework.

Most reasoning errors stem from the common tendency to confound truthfulness
with validity. A scientific fact is a certain or probable truth because there is
substantial objective evidence to support it. On the other hand, an argument is
valid when the conclusion necessarily follows from initial propositions or
premises, that is, from presuppositions or assumptions. A person can derive a
wrong conclusion from verified facts if he reasons incorrectly. He can also derive
an incorrect conclusion by correct reasoning if he employs incorrect propositions
as premises. The solution to this problem of logical inference is the valid
argument's mode of operation. There is only one form of argument accepted in
logic as correct (“strong”): one in which the presuppositions or premises are
(certainly or probably) true facts and the inferences drawn from them are valid.
However, there are three forms of incorrect (“weak”) argument (Silva, 2022a).

25
 Scientific Research: An Introduction
Science and Scientific Method

This may explain why so many people draw incorrect conclusions from scientific
facts, and why trustworthy knowledge grows so slowly in so many fields. The four
argument forms, one correct and three incorrect, are illustrated below.

- Correct argument: the presuppositions or premises are (certainly or

probably) true facts and the inferences (conclusions) drawn from them are
valid.
- Incorrect argument: 1) the conclusion is valid, but the premises are false,
2) the premises are correct (scientific facts), but the conclusion is not
valid, and 3) the premises are false, and the conclusion is not valid.

Many examples could be used to illustrate the fundamental role of scientific facts
and reliable reasoning in the structure of science. In physics prior to Galileo, it
was held that heavy objects fall faster than light objects. This statement is a false
fact and is invalid in form. Heavy objects fall faster than light objects only when
they have a higher density, that is, a lower ratio of resistance to the medium per
unit of volume. A 1 kg lead ball falls faster than a 10 kg pack of uncompressed
feathers because its density is greater. In this argument not only is one of the
assumptions a false fact (density rather than absolute weight determines the rate
of fall), but also the shape is invalid because it does not necessarily follow that
just because it is lighter an object must fall faster (Silva, 2022a).

1.9 BASES OF SCIENCE

Interpretations (descriptions or explanations) of phenomena are usually based on

some presumed prior knowledge. Since some facts are required to prove other
facts, all knowledge systems are compelled to prove basic facts. As these basic
facts cannot be proved, they must be admitted as fundamental conventions,
necessary to any logical or epistemological system. These fundamental facts are
often accepted as indisputable (dogmas) or self-evident. This kind of evidence is,
however, a dubious and often unrealistic basis for establishing valid knowledge.
Science rests on basic assumptions supported by logical consistency with
experience, which scientists employ to interpret the evidence needed to produce
verified facts, that is, to derive scientific knowledge (Silva, 2022a).

These basic assumptions are the postulates of science. These postulates should
not be confused with scientific discoveries. They are just functional tools useful
for their purposes, while scientific discoveries are confirmed by objective
empirical evidence. The postulates can be changed over time, if the evolution of
scientific knowledge demands new forms of reference, as new knowledge
frequently changes the state of previous scientific discoveries.

An examination of the literature reveals that there is no agreement regarding the

number and designation of postulates, as there is still no uniform or typical
treatment of the scientific method. The list of eight postulates that follows is just
an attempt to concretize and add what seems to be generally accepted by
competent authorities as essential presuppositions of the scientific method (Silva,
2022a).

26
 Scientific Research: An Introduction
Science and Scientific Method

1) Every event has a natural antecedent ("cause"). Explanations of events

must be sought in natural causes or antecedents, that is, objectively and
empirically demonstrable phenomena. This postulate is employed in
science in the analysis of causality. Its main function is to direct the search
to explain phenomena to the regularities they apparently obey.
2) Nature is orderly, regular and uniform. The belief that the universe
operates according to certain rules of regularity ("natural laws") is inherent
in the scientific analysis of natural phenomena. In practice, this belief
takes the form of explanations expressed in terms of inferred probabilities
from the particular to the general, or from past experience to the present
and hence to the future. According to this postulate, every phenomenon
has an antecedent, and while many phenomena may appear to be unique
(for example, no two storms have identical characteristics), underlying
such unique or inexplicable events are certain models of forces that, when
understood, will allow for better prediction than would be possible by mere
guesswork. Therefore, direct attention must be given to the search for
qualitative and quantitative relationships that exist between natural
phenomena.

This postulate also expresses the apparent fact that nature is not infinitely
complex. Thus, the ordering of scientific knowledge allows the scientist to
develop theories regarding the interrelationships of phenomena and then
proceed to a broader analysis of the universe as a whole.

The implications of this postulate form the basis of scientific logic applied to
natural phenomena. It allows for generalizations and classifications regarding
phenomena and supports the probabilistic basis of inference in science;
particularly it is indispensable for sampling. It also suggests the possibility of
ever-increasing scope for a more highly integrated general theory, which is the
main aim of all scientific endeavors (Silva, 2022a).

3) Nature is permanent. Although apparently everything changes over time,

many phenomena change slowly enough to allow a reliable body of
knowledge to build up. This postulate supports the cumulative attribute of
science. It implies the belief that an event studied today, though perhaps
indeterminately altered tomorrow, will nevertheless be sufficiently similar
to allow valid generalizations to be derived about it that remain reliable for
a period of time.
4) Every objective phenomenon is knowable. That is, given enough time and
effort, no objective problem is unsolvable. This postulate stems from two
related convictions: a) man's intelligence is capable of unraveling the
mysteries of the universe; b) man's search in the mysteries of objective
phenomena has been so fruitful that apparently no door to knowledge is
immutably closed to the continued efforts of scientific research.
5) Nothing is self-evident. That is, reality must be objectively demonstrated.
This postulate asserts that no reliance should be placed on so-called
"common sense", on tradition, on popular authority, or on any of the

27
 Scientific Research: An Introduction
Science and Scientific Method

customary interpretations of phenomena. Historical examples reveal that

apparent veracity is often very different from objective empirical
verification.
6) Truth is relative (to the existing state of knowledge). Evidence in science is
always relative to the state of scientific knowledge, the data, the methods,
the instruments used, the frames of reference and, therefore, to
interpretation. Therefore, the "truth" in science is simply an expression of
the best demonstrable professional judgments at any given time. This
postulate does not imply that stable knowledge cannot be acquired; but it
recognizes that knowledge is dynamic and that, as knowledge grows in
quality (that is, it becomes more highly verified) and in quantity,
reinterpretations and new conclusions about phenomena become
imperative. This attribute has encouraged the constant re-evaluation of
ideas both old and new and has allowed for the extraordinary growth of
science.
7) All perceptions are acquired by the senses. That is, all knowledge is
acquired from sensory impressions. The elements and instruments of
reasoning (that is, ideas, concepts, constructions, images, etc.) are
shaped by the impressions received by the senses. This postulate also
ensures that the only reliable knowledge is that which is objectively and
empirically verifiable.
8) This postulate originated from the influence of Galileo regarding the
demonstrability of theoretical predictions. The empirical demonstration
became the essential test of the validity of all theoretical speculation
concerning phenomena and resulting predictions.
9) Man can believe in his perceptions, memory and reason as a means of
acquiring facts. This postulate supports the entire rational and empirical
basis of scientific knowledge. It does not imply that any or all perceptions,
memories and reasons are trustworthy. This postulate asserts that the
final resolution of any dispute over phenomena must be based on
accepted rules of reasoning and on sense-perceived data, not mere
notions and ideas. The ultimate belief in the analysis of phenomena must
be based on empirical evidence interpreted according to the rules of
logical reasoning.
___________________________________________________________________________________
© Copyright (2024): Author(s). The licensee is the publisher (B P International).

DISCLAIMER
This chapter is an extended version of the article published by the same author(s) in the following journal.
International Journal of Science and Research, 11(4), 2022. DOI: 10.21275/SR22412084104

28
 Chapter 2
Print ISBN: 978-81-976007-6-0, eBook ISBN: 978-81-976007-1-5

Scientific Research
DOI: [Link]

ABSTRACT

The scientific method is the strategy of science to generate knowledge. Its

implementation is carried out by scientific research. Scientific research is
employed in the three phases of the scientific method: in the initial phase of
synthesis, to identify scientific problems; in the analysis phase, to try to solve
these problems; and in the final synthesis phase, to incorporate the new
knowledge generated into the existing body of knowledge. Thus, scientific
research comprises systemic research, which focuses on global systems, in the
initial and final phases, and analytical research, which focuses on subsystems, in
the analysis phase. New scientific knowledge is achieved by analytical research,
which begins with the formulation of a scientific problem. Once the initial question
is posed, the systematic structure of the scientific approach becomes
immediately evident. This is because scientific research is very precisely
structured and logically organized. This form has been developed over centuries
of scientific problem-solving experience. This chapter summarizes the essential
conceptual and methodological basis of scientific research.

Keywords: Natural phenomena; generation of knowledge; investigation of

nature; tactics of science; observation and reasoning; role of
researcher in science.

2.1 INTRODUCTION

Science aims to increase knowledge and improve understanding of natural

phenomena, with the purpose of their control and, in the failure of this, prediction,
in order to allow the increasing dominion of man over nature. The strategy of
science to produce knowledge is the scientific method. Its implementation is
carried out by scientific research.

The contribution of scientific research to the advancement of knowledge and for

humanity to reach the present stage of civilization is recognized. In fact, one can
clearly identify the close relationship between man's growing dominion over
nature and the scientific effort and the evolution of the methodology and
instruments of science in recent centuries (Silva, 2022b).

Scientific research is not a highly controlled production process. There is no set

of rules that guarantee the success of scientific research. However, one can
clearly detect the indispensability of solid basic training of the researcher, which
includes the understanding of the conceptual and methodological foundations of

29
 Scientific Research: An Introduction
Scientific Research

scientific research, and the knowledge of techniques consolidated by the

experience acquired by the scientific research work carried out in the past.

This chapter provides a brief review of scientific research. The topics that seem
most important are considered: the basic concepts, the two essential elements of
scientific research - observation and reasoning, the process and organization of
the scientific research, the objectives of a scientific research, the stages and
orderly structure of an analytical scientific research, research work and the role of
researchers in science, and statistics in scientific research (Silva, 2022b).

This chapter is based on Silva (2020b) and the contributions listed in the
References.

2.2 BASIC CONCEPTS

2.2.1 Scientific Research

The implementation of the scientific method to produce knowledge is carried out

by scientific research (Silva, 2022b):

- Scientific research is the process of investigating natural phenomena

using the scientific method to discover new facts and develop scientific
theory.

Scientific research also aims at the review of facts, laws and theories in view of
newly discovered facts, and the practical applications of such facts, laws and
theories. Therefore, scientific research is the continuous search for knowledge
and understanding of reality carried out through the scientific method. Its result is
scientific knowledge.

Scientific research can have a purely cognitive objective, that is, the generation
of scientific knowledge without an immediate application purpose, or a practical
objective, that is, the generation of knowledge for immediate application:

- Scientific research with purely cognitive objective is called pure research

or basic research, and with practical objective, applied research or
technological research.

The progress of scientific knowledge basically consists of the permanent and

progressive deepening of knowledge of the complex interconnection of natural
phenomena, which, in turn, generally comprise an extremely complex set of other
more elementary phenomena, also closely related.

2.2.2 Characteristic

One basic property of phenomena in nature is their heterogeneity, which

characterizes what is commonly called "natural variability".

30
 Scientific Research: An Introduction
Scientific Research

The variability of phenomena comes from the heterogeneity of their attributes

and properties (Silva, 2022b):

- The attributes and properties of phenomena are the characteristics of

these phenomena. Each characteristic can manifest under different
alternatives or levels.

Scientific research focuses on a class of phenomena with common particular

characteristics. For example, sex is a characteristic of a sheep herd and of the
animals of a herd. This characteristic can manifest itself in each of these animals
in one of two levels - male and female. Body weight at slaughter is another
characteristic of this herd and of its animals, which, for each animal, can assume
any numerical value in a certain interval of real numbers.

Characteristics are qualitative or quantitative, depending on whether their levels

are related qualitatively or quantitatively. Qualitative characteristics classify units
into classes or categories, according to their levels; quantitative characteristics
distinguish units quantitatively. Thus, in the example in the previous paragraph,
sex is a qualitative characteristic that classifies animals into two classes; body
weight at slaughter is a quantitative characteristic that distinguishes animals
according to the magnitude of their weight.

2.2.3 Unity or System

Each scientific research focuses on a class of phenomena with common

characteristics. A technical-operational characterization of the elementary
constituent of such a class of phenomena, which usually receives the designation
of unit or system, and of its elements and attributes, is convenient (Silva, 2022b):

- A unit or system is a set of related entities that constitute a globally

organized whole dynamically related to the external environment, and
which together perform some function.

A unit is specified by describing the following (Fig. 2.1):

function or objective of the unit,

inputs - elements that enter the unit,
products - elements that leave the unit,
components - internal elements that transform inputs into products,
flow - movement of elements between the components of the unit, and
border or limit - imaginary line that demarcates the scope of the unit, which
includes all its components and elements.

31
 Scientific Research: An Introduction
Scientific Research

Fig. 2.1. Schematic representation of a unit or system

The following examples are illustrative: a) in research of the incidence of

brucellosis in animals in a region, the unit can be an animal, a herd, a farm, or a
county; b) in peach fruit fly control research, the unit may be a plant or an
orchard.

The definition of the unit depends on the research objective. It must be

established in the initial phase of the research, that is, in the formulation of the
problem. Sometimes the definition may be obvious, as in research for the
recommendation of wheat cultivars, where the unit is a crop, and in research of
the effectiveness of a dewormer for the control of helminths in dairy cows, where
the unit is an animal. However, very often, the definition or choice of unit is not so
obvious (Silva, 2022b).

The difficulty in defining the unit or system for research stems from the fact that
in nature there is a hierarchy of systems, that is, systems within systems, in
decreasing order of amplitude, such that a given system is a subsystem with
respect to a higher hierarchical level, and, in turn, contains subsystems at a
lower level. In these circumstances, the definition of the unit must establish the
demarcation of its limit. The units of a class of phenomena are usually complex
and heterogeneous entities. For this reason, the designation of the system is
more suggestive than that of the unit, which, however, is more usual (Silva,
2022b).

Units (systems) are identified and distinguished by a set of characteristics, whose

description constitutes a useful working definition of a unit:

- The units of a class of phenomena comprise a set of characteristics that

identify and distinguish them.

Often, the unit is identified by one of its characteristics or aggregate of

characteristics. However, it must be clearly understood that the complete
definition of the unit must encompass the whole set of characteristics that
comprise it. For example, in research with sheep, the unit comprises the
characteristics of animals (race, sex, weight, wealth, …), environment
(installations, climate, pasture, diseases, predators, …), management (feed,
water, wormers, …) and measurement.

32
 Scientific Research: An Introduction
Scientific Research

2.2.4 Target Population

Scientific research aims to solve problems concerning the units of a collection of

units of specific interest (Silva, 2022b):

- In scientific research, the target population, or more simply, population, is

the well-defined collection of units (systems) of interest for which it is
desired to infer. The number of units is called the size of the population.

The target population is defined by specifying its units or the conditions for units
to be part of it. The specification of the population, as well as of its units, is
determined by the research objectives and must be established in the formulation
of the research problem. Populations existing in nature are finite populations,
whose size is expressed by a natural number, very often large and unknown.
Furthermore, such populations have a dynamic constitution, as a result of the
mutability of the systems that comprise them over time. In some research, the
target population consists of units existing at the time the research is carried out.
A population in these circumstances, whose units can be identified, is a real
population. Very often, however, the target population comprises units that do not
exist at the time the research is being carried out, but which, presumably, may
exist in the future. A population whose units are not identifiable, but defined by
the conditions to belong to it, is a conceptual population. In wheat breeding
research, for example, the units of the target population are not the wheat crops
existing in the region of interest at the time the research is carried out, but the
crops that will exist in that region in the future.

2.2.5 Sample and Sampling

In some situations, the research is carried out on all units of the target
population. Research in these circumstances is called a census. Very often,
however, it is impracticable or inconvenient to conduct the research on all units of
the target population. In these situations, research is carried out on a set of units
chosen from the target population or constructed to represent it (Silva, 2022b):

- A subset of the units of a target population or a set of units constructed to

represent it is a sample of this population. The number of sample units is
the sample size. The process of choosing or building a sample is called
sampling.

The characteristics of the sample are the same as those of the target population.
The levels of each sample characteristic can be the same as the target
population or a subset of it.

The choice of the sample can be determined by an objective and random or

subjective and arbitrary sampling process.

33
 Scientific Research: An Introduction
Scientific Research

- The sample is a random sample if it is determined by random sampling,

that is, by a process that assigns to each unit of the target population a
known (non-zero) probability of constituting the sample.
The execution of a research through a sample aims to obtain information about
the target population that is supposed to be represented by this sample.
Inferences from the sample usually do not coincide with inferences that would be
derived directly from the target population.

- The discrepancies between the inferences derived from a particular

sample and those that would be obtained from the target population, that
is, the inference error resulting from the observation of a sample instead of
the target population, constitute the sampling error.

The average of inferences from all possible samples from a target population of a
given size should coincide with the inferences that would be provided by the
population itself.

- Inferences derived from a sample to the target population are valid

(unbiased) if the average of the inferences that would be provided by all
samples of the same size from that population are the same as those
that would be provided by the population itself.

In general, two sample properties are desirable: validity and representativeness:

- The sample is valid if it provides valid inferences for the target

population; it is a representative sample if the distributions of the
relevant characteristics in the sample are equal to the respective
distributions in the target population, except for random variation due to
sampling.

The sample is valid if it is a random sample; it is representative if it has a

structure consistent with the structure of the target population and supplies
plausible inferences. A valid and representative sample is critical for inferences.
When the sample is constituted by units of the target population, these properties
can be achieved using adequate design and random sampling. Very often,
however, random sampling is unfeasible or undesirable, due to the impossibility
or inconvenience of considering all units of the target population in the sample
selection process. In these circumstances, the sample is determined by non-
random sampling, in which the researcher uses judgment to achieve the best
possible representation. This is the case with conceptual populations, whose
units do not exist at the time of sample selection. In these circumstances, in
some situations, the sample units are not units of the target population, but units
built to simulate them, especially for carrying out the research (Silva, 2022b).

It is important to characterize the scope for which the sample inferences are
valid:

34
 Scientific Research: An Introduction
Scientific Research

- The collection of units for which the sample provides valid inferences is
called the sampled population.

Inferences from the sample to the target population are valid when the sampled
population is the target population itself. This is the case for random sampling
research. However, in non-random sampling situations, the validity of inferences
for the target population must be subjectively evaluated.

In general, the process of inference from the sample to the target population
comprises two steps: generalization from the sample to the sampled population
and generalization from this population to the target population. Generalizations
referring to the first step, i.e., from the sample to the sampled population, are
conceptually valid and can be derived by objective statistical procedures.
Extensions of these generalizations to the target population are valid if the
disparity between the sampled population and the target population, that is, the
sampling error, is irrelevant. This is the case in research with random sampling.
However, in research with non-random sampling, the judgment of the validity of
this second step of generalization to the target population is necessarily
subjective (Silva, 2022b).

2.3 OBSERVATION AND REASONING

The generation of knowledge through scientific research essentially takes place
through observation and reasoning:

- Observation is the sensory perception by which knowledge of phenomena

is obtained. Reasoning is the mental elaboration that conceives and
discovers the meanings of these phenomena, their interrelationships and
relations with the existing body of scientific knowledge, insofar as the
present knowledge and the researcher's skill allow.

A basic postulate of science is that all information concerning phenomena is

derived from sense impressions. The data of science are mental impressions of
sensory experiences, that is, ideas derived from the senses, mainly from sight,
hearing, smell, taste and touch. A man reacts to sense impressions by mental
manipulations, and these reactions or ideas constitute the foundation of
knowledge. Knowledge generation in science is essentially a process that
comprises (Silva, 2022b):

a) stimulus reception of the phenomenon that is studied,

b) mental manipulation of the impressions of this stimulus to interpret it,
c) combining these impressions with other earlier impressions and their
interpretations with the aid of memory, and
d) deduction of a conclusive interpretation of the phenomenon.

This process involves two basic elements: the sensory organs, aided by
instruments, and the mind. The sensory organs employed depend on the
properties of the phenomenon, while mental manipulations are related to the
researcher's knowledge, mental ability, awareness, interest and attitude.

35
 Scientific Research: An Introduction
Scientific Research

2.3.1 Observation

Man has limited power of observation. For this reason, it is necessary to limit the
observations to the relevant characteristics. Some characteristics can be
observed with the use of simple instruments to aid the senses, such as a scale,
ruler and test tube. However, others require the use of more sophisticated
observation instruments, such as a stethoscope, microscope and telescope
(Silva, 2022b).

Before proceeding with observation, the researcher must clarify and establish
basic definitions relevant to scientific observation, particularly the following:

a) the phenomena that must be observed, the relevant behavior of these

phenomena and the significant facts that must be addressed,
b) the places, moments and conditions in which the observation must be
carried out, and the way of structuring and describing the observed
situation,
c) the relevant characteristics of the phenomena to be observed,
d) the operational definitions and symbolic representations of these
characteristics, and the procedures and instruments that should be used
for measurement,
e) the stabilization and standardization of observation conditions to allow
verification.

The observation will be truly scientific only when all these definitions are
established. However, it will only be successful if it achieves the necessary
accuracy, that is, if it portrays the phenomenon as it really is. Accuracy demands
some basic requisites:

a) sensitivity,
b) objectivity,
c) use of sensory aids and
d) definition of the relevant characteristics to be observed.

The sensitivity to observe is a combination of experience and insight. Experience

directs the observation of phenomena to their empirical properties and interprets
them rationally; insight refers to the ability to perceive these properties through
discernment and understanding of their inner nature and reality. Objectivity is
achieved using calibrated and impartial observation instruments and with
impersonal attitudes, actions and judgments. The use of sensory aids is essential
to clarify and discern what is observed. The definition of relevant characteristics
implies questions of conceptualization, validity and relevance. Decisions on these
issues must be made in accordance with the research objectives and are closely
related to the problem and the scientific hypothesis (Silva, 2022b).

The observation of a characteristic of a set of units comprises the description and

representation of that characteristic through symbols that express its levels and
allow classifying or quantifying these units. This process is called measurement.

36
 Scientific Research: An Introduction
Scientific Research

[Link] Measurement

- The measurement of a characteristic is a process of description and

representation of this characteristic that expresses its levels by symbols
that present among themselves the same relevant relationships existing
between those levels.

Thus, measurement demands the definition of a correspondence rule between

the levels of the characteristic and the symbols that correspond to them (Silva,
2022b):

- A rule or function that establishes a one-to-one correspondence between

the set of levels of a characteristic and a set of symbols is called a
variable. Each of these symbols is a level of this variable.

Usually, the variable receives the same name as the characteristic it expresses,
and often these two terms are used interchangeably. However, characteristics
and variables are different concepts. In fact, the same characteristic can be
expressed by many alternative variables, depending on the measurement level
adopted and the set of symbols chosen. The choice of one of these alternatives
is arbitrary and depends on convenience and availability of measurement
resources. For example: a) In a research with sheep, the sex characteristic can
be expressed by a variable that makes an animal correspond to the initial letter of
its sex, that is, the symbol m or f, depending on whether the animal is male or
female, or a variable with levels 0 and 1, which assigns the symbol 0 to one of
the sexes and 1 to the other, or by any variable with two levels; the characteristic
amount of parasites in the feces can be expressed by a variable that assumes
values of a subset of non-negative integers with zero lower end; and the body
weight of the animal at slaughter by a variable with values of an interval of real
numbers (a ; b), where a and b are, respectively, numbers lower and higher than
the expected minimum and maximum weights. b) In wheat research, the
characteristic degree of Gibberella infection can be expressed by a two-level
variable - without and with infection, or a variable that expresses the degree of
infection, such as no infection, weak, medium, strong and high infection; or a
variable that expresses the percentage of infection with values in the interval [0;
100] (Silva, 2022b).

The measurement of a characteristic involves two problems: what to measure

and how to measure it. The first problem, that is, what to measure must be
solved according to the research objectives and is closely related to the scientific
problem and hypothesis. A characteristic is an abstract notion or concept of a
property of the units of a target population, which can be simple or complex,
objective and easily interpretable or subjective and of dubious interpretation.
Therefore, the measurement of a characteristic requires the translation of a
theoretical concept into an operational equivalent. Disagreements may arise from
different operational definitions, presumably of the same characteristic. This can
occur when a concept expresses a complex characteristic. Suppose, for
example, that a theory holds that the productivity of a cultivated plant species is

37
 Scientific Research: An Introduction
Scientific Research

positively related to the quality of the seed. The difficulty is finding agreement
regarding an operational equivalent of seed quality. Several useful and
convincing operational definitions can be proposed, but there may be no logical
way to determine which of these definitions is the most valid (Silva, 2022b).

The second problem of measurement, that is, how to measure, is related to the
first: what to measure. The problem now is how to classify or quantify, according
to a characteristic, units presumably already classified for that characteristic. This
classification or quantification is obtained by representing the characteristic
through a variable.

Observation depends on some degree of subjective judgment. For this reason, it

is important to establish the conditions of observation in order to avoid observer
bias. Elaborate strategies must often be established to enable the observer to
avoid his own bias and obtain the correct record of the facts (Silva, 2022b).

Measurement Scales:

Measurement of a characteristic is an attempt to find a correspondence between

the alternatives of the characteristic and the levels of a variable, that is, a set of
symbols or numbers that properly considers the relationships between these
alternatives and the operations that can be performed on them. Variables are
classified into four measurement levels (measurement scales) according to the
relationships between their levels and the operations allowed between these
levels: nominal, ordinal, interval and ratio (Silva, 2022b).

A nominal variable classifies units into classes or categories as to the

characteristic it represents, without establishing any relationship of magnitude or
order. This attribute only differentiates units that belong to different classes. Two
units are assigned the same symbol in relation to a variable if they have the
same level of that characteristic. Thus, all units of a class are equivalent with
respect to this characteristic. This is the property of equivalence or equality of
classes or levels. Examples are sex (male, female), breed of sheep (Romney
March, Corriedale, Suffolk, ...), and numbers of boxes in a sheep facility (1, 2, 3,
...). Note that the levels of a nominal variable are just symbols and, even if
symbols are numbers, there is no order relationship between them. Therefore,
arithmetic operations on these numbers have no meaning with respect to the
objects they identify. The nominal scale allows only a few elementary operations,
such as the count of the number of units (frequencies) of the classes (number of
animals of each breed, for example) and determining the most numerous class,
that is, the mode of distribution. Statistical inferences regarding the distribution of
these frequencies can be carried out. However, because a nominal scale does
not infer degree or quantity, it only classifies the units, and classes cannot be
manipulated mathematically (e.g., by adding or subtracting numerical equivalents
from those classes). Consequently, most of the usual statistics such as mean
and standard deviation are meaningless, as operations for their determination
are not permissible (Silva, 2022b).

38
 Scientific Research: An Introduction
Scientific Research

Nominal variables have the following three properties regarding the equality or
equivalence of units in the classes: 1) reflexivity - a unit is equivalent to itself, that
is, if a is the value of a unit of class A, then a=a; 2) symmetry - if unit A is
equivalent to unit B, then unit B is equivalent to unit A, that is, if a and b are
values of two units A and B, respectively, then a=b implies b=a; 3) transitivity - if
A and B are equivalent units and B and C are equivalent units, then A and C are
equivalent units, i.e. if a, b and c are the values of three units A, B and C,
respectively, then a=b and b=c implies a=c (Silva, 2022b).

An ordinal variable classifies units into classes or categories according to the

characteristics they represent, establishing an order relationship between the
units. For example, animals in a herd can be classified by the degree of tick
infestation, assigning the number zero to indicate no infestation and the numbers
1, 2 and 3 to indicate increasing levels of infestation, such as low, medium and
high. Thus, for example, if a unit belongs to category 2, it is known that its degree
of infestation by ticks is lower than that of a unit belonging to category 3, as there
is the same order relationship between these categories. Examples of ordinal
variables are also: age with child, youth, adult and elderly levels, and academic
performance grade with levels A, B, C, ... (Silva, 2022b).

The ordinal scale maintains the properties of the nominal scale regarding the
equivalence of units assigned to the same order: reflexivity, symmetry and
transitivity. In addition, the ordinal scale has the property of magnitude of
classes, which implies the property of asymmetry of classes. Therefore, units can
be designated not only as equivalent to other units but also as non-equivalent.
For example, an ordinal scale can designate that class A is higher than class B,
and therefore that class B is smaller than class A.

The additional properties of the ordinal scale characterize its superiority over the
nominal scale. However, arithmetic operations cannot be performed with symbols
that only characterize order and vaguely designate quantity. Consequently,
statistical descriptions are still limited. Position measures are restricted to the
median, quartile, percentile, and other measures dealing with orders. Statistical
inference procedures are restricted to those specifically appropriate for order
data (Silva, 2022b).

An interval variable orders the units according to the measured characteristic

with numeric values. Thus, it allows the determination of differences between
measured units. However, the zero value of a variable with an interval scale does
not mean the absence of the measured characteristic. Consequently, the origin
or zero point of this scale is arbitrary. In addition, ratios and proportions are
meaningless.

Examples of interval variables are temperature measured in degrees Celsius or

Fahrenheit. A day with zero temperature measurement on the Celsius or
Fahrenheit scale is not a day without temperature. Each of these two scales
assigns an arbitrary zero value, which is just a reference point. Furthermore, if in
one day the temperature is 15 degrees centigrade and another 30 degrees, this

39
 Scientific Research: An Introduction
Scientific Research

does not mean that on the second day the heat is twice as high as the other day;
in fact, this ratio is only verified on this measurement scale. The Fahrenheit scale
is related to the Celsius according to the expression TF=(9/5)TC+32. Thus, the
value of zero degrees on the Celsius scale corresponds to 32 on the Fahrenheit
scale, and the values of 15 and 30 degrees on the Celsius scale correspond to
59 and 86 on the Fahrenheit scale, that is, to a difference of 15 degrees on the
Celsius scale corresponds a difference of 27 degrees on the Fahrenheit scale
(Silva, 2022b).

The interval scale allows arithmetic operations. Consequently, it enables most

descriptive statistical measures such as mean, standard deviation, and
correlation coefficient. For example, if the average temperature of a city in a
month is determined in degrees Celsius or degrees Fahrenheit, the days of the
month with above-average temperatures are the same under each of these two
scales. However, some statistical measures, such as the coefficient of variation,
can be misleading when applied to interval-scale variable data.

A ratio variable ranks the units in terms of the measured characteristic, has a
constant unit of measurement and an absolute zero point, which is a unique
origin. So, unlike the interval scale, the ratios have meaning. This is the most
elaborate of the measurement scales, in the sense that it allows all arithmetic
operations. It is the most common measurement scale in the physical sciences,
such as scales to measure length, width, height, weight and volume. As the
designation suggests, equal ratios between values on the ratio scale correspond
to equal ratios between the measured units. Thus, ratio scales are invariant
under positive proportion transformations, that is, transformations of the form y=
c x, c>0. For example, if one plant is 3 m tall and the other 1 m, it can be said
that the first plant is 3 times taller than the second. This is because, if the heights
of the two plants are converted into centimeters, their measurements will be,
respectively, 30 cm and 10 cm, which are in the same 3:1 ratio. One can convert
measurements from one ratio scale to another ratio scale merely by multiplying
by an appropriate constant (Silva, 2022b).

The ratio scale contains all the information of the lower order scales, namely,
class equality, order and difference equality, and more. All descriptive statistics
can be determined for data of one variable expressed in ratio scale.

Variables are also classified into two classes: qualitative and quantitative.
Variables with nominal and ordinal measurement scales are qualitative variables
because their levels are qualitatively different. They are also called categorical
variables, as their levels designate categories or classes in which the units are
classified according to the characteristics represented. Interval and ratio scale
variables are quantitative variables because their levels are quantitatively related.

Variables can also be classified as discrete and continuous. A variable that takes
on isolated values from a set of real numbers, that is, values from a discrete set
of real numbers, is a discrete variable; a variable that can take on any value in an
interval of real numbers is a continuous variable. Qualitative or categorical

40
 Scientific Research: An Introduction
Scientific Research

variables, that is, of nominal or ordinal measurement scale, are, by definition,

discrete and finite, that is, they assume a finite set of distinct values. Variables
with an interval or ratio measurement scale can be discrete, assuming a finite or
infinite set of distinct values, or continuous. For example, variables that express
the amount of fruit on a tree and the number of piglets in a litter are discrete and
finite; variables that express the weight and height of an animal or a plant are
continuous variables (Silva, 2022b).

Note that measurement with an interval or ratio scale variable can be considered
as a refined form of classification. If lambs in a herd are weighed with a scale to
the nearest grams, the result is a distribution of the animals among discrete
weight categories, where neighboring categories differ by one gram. If weighing
is carried out on a scale to the nearest milligram, the number of possible
categories increases considerably, but the principle remains the same. Thus, in
practice, the measurement process, even of conceptually continuous
characteristics, constitutes a classification of units.

Choice of Measurement Scale:

For some characteristics, the choice of measurement scale is limited in nature.

More often, however, a characteristic may alternatively be expressed by more
than one variable with measurement scales of different levels of precision. For
example, a) the sex and race of an animal are necessarily expressed on a
nominal scale; b) The seed size of a cultivar can be expressed by a two-level
nominal variable: 0 - abnormal, 1 - normal, or by an ordinal variable: 1 - very
small, 2 - small, 3 - medium, 4 - large and 5 - very large, or by a ratio variable
whose levels are the numbers of a certain range of real numbers. If a
characteristic can be alternatively expressed by variables of nominal, ordinal,
interval and ratio scales, the ordering of the amounts of information provided by
individual measures by these variables is as follows: nominal < ordinal < interval
< ratio (Silva, 2022b).

The choice of measurement scale must consider the importance of the

characteristics and theoretical and practical considerations. From a theoretical
point of view, the highest-level measurement scale should always be used, that
is, the ratio scale. However, practical considerations must also be considered,
particularly the importance of the variable and the availability of resources, such
as measurement instruments and qualified personnel to perform the
measurement. Statistical considerations are also important, as data from the
research will have to be subjected to statistical analysis. In general, the statistical
methodology is more developed and popularized for normally distributed
variables, which are continuous variables with a ratio scale. In many studies,
intrinsically quantitative characteristics are expressed by ordinal scale variables
whose levels are subjectively assigned by an evaluator. This is often the case for
characteristics that comprise the degree or intensity of a property as measured
by visual assessment, such as the degree of disease infection or pest infestation.
To represent such a characteristic, it is common to define a variable with the
lowest level equal to zero to express the null degree or intensity, such as the

41
 Scientific Research: An Introduction
Scientific Research

absence of the symptom; increasing intensities are expressed by the following

integers, i.e. 1, 2,... ; for example, 0 - no infection, 1 - mild infection, 2 - medium
infection, 3 - high infection, and 4 - very high infection. These variables are
generally inconvenient, as they express a continuous characteristic very
imprecisely and because of the difficulties that arise for statistical inference
procedures. Continuous characteristics should be expressed, whenever possible,
by continuous variables, even in subjective measurements. For example, the
evaluation of the intensity of rust infection in wheat leaves can be done in
percentage, with a continuous measurement scale of values between 0 and 100,
using as patterns figures of leaves with infection intensities corresponding to
arbitrary points on this scale. Infection intensities are measured by comparing the
appearance of leaf samples with these standards and interpolation on this
continuous scale (Silva, 2022b).

Precision and Accuracy of a Measurement Process:

- The precision of a measurement process of a characteristic refers to the

proximity of measurements of that characteristic carried out repeatedly in
the same unit.

The measurement of a characteristic is more precise the closer its different

measurements referring to the same unit are located. Thus, low precision means
a large variation of repeated measurements; high precision means a small
variation of repeated measurements. For example, if repeated measurements of
the weight of an object carried out with scale A are quite close and those carried
out with scale B vary considerably, the measurement process with balance A has
higher precision than with balance B.

The precision of a measurement process does not refer to correct assessment: a

measurement process of a characteristic can be highly precise, determining
successive measurements that are very close to each other, but far from the true
value of that characteristic. A measurement process in these circumstances and
the measures it determines are said to be biased. The difference between the
mean value of repeated measures of a characteristic in one unit and the true
value of that characteristic is called bias. A biased measurement process
systematically overestimates or underestimates the true value of the
characteristic (Silva, 2022b).

- The accuracy of a measurement process of a characteristic refers to the

proximity of measurements of that characteristic carried out repeatedly in
the same unit in relation to the actual value of that characteristic.

A measurement process is more accurate the closer to its true magnitude are
repeated measurements performed on a unit by this process.

Accuracy is higher the higher the precision and the lower the bias. High accuracy
corresponds to little or no bias and high precision. Low accuracy can result from
a large bias and high or low precision, or a null or low bias and low precision. For

42
 Scientific Research: An Introduction
Scientific Research

example, weighing performed with a low precision balance has low accuracy,
regardless of the presence or absence of bias; weighing performed with a high-
bias balance also has low accuracy, regardless of the precision level of the
weighing. Weighing performed with a balance has high accuracy only if it has
high precision and low bias (Silva, 2022b).

2.3.2 Reasoning

Facts are the essential building blocks of science. However, they must be
arranged in useful and interrelated structures. The most essential tool of science,
along with verified fact, is the system of valid logical reasoning about facts that
allows reliable conclusions to be derived from them. These conclusions are
propositions about the interrelationships of facts, which constitute scientific
principles, laws and theories (Silva, 2022b).

At the heart of logical reasoning about facts is a system of rules and prescriptions
whose correct use is fundamental to every scientific endeavor.

- The reasoning process used in scientific research starts with one or more
propositions and proceeds to one or more other propositions whose
veracity is believed to be implied by the veracity of the first set of
propositions. This psychological process is called inference.

The validity of this implication depends on the logical relationship between the
propositions. The absence of an implicative relationship between the propositions
can lead to false inference. Two processes of inference are fundamentally
distinguished: deductive inference and inductive inference:

- Deductive inference is the process of reasoning in which one starts from a

general proposition or set of propositions and proceeds to a specific
proposition or set of propositions. Or, put another way, it is the process of
deriving knowledge of a specific member of a class from general
knowledge concerning all members of the same class.
- Inductive inference is the process of reasoning in which one starts from a
specific proposition or set of propositions and proceeds to a general
proposition or set of propositions. That is, it is the reasoning process by
which knowledge of some members of a class is applied or extended to all
unknown members of the same class (Silva, 2022b).

The discovery of the nature of inference came relatively late in human culture.
Aristotle (384-322 BC) was the first to stress the systematic nature of science,
stating that it can be developed by reason alone. He established the logical
scheme of deductive reasoning called syllogism. This method of deductive
inference starts with two propositions called the main premise and secondary
premise that are logically related in such a way that from them a third proposition,
called the conclusion, can be derived. For example:

43
 Scientific Research: An Introduction
Scientific Research

Main premise: All living plants absorb water (induction).

Secondary premise: This peach tree is a living plant (observation).
Conclusion: Therefore, this peach tree absorbs water (deduction).

The inductive method is used to generalize from known and verified phenomena
of a given class to unknown and not yet verified phenomena of the same class.
Induction involves formulating a generalization. This generalization step is called
“inductive jump”.

Inductive reasoning is commonly employed whenever a judgment is made about

a situation based on experience with a presumably similar previous situation. For
example, when lightning is observed and then the sound of thunder is expected,
and when a cultivar is selected for planting on the assumption that it will have a
similar yield to the previous years. The bases of these claims are the postulates
of uniformity and permanence of nature. Although strictly every event is unique
and therefore not repeatable in the future, from a practical point of view many
events show similarities in some essential characteristics. Thus, the observation
of the succession of events provides a basis for the evaluation, in terms of
chance, of the possibility of the occurrence of an event when other events that
preceded it in the past occur with known frequency (Silva, 2022b).

In summary, induction in science is based on incomplete evidence, due to the

impossibility of considering all units of the target population. The conclusions are
only probable, to a greater or lesser extent, depending on the number of units
considered and the precautions taken in their selection. The relevant fact is that,
in well-planned and executed research, evidence of inductive inference can be
established on a probabilistic basis by statistical methods.

2.4 SCIENTIFIC RESEARCH PROCESS

The generation of knowledge about the units of a target population through the
scientific method comprises three phases: synthesis, analysis and synthesis. The
initial synthesis phase consists of the identification of scientific problems through
the successive decomposition of these systems into sufficiently simple
subsystems to which important specific problems correspond that may be
susceptible to attempts at solutions based on existing scientific knowledge and
with the available resources. In the analysis phase, for each of these scientific
problems, a conjecture of solution is chosen that can be verified empirically by
scientific research. In the final synthesis phase, the new knowledge generated is
incorporated into the body of knowledge about the systems (Silva, 2022b).

Scientific research encompasses all phases of the scientific method: it is used to

identify scientific problems, to try to solve these problems, and to incorporate the
new knowledge generated into the existing body of knowledge. Thus, it
comprises systemic research, which focuses on global systems, in the initial and
final phases, and analytical research, which focuses on subsystems, in the
analysis phase.

44
 Scientific Research: An Introduction
Scientific Research

Analytical research is necessarily carried out in the scientific research, but

systemic research is little used.

The scientific research process is dynamic and reiterated over time. When new
knowledge generated from research is incorporated into current systems, new
research problems are identified, and the process is reiterated (Silva, 2022b).

2.5 OBJECTIVES OF SCIENTIFIC RESEARCH

The objective of systemic research, in the phases of synthesis of the scientific

method, is exploratory. The objective of analytical scientific research is the
verification of a conjecture or solution to a particular scientific problem. Thus,
depending on the purpose of the scientific problem that gives rise to it, the
objective of analytical research can be descriptive or explanatory. In practice,
these three objectives or functions are not mutually exclusive, and scientific
research can often combine more than one of these objectives. The essential
characteristics of these three objectives are explained below (Silva, 2022b).

2.5.1 Exploratory Research

- Exploratory research focuses on systems globally for the examination and

understanding of their structure and constitution and the identification of
their relevant characteristics and their interactions. For this reason, it is
called systemic research.

Exploratory research is suitable for the initial and final synthesis phases of a
scientific method cycle. In the initial synthesis phase, the systems are globally
focused to understand them as a complex whole in order to identify relevant
characteristics and problems that may be amenable to analytical research, to
obtain suggestions of conjectures of solutions (hypotheses) of these problems,
and to investigate the feasibility of research procedures and techniques. At this
phase, exploratory research is also useful for identifying and establishing
research priorities and alternative courses of action for fruitful research. In the
final synthesis phase, exploratory research is employed for practical verification
of responses to technologies generated by analytical research when integrated
into real systems (Silva, 2022b).

Exploratory research is particularly important in emerging or developing areas of

science, particularly in areas where fruitful theories have not yet been formulated,
important characteristics have not been identified, and research demand is
unknown. It can also be important in the initial phase of a large-scale research
program to indicate its feasibility and chance of success.

For focusing on systems globally, exploratory research has a precarious

theoretical basis and is poorly structured. Particularly, does not require the
formulation of scientific problems and hypotheses.

45
 Scientific Research: An Introduction
Scientific Research

2.5.2 Descriptive Research

- Descriptive research aims to characterize the units of a target population,

through their description and representation.

Descriptive research can also aim to identify and describe relationships between
characteristics. Its function is fundamentally to portray and report. It focuses on
the characteristics of subsystems of the systems that constitute the object of
research.

Descriptive research focuses on measuring characteristics and determining their

important properties, such as means and variances, or frequency distributions. It
can also aim at identifying association relationships, that is, covariation or
correlation of characteristics. In some research, this information is used for
prediction purposes. However, descriptive research may provide a basis for
prediction but is not appropriate for explaining the nature of characteristic
relationships.

Unlike exploratory research, descriptive research is characterized by the a priori

formulation of the research problem. Usually, the researcher already has
substantial knowledge to formulate the research problem, possibly as a result of
exploratory research. This knowledge allows him a clear definition of the
research objectives, particularly the relevant characteristics that must be
considered (Silva, 2022b).

2.5.3 Explanatory Research

Explanatory research focuses on the relationship between two subsets of

characteristics of units in a target population, one of which expresses the
performance or behavior of the units and the other supposedly affects that
performance or behavior. In this causal relationship, these two subsets of
characteristics are supposed to constitute the effect and the cause.

- Explanatory scientific research aims at inferences about the relationship

between a subset of the class of characteristics of the units of a target
population that express the performance of these units (response
characteristics) and a class of characteristics whose control and alteration
can, supposedly, imply the improvement of that performance (explanatory
characteristics), in the presence of the third class of characteristics of the
units (extraneous characteristics).

In other words, the function of explanatory research is to assess and express

significant and predictable causal relationships between a subset of
characteristics of the units of a target population with another subset of
characteristics. It must be remembered that in science these relationships can
only be expressed in terms of probable relationships, not in terms of exact
relationships (Silva, 2022b).

46
 Scientific Research: An Introduction
Scientific Research

2.6 STAGES OF ANALYTICAL SCIENTIFIC RESEARCH

Analytical scientific research starts from a particular problem or question

concerning the units of a target population. Once the initial question is
established, the systematic nature of the scientific method dictates the structuring
of the research in a precise and logically arranged form in several sequential
stages. From the initial question to its final answer, analytical scientific research
proceeds through eight main stages of operation, listed below (Silva, 2022b):

• First stage: Identification and formulation of the scientific problem or

research problem.
• Second stage: Formulation of the scientific hypothesis or research
hypothesis.
• Third stage: Literature review.
• Fourth stage: Construction of the research plan.
• Fifth stage: Data collection.
• Sixth stage: Analysis and interpretation of data.
• Seventh stage: Derivation of conclusions, which may lead to the
confirmation or rejection of the original hypothesis and confirmation or
questioning of the results of another research.
• Eighth stage: Presentation of results through a report, and dissemination
of these results.

Some relevant observations regarding this structure of analytical research

process deserve to be highlighted: a) It is arbitrary; research methodology texts
list different numbers of stages. However, an examination of such lists indicates
that all eight stages are involved. b) Stages do not necessarily occur in all
research in the order indicated. An experienced researcher may be studying the
literature at the same time as he is formulating his hypothesis and designing his
research plan. c) Stages are not always rigidly established at the beginning of the
research. Well-planned research can allow for adjustments or changes to be
made as it progresses. d) All stages are equally important in contributing to the
results of the research but do not involve the same amount of time, cost and
effort.

The eight stages are inherent to the analytical function of science and constitute
a common framework that ensures the basic attributes of the scientific method.
However, scientific research with a descriptive objective may comprise only part
of the eight stages.

In the following sections, the first two steps are considered, which define the
research objectives.

2.6.1 Identification and Formulation of the Scientific Problem

Analytical research begins with the identification and formulation of a scientific

problem, that is, a specific problem concerning a subsystem of the systems of

47
 Scientific Research: An Introduction
Scientific Research

the target population that can be solved by scientific research with available
resources (Silva, 2022b):

- A scientific problem or research problem is a problem or question

concerning characteristics of a target population and their relationships
that is answerable through the scientific method with the techniques,
procedures and resources available.

The identification and formulation of research problems can originate from

research in the initial phase of synthesis of a cycle of the scientific method, from
the literature and from the observation of the systems object of the research. It
can also be formulated as a theory-based prediction.

The first consideration for choosing a scientific problem must be its relative
importance; the second is the possibilities of its resolution considering the
available resources.

A research problem is formulated for descriptive or explanatory purposes. A

descriptive problem refers to the lack of, or demand for, knowledge about the
units of a target population. For example: “What technologies are used in the
vineyards of region A?”, “Which pathogenic microorganisms contaminate the
bovine milk produced in the dairy basin of region B?”, “What are the most
relevant problems of wheat crops in the region C?”. An explanatory problem is a
question concerning causal relationships between characteristics of the units of a
target population. For example, “What is the relationship between the incidence
of mastitis in dairy cows and the age of the animal?”, “What is the relationship
between the incidence of wheat blight and relative humidity?”, “What is the
relationship between productivity and season?”

The correct formulation of the scientific problem is crucial, not only because it
constitutes the definition of the research objective, but also because of its
intrinsic difficulty. The importance of this first stage of scientific research is
highlighted by the frequency of problems that are imperfectly or inadequately
formulated as a result of faulty information and reasoning. These failures can be
avoided by a careful, detailed and complete formulation of the problem, preceded
by adequate knowledge of its background, particularly its origin and importance,
and the destination of the research results that it will give rise to.

Although there are no general rules for formulating a scientific problem, certain
requirements must be satisfied for a problem to be considered a well-formulated
scientific problem. Some of these requirements are set out below (Silva, 2022b):

a) The problem must be inserted in an accessible body of scientific

knowledge,
b) Complex problems must be decomposed into a set of specific problems
that give rise to a research program,
c) All relevant characteristics related to the problem must be identified and
considered,

48
 Scientific Research: An Introduction
Scientific Research

d) The existence of the problem must be ensured by a clear definition of the

attributes of the problem and an explicit delimitation of its area of
existence,
e) The problem must be well conceived and delimited,
f) The problem must be understood before it is formulated,
g) It must be treatable by existing scientific research techniques and
available resources,
h) Must be answerable in objective terms,
i) Must be formulated clearly so that it has the same meaning for all
intelligent and well-informed people,
j) The formulation of the problem must delimit its scope, which demands a
clear characterization of the target population and its units,
k) Alternative or substitute formulations should be considered if the original
formulation is not feasible,
l) Among the alternative formulations of the problem, the one corresponding
to the most feasible and most efficient solution must be chosen,
m) The problem must be formulated in a systematic way,
n) The formulation of the problem should not influence the researched
phenomena.

2.6.2 Formulation of the Scientific Hypothesis

After the formulation of the problem, the next step is the inquiry regarding the
nature and connections of characteristics that lead to the idealization of one or
more paths for the solution or answer to the problem. Each of these idealized
paths constitutes a scientific hypothesis (Silva, 2022b):

- A scientific hypothesis or research hypothesis is a proposition of a solution

or answer to a research problem that allows an empirical confirmation test
through scientific research, with available techniques, procedures and
resources.

The research hypothesis can be suggested by the literature. Less often, it comes
from reasoning based on occasional observation of events. More commonly,
however, it is formulated as a prediction based on theory.

The function of the scientific hypothesis is to extend scientific knowledge beyond

the present frontiers of theoretical knowledge. So, the scientific hypothesis is
more than a connection between speculation and verification; it is an essential
factor in the growth of scientific knowledge.

It should be noted that exploratory research, in the initial and final phases of a
cycle of the scientific method, when the relevant characteristics of the target
population units and their relationships are not yet known, do not include problem
and hypothesis formulation. Its objective is precisely the identification of relevant
problems (scientific problems) for research and fertile conjectures for the solution
of these problems (scientific hypotheses).

49
 Scientific Research: An Introduction
Scientific Research

In line with the corresponding scientific problem, the function of a scientific

hypothesis can be either descriptive or explanatory. A descriptive hypothesis is
stated for a descriptive problem. It declares the existence of an empirical
uniformity, that is, of a fact or relationship that is commonly known but has not
yet been empirically verified, or which is not yet known but whose existence is
suspected. As science attempts to build a systematic body of verified knowledge,
it is natural that a great deal of its effort is spent on verifying hypotheses that
appear obviously true but have not yet been empirically verified. An explanatory
hypothesis refers to a causal relationship of characteristics. Seeks to find out if
changing one or more characteristics implies changing one or more other
characteristics (Silva, 2022b).

Some simple research concerning descriptive problems can be conducted

without reference to a hypothesis framework. However, most descriptive
research implies an underlying hypothesis.

For the same problem, one or more hypotheses can be formulated that differ in
the level of complexity of the problem and in the proposed solution. A simple
hypothesis may be a mere generalization of a particular empirical observation.
More complex hypotheses may postulate connections between events, or
elaborated chains of chance relationships. The formulation of an adequate
hypothesis depends on the knowledge and experience of the researcher.
However, creativity is of the utmost importance for the construction of useful and
fertile hypotheses.

There is no set of rules that can guarantee the formulation of the most
appropriate and fruitful hypothesis. However, the following attributes are
essential for a conjecture to be a scientific hypothesis (Silva, 2022b):

a) the hypothesis must have meaning and be formally correct,

b) it must be based on prior knowledge and be compatible with the body of
scientific knowledge,
c) must be logically related to the theory from which it is derived,
d) must respond to the problem that gave rise to it,
e) it must be conceptually clear and unambiguous,
f) it must be plausible, that is, it must be logically possible,
g) must be expressed in objective and operational terms,
h) should be as specific and simple as possible,
i) it must be empirically verifiable, that is, it must suggest research by which
it can be tested,
j) must allow a decision regarding the problem,
k) it must allow a reliable means of predicting unknown events,
l) it must include the criterion of its confirmation, that is, it must be stated in
terms that do not leave doubts regarding the answers considered as
confirmation.

Very often, more than one hypothesis can be formulated as a likely solution to a
research problem. In these circumstances, these requirements should be

50
 Scientific Research: An Introduction
Scientific Research

employed as criteria to suggest which hypothesis should be chosen as the most

appropriate and fruitful.

A scientific hypothesis is never definitively and absolutely proved. An empirically

confirmed scientific hypothesis is a scientific fact, that is not necessarily the exact
expression of reality. A hypothesis is only confirmed or not confirmed, depending
on the degree of confirmation established. A scientific fact, that is, a confirmed
hypothesis, is a new hypothesis, presumably with a higher level of reliability than
the original.

2.7 RESEARCH WORK AND THE ROLE OF THE RESEARCHER IN

SCIENCE

The scientist or researcher is the individual who rigorously employs the scientific
method in the pursuit of knowledge and who ultimately builds science. The
researcher must actively seek out and investigate nature to discover orderly
relationships. To do so, he must have some important qualities, among them:
curiosity, patience, objectivity, collaboration, communication, creativity,
discrimination, enthusiasm, accuracy, firmness, honesty, imagination, liberality,
morality, perseverance and concern for professional development (Silva, 2022b).

Originally, research work was an individual initiative. The advancement of

scientific knowledge has made the work of scientific research highly complex. As
a result, scientific research began to require a multidisciplinary team that must be
made up of specialists with cooperative action in the different phases of its
execution. Naturally, the more intense participation of each member is required in
specific phases.

It is highly relevant to the success of research, particularly in applied science,

that researchers are in contact with real problems and that they establish
appropriate strategies so that research is relevant and has its results applied in
practice. In fact, the fundamental objective of scientific research is to serve the
community in which it operates. For this purpose, the knowledge and experience
of the research team are needed, particularly (Silva, 2022b):

- In-depth knowledge of the research area, in particular the target

population units. In agriculture, for example, knowledge of production
systems, the economic and social conditions of agricultural companies,
market requirements, changes in product costs and value, appreciation of
product quality, etc.,
- knowledge of the results of research conducted previously in the region or
in other similar regions; in particular, knowledge of the existing literature,
and
- knowledge of the experience of the most progressive producers, acquired
over years of observation and experience.

51
 Scientific Research: An Introduction
Scientific Research

This knowledge is essential for a broader view of the problems that demand
research and their possible solutions and for the basis of research planning.
These and other requirements may prevent the execution of fruitless research.

Much research is ineffective and fails to achieve its objectives as a result of

limited resources and structural conditions for the research activity and
deficiencies in the guidance and discipline of this activity. On the other hand,
research institutions have accumulated many unanalyzed, uninterpreted and
undisclosed research results. Furthermore, it is not uncommon for experiments to
be performed repeatedly over a long period of time without the necessary
evaluation of their development.

The logical deduction, all too often recognized but rarely taken seriously, is the
little attention that has been devoted to planning and analyzing research findings.
In fact, it is generally observed that researchers devote very little time and effort
to research planning, which is the starting point, and the usefulness of the
research results depends on it. Most technological research is intended to
generate well-defined recommendations. It is therefore essential that such
recommendations are derived with a high level of security. These goals can only
be achieved by rational planning, which considers the research objectives, uses
existing resources more efficiently and adopts the most appropriate available
methodology to achieve these objectives (Silva, 2022b).

2.8 STATISTICS IN SCIENTIFIC RESEARCH

2.8.1 Scientific method and Statistics

The success of the physics of Isaac Newton (1643-1727), supported by applied

mathematics, developed and strengthened the deterministic view of science in
the 17th, 18th and 19th centuries. However, with the advent of quantum
mechanics by Max Karl Ernst Ludwig Planck (1858-1947) and the uncertainty
principle of Werner Karl Heisenberg (1901-1976), even physics, the leader of the
exact sciences, ceased to be deterministic.

Statistics is part of the modern scientific approach to uncertainty. This approach

assumes that "laws" can only predict "expectations" and those actual
observations can differ from these by "random errors". It is the study of these
errors that enables predictions under uncertainty. Prediction under uncertainty is
made possible by Applied Statistics, based on probability calculus. Applied
Statistics is vitally dependent on concepts from the areas to which it is applied.
For this reason, it is commonly identified with these application areas, assuming
specific designations, such as experimental statistics, biometrics, biostatistics,
econometrics and sociometry (Silva, 2022b).

Statistics is an integral part of the scientific method. The development of

Statistics has resulted from the demand from scientific and technological
progress. It is closely related to the development of the scientific method and the
advancement of different areas of science. The progress of science in the last

52
 Scientific Research: An Introduction
Scientific Research

two centuries has been the source for the extraordinary development of
Statistics.

It is notable that the development of modern Statistics was greatly influenced by

agricultural research. In the second decade of the 20th century, research at the
Experimental Station in Rothamsted, England, which began in 1843, generated a
considerable amount of data. The mathematician Ronald Aylmer Fisher (1890-
1962) was assigned to analyze this numerical information. During just 14 years at
Rothamsted, Fisher developed the theory and methods he needed and achieved
the practical application of his theory of statistical inference, which became the
foundation of modern Statistics. His work led to conclusions relevant to scientific
research, including that the amount of information generated by the inferences of
research cannot be greater than that contained in the data. Therefore, the
research planning and data generation process took on a fundamental
importance (Silva, 2022b).

From Fisher's contributions to agricultural scientific research, the new statistical

methods began to apply to other areas of science and technology. Scientific
developments in different areas, in turn, demanded new statistical methodologies
that also became, in general, applicable to other areas.

The development of the conceptual and methodological basis of scientific

research, of plans appropriate to the circumstances of each research, and of
more objective and efficient methods of data analysis, have increased the
possibilities, speed and reliability of scientific research. However, the
effectiveness of scientific research depends not only on specialized knowledge in
the research area but also, fundamentally, on knowledge of the scientific method,
which includes the statistical method.

This methodological development has resulted, in large part, from the vast
increase in computing capacity that became available, mainly from the mid-20th
century onwards. The growing computing facilities have allowed researchers
greater possibilities for decisions regarding the best use of available research
resources; in particular, to better plan the research and analyze its data more
adequately, achieving more efficient exploitation of the information provided by
the research (Silva, 2022b).

These developments have had a considerable impact on researchers' activities,

the effective exercise of which demands changes in interest and attitude. A
fundamental requirement is permanent training and updating, particularly for
understanding the methods and procedures implemented in computing
“packages”. In fact, the powerful facilities available for data analysis cannot be
used properly without understanding the methods implemented and, if used
incorrectly, can lead to misleading results.

A notorious fact is that, despite the consolidation of the scientific method,

particularly the statistical method, basic concepts are still not within the domain of
many researchers. Consequently, the impact of this methodological development

53
 Scientific Research: An Introduction
Scientific Research

is still not felt as widely as might be expected. This delay is not explained only by
the natural delay in incorporating theoretical results into practical activity. The
divorce that persists between theory and practice has several origins, among
them: failures in teaching and in the transmission of knowledge through literature,
lack of infrastructure and material resources in research institutions and lack of
vocation of researchers. From the first and the last origin, another no less
important cause arises the ritualistic attitude of many researchers who tend to
adhere to, copy and use ideas, concepts, methods and procedures mechanically,
without a critical attitude (Silva, 2022b).

2.8.2 Use of Statistics in Scientific Research

Statistical methods cannot prove a proposition or the result of scientific research.

What statistical methods provide is the assessment of the probability of error of a
proposition, or of the confidence that can be placed on the result of a scientific
research.

Suppose, for example, that scientific research is carried out to estimate the effect
of dietary supplementation on the growth of lambs. The research may indicate
that the supplement increases an animal's body weight at weaning by an average
of 7.5 kg. This result, however, does not provide any proof that similar growth, or
even any growth, would be observed if the research were repeated under similar
conditions. Considerable doubt will remain about this result. The use of a
statistical-based planned research followed by the appropriate statistical analysis
of the results will not eliminate the doubt of the research result, but it will allow to
assign a value to the reliability of this result and to estimate how often such a
result will arise if the effect does not exist, and to use it as a basis for
conclusions. For example, if it can be shown that differences as extreme or more
extreme than observed would occur less than once in a thousand if the
supplement had no effect, then there would be reasonable safety in disregarding
the possibility of the ineffectiveness of supplementation and concluding that it is
effective. However, the possibility will remain that it is ineffective, but now some
measure of the possibility will be available and can be used to demonstrate the
plausibility of the existence of a supplement effect (Silva, 2022b).

It is usually recognized that the use of statistical methods can greatly contribute
to the efficiency of scientific research and, particularly, to the validity of the
inferences it derives. However, the lack of knowledge of the fundamentals of
statistical methods, especially about the requirements for the validity of their
applications, often leads to their misuse. Statistics is often used without attention
to the assumptions required for the validity of the procedures employed. For
these and other reasons, it is necessary for the researcher to clearly understand
the techniques he employs.

2.8.3 Knowledge of Statistics by the Researcher

Researchers from different areas must have mastery of knowledge of scientific

research methodology that is essential for understanding and making decisions

54
 Scientific Research: An Introduction
Scientific Research

and actions related to the usual aspects of planning, conducting and statistical
analysis of research under their responsibility. This domain of knowledge should
cover statistical methods and computational resources, including “packages” for
statistical analysis (Silva, 2022b).

It is up to statistical specialists to complement the domain of the necessary

knowledge about methods that require specialized training and a more advanced
theoretical base in Statistics. Particularly, it is the responsibility of the statistician
to carry out studies and research aimed at adapting and developing new
methods and techniques.
___________________________________________________________________________________
© Copyright (2024): Author(s). The licensee is the publisher (B P International).

DISCLAIMER
This chapter is an extended version of the article published by the same author(s) in the following journal.
International Journal of Science and Research, 11(9), 2022. DOI: 10.21275/SR22914021617

55
 Chapter 3
Print ISBN: 978-81-976007-6-0, eBook ISBN: 978-81-976007-1-5

Scientific Research Methods

DOI: [Link]

ABSTRACT
The function of the research method is to organize research procedures to
minimize error, save effort, and supply an efficient collection of evidence relevant
to the research objectives. The literature presents a wide range of research
methods, with different procedures and names that vary according to the area of
science. In this chapter, the most common research methods in biology and
particularly in agriculture are considered in a unified and rational way. Due to
their characteristics, each of these methods is more suitable for one of the three
objectives of scientific research: exploratory, descriptive and explanatory.

Keywords: Exploratory; descriptive and explanatory research methods; case

study; prototype study mathematical modeling; census; sample
survey; experiment; observational study; strategic use of research
methods.

3.1 INTRODUCTION
The objective of scientific research is to verify a scientific hypothesis about a
scientific problem. According to the purpose of consideration of the problem, the
research objective can be exploratory, descriptive, or explanatory. In practice,
these three objectives, or functions, are not mutually exclusive, and scientific
research can often combine more than one of these objectives.

In general, the objective of exploratory research is to identify problems of the

units of a target population, suggest hypotheses and discover and investigate the
feasibility of themes, areas and techniques for more intensive research;
descriptive research aims to identify and describe the relevant characteristics of
those units, and important relationships between these characteristics;
explanatory research aims to derive inferences about cause and effect
relationships between characteristics of the units of the target population in order
to provide information for the control or prediction of a subset of characteristics
based on alteration or knowledge of another subset of characteristics (Silva,
2022c).

Whatever its objective, scientific research must be performed by one of several

alternative research methods. The function of the research method is the
organization of research procedures to minimize error, save efforts and provide
efficient collection of evidence relevant to the research objectives.
The literature presents a wide range of research methods, with differences in
procedures and designations that vary with the area of science. This chapter

56
 Scientific Research: An Introduction
Scientific Research Methods

provides a brief review of the important methods especially for biological and
agricultural sciences. Methods for the three objectives of scientific research are
considered: exploratory, descriptive and explanatory, and their relationships and
strategic use in scientific research (Silva, 2022c).

In the following sections, research methods are presented and illustrated with
respect to the purposes for which they are most specifically proper. Thus, the
next two sections present the methods of exploratory and descriptive research.
The following section considers the methods of explanatory research, with
consideration of two specific topics of explanatory research: relationships of
characteristics and control of the sample. The next section addresses
explanatory research methods and highlights the distinctions between these
methods that have relevant implications for inferences related to causal
relationships of characteristics. In the last section, the use of exploratory,
descriptive and explanatory research methods is discussed, to characterize that
they should be used strategically in the initial synthesis, analysis and final
synthesis phases of the application of the scientific method for the generation of
knowledge (Silva, 2022c).

This chapter is based on Silva (2020c) and the contributions listed in the
references.

3.2 EXPLORATORY RESEARCH METHODS

Exploratory research methods have greater flexibility and versatility and are less
structured than descriptive and explanatory methods. However, they do not
supply the basis for generalization provided by some descriptive research
methods, nor the basis for inferences related to causal relationships that
characterize the methods of explanatory research.

Exploratory research methods are proper for use in the initial phase of a cycle of
the scientific method, for the search for information that allows the identification
of scientific problems and fertile hypothesis, and in the final phase, for the
incorporation of the knowledge generated by analytical research into the existent
body of knowledge. These methods are particularly useful for identifying
characteristics, areas and important themes for research on systems of a target
population about which there is still not enough knowledge to formulate scientific
problems and, consequently, also scientific hypotheses (Silva, 2022c).

The most common exploratory research methods are case studies, prototype
studies and mathematical modeling.

3.2.1 Case Study

- The case study comprises the choice of a subset of a few units (cases) of
the target population with the characteristics relevant to the purpose of the
research, and the description of a large subset of characteristics of these
units, according to a pre-established plan.
The case study method is appropriate for the intensive description of a few units,
based on direct observations and interviews. Cases can be typical units that

57
 Scientific Research: An Introduction
Scientific Research Methods

characterize the diversity of the target population, or units of some particular

interest. Information regarding all characteristics relevant to the identification and
detailed description of these units is recorded (Silva, 2022c).

This research method originated in social sciences but is also used in other
areas of research, such as public health, industry and agriculture.

A case study is a proper research method to provide ideas, identify problems and
suggest hypotheses that can be verified by explanatory research. For this
reason, it is very useful in the initial synthesis phase of a cycle of the application
of the scientific method in research. A case study is also especially useful for
describing rare events and for the provision of counter example of some
universally accepted principles.

Case studies can follow descriptive research. In some studies, a sampling survey
is initially conducted to identify typical clusters of the units of the target
population. After this typification phase, a case study follows with one or a few
units of each of these clusters. For example, in research with the purpose of
deepening the knowledge of the production systems in use in a region, a survey
is carried out to identify the groups of typical systems in use; then, a system is
chosen from each of these groups and is followed by a time interval, when data
are recorded for a large number of characteristics of these systems (Silva,
2022c).

3.2.2 Prototype Study

- The study of prototypes consists of assembling one or a few units

(prototypes) with a set of chosen characteristics and carrying out the
observation and measurement of the relevant characteristics of these
units, according to a pre-established plan.

The prototype study emerged in industrial research. The purpose of a prototype

study is primarily problem detection. In the industry, prototypes are physical
models of machinery or equipment that result from the aggregate of components
developed and produced independently by various specialized suppliers. The
prototype study aims mainly at evaluating the performance of these models,
before large-scale manufacturing and commercialization (Silva, 2022c).

Unlike the case study, in which the units exist before the research is carried out,
in the prototype study the units are assembled especially for the research.

The study of prototypes is an especially useful exploratory research method to

provide the synthesis of explanatory analytical research results through their
integration into real systems and to provide the identification of problems and
hypotheses that can be verified by explanatory research. For example, a unit or
system of production of pigs is assembled, incorporating the technologies
generated by the research to observe the results of the interactions of these
technologies when integrated into the global system, and the performance of the

58
 Scientific Research: An Introduction
Scientific Research Methods

new system. In these circumstances, this unit is a laboratory for experience and
observation for the identification of problems and suggestion of hypotheses that
can be verified by explanatory research (Silva, 2022c).

3.2.3 Mathematical Modeling

The method of mathematical modeling also originated in industrial research, from

the advent of electronic computing. It is distinguished from other research
methods for being a conceptual or formal method, that is, the units or systems
are not real or empirical, but abstract or conceptual (Silva, 2022c).

- Mathematical modeling consists of representing a set of real units

(systems) by a conceptual model, that is, a mathematical model, or
simulation model, which expresses the relationships between the relevant
characteristics of the units. The model is then submitted to validation
through numerical experimentation, with the aid of an electronic computer,
and a comparison of the results with results obtained by other research
methods.

The mathematical model is assembled based on information provided by other

research methods - surveys, case studies, experiments, etc., and even by
indirect sources, including opinions.

The mathematical modeling method involves the identification of relationships

between the relevant characteristics of the units of interest, the mathematical
representation of these relationships, the construction of a global model to
represent (simulate) the functioning of these units and their changes in time, and
the use of this model to derive new knowledge about dynamic interactions
between characteristics of these units.

For example, a sheep research unit can assemble a model of lamb production
systems, or part of these systems, with the characteristics and relationships of
essential characteristics. A first model can be built based on the user's regular
practices, research results and opinions of producers, extensionists and
researchers. This model can be progressively improved with the incorporation of
new characteristics and relationship of characteristics, and reformulations of
characteristic relationships, based on new technologies that are generated by the
research. Its adequacy can be checked periodically, by comparing the results it
provides with the results of real systems of the target population (production units
of the region) (Silva, 2022c).

Because the method of mathematical modeling focuses on the system globally, it

is particularly useful for verification of the behavior of new technologies
generated by research when integrated into global systems, identification of
relevant interactions between characteristics of systems, and critical
characteristics that require research. It can be also especially useful for other
purposes related to scientific research, such as identification of research

59
 Scientific Research: An Introduction
Scientific Research Methods

problems, suggestion of hypotheses, indication of research priorities and

derivation of extrapolation of explanatory research results (Silva, 2022c).

3.3 DESCRIPTIVE RESEARCH METHODS

In contrast to the flexibility of exploratory research, descriptive research requires

clear definitions of the target population and research objectives, the
specification of the questions it aims to answer, and careful planning regarding
the sources of information to be used and the procedures for information
collection (Silva, 2022c).

- Descriptive research methods, also called surveys, are suitable for

research to identify, describe and represent the relevant characteristics of
the units of a target population and their correlations.

A survey can be conducted on all units of the target population or on a fraction of

that population, that is, on a sample of the target population:

- If the survey involves the observation of all units of the target population,
the research method is called a census. If it is conducted on a sample of
the target population, it is called a sample survey.

The census is only applicable in situations where the target population is finite,
and its units are identifiable and available for information collection. Even in
these circumstances, for economic, ethical or other reasons, sample survey may
be more convenient and appropriate. For this reason, it is much more frequently
used.

In descriptive research, observations and measurements are made of

characteristics of the units that manifest themselves without interference from the
researcher. The researcher is limited to the recording of information according to
a pre-established plan, very often through a questionnaire.

At the elementary level, descriptive research focuses only on the measurement

of characteristics and the determination of their important properties, such as
means and variances, or frequency distributions. At a more advanced level,
descriptive research can aim also at identifying and describing association
relationships between characteristics. For this reason, descriptive research
methods are sometimes called "correlational methods". However, descriptive
research is not suitable for deriving inferences about causal relationships, that is,
cause-and-effect relationships between characteristics (Silva, 2022c).

The sampling process determines the sampling design, that is, the structural
relationship between the sample and the target population and he chances of
selecting units from the target population. There is a wide variety of sampling
designs. The most important distinction between these designs refers to the
sample selection process, which can be objective and probabilistic, or subjective
and non-probabilistic:

60
 Scientific Research: An Introduction
Scientific Research Methods

- If the sampling process is objective and establishes a known probability

that each unit of the target population is included in the sample, it is called
probabilistic sampling or random sampling; otherwise, it is called non-
probabilistic sampling or non-random sampling.

A target population can be a real or conceptual population. Real populations are

finite, that is, they have a certain number of units that, in many situations, can be
identified and listed. Sampling processes for these populations can be objective
and random, or subjective and non-random. Conceptual populations are of
unknown size; they are often very large and, in these circumstances, for
convenience, they are usually considered infinite in statistical inference
procedures (Silva, 2022c). For this reason, sampling processes for conceptual
populations are necessarily subjective and not random. The most used sampling
designs are considered below.

3.3.1 Probabilistic Sampling

Probabilistic sampling has the following properties:

each unit of the target population has a known probability of being selected for
the sample, the sample is selected by some method consistent with these
probabilities, and objective inferences can be derived for the target population by
statistical procedures that consider these selection probabilities.

Probabilistic sampling design is also convenient because a) it allows the

determination of the sampling error, i.e., the degree to which the sampled
population differs from the target population; b) this information allows the
objective assessment of the validity of the sample. Non-probabilistic sampling
design does not allow this evaluation.

The most common probabilistic sampling designs are simple random sampling,
stratified random sampling, random cluster sampling, cluster-stratified random
sampling, random sampling in stages, systematic random sampling and multiple
random sampling (Silva, 2022c).

Simple Random Sampling:

The simple random sampling process is the purest form of random sampling: the
sample units are randomly chosen from the target population without any
restriction:

- A simple random sampling design consists of selecting the units of the

target population in an unrestricted manner and such that all units have an
equal probability of constituting the sample. In other words, a simple
random sampling design of size n consists of selecting the sample with
the property that all subsets of n units of the target population have the
same probability of selection.

61
 Scientific Research: An Introduction
Scientific Research Methods

The following surveys illustrate the use of simple random sampling: a) survey of
the incidence of diseases in animals of a herd of beef cattle on a farm in which
the sample consists of a group of animals randomly selected from the register of
animals of this herd; b) survey of the opinion of the members of a cooperative of
wheat producers to identify the most relevant problems for research whose
sample consists of farmers randomly chosen from the register of members of this
cooperative (Silva, 2022c).

Simple random sampling is particularly proper for relatively homogeneous and

not very large target populations. As it requires the identification and listing of all
units of the target population, it can become laborious or impractical for large
populations. It is also inconvenient when the target population comprises
subpopulations or groups of units among which characteristics relevant to the
purposes of the research vary markedly. In these circumstances, the random
choice of the sample without restriction may result in failure to represent variation
between these groups (Silva, 2022c).

Stratified Random Sampling:

A proper sampling process to constitute a representative sample of a target

population comprising considerably heterogeneous units may be to classify these
units into groups of homogeneous units and randomly and independently select a
subset of units from each of these groups:

A stratified random sampling design of size n consists of classifying the units of

the target population into k groups (strata) and then selecting a simple random
sample of ni size of the i-th stratum so that n1+n2+...+ nk = n.

The numbers of units selected from the strata of the target population can be
equal or different (Silva, 2022c):

- If the stratified random sampling process selects the same number of units
from each of the strata of the target population, the sampling process is
called equal allocation. If the numbers of units selected from the strata are
proportional to the sizes of the strata, then it is called proportional
allocation.

In a stratified random sampling design, the sampling error comes from the
variation between units within strata; it is not affected by the heterogeneity of the
strata. Thus, when the variation of characteristics relevant to the research aims is
considerably higher between the strata than between the units within the strata,
stratified random sampling leads to smaller sampling errors and more accurate
statistical inferences than simple random sampling. These advantages are more
salient when the sizes of the strata are sharply different.

62
 Scientific Research: An Introduction
Scientific Research Methods

Stratified random sampling may also be convenient when information regarding

individual strata is desirable and when collecting information separately for each
stratum is easier for physical or administrative reasons.

The following examples illustrate surveys with stratified random sampling: a)

survey of the incidence of diseases in animals of a herd of beef cattle on a farm
whose sample consists of a group of animals randomly selected from the register
of animals of each of the three breeds of this herd; b) survey to characterize the
quality of lamps received by a network of appliance stores from various
manufacturers in which are drawn to test some lamps from each of the
manufacturers; c) survey of the opinion of the members of a cooperative of
wheat producers to identify the most relevant problems for research whose
sample consists of randomly chosen farmers from the register of members of this
cooperative of each of three property sizes.

Stratified random sampling, as well as simple random sampling, requires the

identification and listing of all units of the target population. When the size of the
target population is too large, the preparation of this list can be too laborious or
impractical. Other drawbacks of the use of these sampling processes in these
circumstances are the high costs and administrative difficulties arising from the
collection of information from dispersed units.

Random Cluster Sampling:

When the size of the target population is too large, random sampling can become
easier and more convenient for preparation, cost, and administration when units
constitute relatively homogeneous natural groups. In these circumstances, a
proper sampling process to constitute a valid and representative sample can
comprise the random selection of a subset of these groups, instead of a subset
of individual units (Silva, 2022c):

- A random cluster sampling design consists of classifying the units of the

target population into groups (clusters) and then extracting a simple
random sample of the clusters.

The sample can consist of all units of the selected clusters or a simple random
sample of each of these clusters. In the latter case, equal allocation or
proportional allocation to cluster sizes can be adopted. Proportional allocation is
most often used.

In many situations, clusters comprise geographically close units and are areas
corresponding to divisions of a region that constitute the target population. In
these circumstances, the design is usually called random sampling by area.

Examples of characteristics that may be appropriate for cluster formation are a)

farm in a survey of the incidence of diseases in animals in a region in which the
unit is an animal; b) county in a survey of technologies adopted by farmers in a
region in which agricultural property is the unit; c) family in a survey of the

63
 Scientific Research: An Introduction
Scientific Research Methods

socioeconomic situation of the inhabitants of a neighborhood of a city in which

unity is the individual. A random cluster sampling illustration is provided by a
survey of consumption in a city in which the sample unit is the family defined as
the set of people who inhabit the same residence, the blocks are identified on a
city map and a certain number of blocks is chosen at random for collecting
observations. Next, a subset of the residences of each of the chosen blocks is
selected by simple random sampling with proportional allocation (to the number
of residences of these blocks).

This random sampling process requires only the identification of clusters and the
listing of the units of the clusters selected for the sample. For this reason, it is
convenient for considerably large target populations, especially when a complete
list of units is not available and obtaining it would entail a lot of work and high
cost.

When the variation of relevant characteristics is lower among clusters than

between units within clusters, random cluster sampling produces less sampling
error and more precise statistical inferences than simple random sampling. Thus,
in the planning of a random cluster sampling survey, greater homogeneity
between clusters should be sought than between units within clusters. This is the
opposite of what should be sought with the stratified random sampling process,
where greater homogeneity within the strata than between the strata is
convenient (Silva, 2022c).

Stratified and Cluster Random Sampling:

In some situations, a stratified and clustered random sampling design may be

convenient (Silva, 2022c):

- In a stratified, cluster random sampling design the target population is

divided into groups (strata); each of these groups is subdivided into
subgroups (clusters); then, from each of the strata, a simple random
sample of clusters is extracted, and then, from each cluster (from each
stratum) a subset of units is randomly selected by some appropriate
criterion.

The two surveys that follow exemplify stratified and cluster random sampling
design: a) Survey of consumption in a city in which the sample unit is the family,
i.e., all people living in the same residence, the city is divided into zones (strata)
and the blocks (clusters) of each of these zones are identified; then, a simple
random sample of the blocks of each of the zones is extracted, and a subset of
families from each selected block of each city zone is randomly chosen. b)
Survey of opinions of wheat growers of a wheat region in which this region is
divided into counties (strata) and each county into areas (clusters); then, a simple
random sample of areas within each county is selected with the number of areas
selected proportional to the number of wheat grower areas in the county, and
from each area a sample of wheat growers is randomly chosen.

64
 Scientific Research: An Introduction
Scientific Research Methods

Random Sampling in Stages:

Stratified and cluster random sampling design is a two-stage sampling design;

the first stage consists of a simple random sampling of clusters within each
stratum; the second stage, in the random sampling of the units of each cluster
selected in the first stage. This idea of stage sampling can be conveniently
extended to very large and complex population situations.

For example, a national survey on employment and unemployment, with the

survey in homes. It would not be practical to obtain and maintain a list of the
country's residences. In addition, data collection through interviews in very
dispersed households would have an exorbitant cost. For these circumstances, it
is interesting to design a sampling in stages, such as the design in the three
stages that follow: 1 - extraction of a random sample of the counties of each
State of the country; 2 - selection of a random sample of the blocks of each of the
chosen counties; and 3 - selection of a random sample of the residences of each
of the blocks (Silva, 2022c).

Systematic Random Sampling:

Systematic random sampling is a probabilistic sampling process used with some

frequency when it is possible to list or sort all units of the target population (Silva,
2022c).

- A systematic random sampling design of a sample of size n from a

population of size N consists of randomly choosing the c-th unit of the
subset of the first k = [N/n] units, and then taking each of the k-th units
from it, so that the sample results from the units of the target population of
orders c, c+k, c+2k,... ([r] denotes the largest integer that does not
exceed the rational number r).

In this design, the random choice of the first unit of the sample determines the
whole sample. Suppose, for example, the extraction of a sample of 10% from the
list of the 620 associates of a cooperative of milk producers. To do this, a number
is randomly selected from the set of the first 10 integers, 6 for example, and then
the producers identified in the list with the numbers 6, 16, 26 and so on, up to
616 are taken.

Systematic random sampling considerably reduces the number of subsets of

units of the target population that can be chosen to constitute the sample but
ensures that all units of the population have a chance to constitute the sample.

This design has three advantages over the simple random sampling design: it
considerably reduces the number of subsets of units of the target population that
can be chosen to constitute the sample, it ensures that all units in the population
have a chance of constituting the sample, it is easier to plan, since only a random
number is required, and distributes the sample more evenly over the target
population. For this last reason, the systematic random sampling design

65
 Scientific Research: An Introduction
Scientific Research Methods

sometimes leads to smaller sampling error and more precise inferences than the
simple random sampling design. However, it has disadvantages. An important
disadvantage is that if there is any hidden ordering of the units of the target
population according to the levels of some relevant characteristic, a severely
biased sample may result. Thus, for example, a) a systematic sample of the
houses of a city may have too large, or too small, numbers of corner houses; b) a
systematic sampling of the plants in an orchard may have the plants selected
from the same positions along each of the rows of the orchard (Silva, 2022c).

Multiple Random Sampling:

In some circumstances it may be convenient to select the units for the sample in
stages (Silva, 2022c):

- In a multiple-sampling design, the sample consists of units that are

selected from the target population in successive stages. Depending on
the results observed at each step, subsequent steps may be waived.

This sampling process is often employed in sampling inspection for product

testing or quality control. An example is provided by the following double
sampling process for a decision regarding acceptance or rejection of a batch of
an industrial product: A random sample of the units of the lot is chosen; if the
number of defective units in this sample is at most 5%, the lot is accepted; if it is
at least 10%, the lot is rejected. If the number of defective units is between 5 and
10%, a second sample is required. So, if the number of defective units in the
combined sample of these two samples is at most 10%, the lot is accepted; if the
number of defective units is greater than 10%, the lot is rejected.

Multiple random sampling starts with a relatively small sample and proceeds with
the successive increment of the sample size in stages until the decision criterion
is achieved. Thus, it is a convenient sampling process because it allows the
reduction of the sample size and the reduction of cost and time.

An extreme case of multiple sampling is sequential sampling, in which the

sample is increased by unit per unit until a conclusion is reached about accepting
or rejecting a given hypothesis referring to the target population. Sequential
sampling allows to minimize the average sample size in the long term.

3.3.2 Non-probabilistic Sampling

In some circumstances, it may be impractical or inconvenient to consider all units

of a real population in the sample selection process. For example, a) in a survey
of a state's voter preference over candidates for governor in the next election, it
is impractical to sample a subset of individuals selected from the list of eligible
voters; b) in a survey of the opinion of the citizens of a county on a particular
topic, it may be impractical or inconvenient to consider all the inhabitants of that
county in the sample selection process. Similar circumstances occur when the
units of a real population cannot be all identified or are not all accessible, such as

66
 Scientific Research: An Introduction
Scientific Research Methods

in a survey of the fish population of a lake, in a survey for botanical

characterization of plant species native in a region, and in a survey of animals
from a biological reserve. In such circumstances, the use of a probabilistic
sampling design becomes impractical or unfeasible. The use of probabilistic
sampling design is also unfeasible when the target population is a conceptual
population, whose units are not all available at the time of research execution. In
this case, the target population can be specified only conceptually, that is, by
describing the properties of the units that constitute it (Silva, 2022c).

In all these situations, the choice of sample should be based on human criteria
and judgment. In these circumstances, the probability of each unit of the target
population being chosen to constitute the sample is not known. In fact, in many
situations, a considerable fraction of these units has no chance of being selected
for the sample.

As these sampling processes have no probabilistic basis, they do not allow the
determination of the sampling error, and they do not allow the derivation of
inferences to the target population by statistical procedures. Such inferences
should be based on subjective judgment. Consequently, they are generally
subject to bias, which cannot be objectively assessed.

The most common non-probabilistic sampling designs are judgement sampling,

convenience sampling, quota sampling, systematic non-random sampling,
haphazardly sampling at random and mechanical sampling (Silva, 2022c).

Judgement Sampling:

- In a judgement sampling design, the sample is constituted based on the

judgment of the researcher who seeks to select a subset of units that is
representative of the target population.

Very often, the sample is chosen by a researcher who, because he is considered

to have a deep knowledge of the relevant characteristics of the target population
units, is judged qualified to select a sample that represents these characteristics
and the variability present in the population.

The following research illustrates a judgement sampling design: a) Survey of the

opinions of the workers of an industry to increase productivity in which only a part
of the workers should be interviewed; the choice of these workers is made by the
production manager because he is considered the most qualified to obtain an
appropriate representation of all workers. b) Survey of characteristics of bean
production systems in a region in which a subset of counties whose producers
should be included in the sample should be chosen; the choice of counties is
made by a group of researchers (Silva, 2022c).

Convenience Sampling:

- In a convenience sampling design, the sample consists of units selected

from the target population by some convenience criterion, very often
related to the accessibility of the units.

67
 Scientific Research: An Introduction
Scientific Research Methods

Convenience sampling is often used to obtain approximate information about the

target population in a short time and at low cost. It is exemplified by the following
surveys: a) survey of information about families when the interviewer chooses
families close to their residence or workplace; b) survey of opinions of the
residents of a district in which the interviewer decides to interview the 200 people
who are present in a certain event next weekend; c) survey of characteristics
related to the quality of cement from an industry in which the sample consists of
bags chosen from the top of one or more trucks with a load of cement from that
origin (Silva, 2022c).

Quota Sampling:

This non-probabilistic sampling process is like the stratified random sampling

design. It differs from this design in that the units are selected from each of the
strata by a non-random process. The researcher first identifies the strata and
proportions in which they are present in the target population and then uses non-
random sampling to select the required number of units of each of the strata
(Silva, 2022c):

- A quota sampling design consists of: 1) division of the target population

into subpopulations or strata representing the various categories in which
their units are classified according to some relevant characteristic; 2)
determination of the sample fraction size (quota) corresponding to each of
these strata; and 3) delegation to the executor of the survey in each
stratum the selection of the units until completing the quota as indicated to
him.

The choice of units of each stratum is usually carried out by judgement or

convenience sampling. Frequently, unaffordable or difficult-to-access units are
overlooked.

Illustrations of sampling survey by quota: a) survey of information from rural

producers of a region through interviews in which the region is divided into
sections and each section is assigned to an interviewer, and the interviewer is
responsible for choosing a certain number of producers from his section for the
interviews; b) survey of the socioeconomic conditions of the families of a district
in which the district is divided into segments of streets and the interviewers are
assigned the choices of the residences of the corresponding segments so that
the predetermined numbers of residences are completed (Silva, 2022c).

Systematic Non-random Sampling:

This sampling process is usable in situations where the units of the target
population will become available one by one over time:

- A systematic non-random sampling design, also called interval sampling

design, consists of taking one unit at the end of each repetition of a given
time interval, as the units of the target population become available, or

68
 Scientific Research: An Introduction
Scientific Research Methods

taking each unit that arises immediately after a subset of a certain number
of units.

This sampling process requires that the starting point, that is, the initial instant or
the first unit be selected by some objective criterion. It is sometimes used in the
industry to monitor the operation of equipment and production processes and for
product quality control, and in surveys involving interviews of people. Examples
of this sampling method are a) survey to verify that the set of equipment or the
production process of an industry is working according to specifications in which
a production line item is selected for testing every time interval (e.g. every 60
minutes), from a pre-established time; b) survey of the opinion of the customers
of a store in which is selected every tenth person who enters the store, from the
opening of the store; c) survey to collect suggestions from the residents of a
street regarding a remodeling project, in which the sample consists of the
occupants of each tenth house of this street, from a given corner (Silva, 2022c).

Haphazardly Sampling:

Haphazardly sampling is a process of non-random samples usable when the

units of the target population are arranged in some container, compartment or
facility, and are not identified or is difficult to access selected units (Silva, 2022c):

- In a haphazard sampling design, the sample consists of units selected

from various points of the target population chosen by some procedure
that simulates chance.

This non-random sampling seeks to constitute the sample by units chosen by

process, supposedly corresponding to the random choice, without, however, a
probabilistic basis. This can be proceeded in several ways. For example, if the
units are arranged in some container, taking the units at hand, in various
positions; if the units are in some large compartment or installation, traveling
through the compartment or installation and taking the units within reach, at
various points.

The following example illustrates situations of use of haphazard sampling: a)

survey of the characteristics of the screws contained in a box made by means of
a sample of some screws caught in each of ten points of the box; b) survey of
characteristics of chicks arranged in several cages each with 100 chicks in which
from each of these cages the researcher picks five chicks within reach of his
hand; c) survey to characterize the quality of meat stored in a cold room whose
sample consists of parts of meat collected in various positions along a walk
through the warehouse.

The haphazard sampling can be a reliable sampling process if the target

population is homogeneous or there is no tendency of systematic grouping of the
units, and there is no possibility of the researcher being influenced by some
characteristic of the units (Silva, 2022c).

69
 Scientific Research: An Introduction
Scientific Research Methods

Mechanical Sampling:

This non-random sampling process is usable when the target population

comprises a solid, liquid, or gaseous volume:

- A mechanical sampling design consists of constituting the sample of a

target population comprising a solid, liquid or gaseous volume by parts
taken at various points of this volume chosen in a supposedly random
manner.

Mechanical sampling is usually conducted with the use of special instruments

such as claws, shells and probes. This sampling process is similar to haphazard
sampling and should be used with the same care as indicated for that sampling
process (Silva, 2022c).

3.4 EXPLANATORY RESEARCH METHODS

3.4.1 Characteristic Relationships

Descriptive research can provide the identification of association relationships

between characteristics. Although descriptive information is often useful for
predicting relevant characteristics, it is usually desirable to identify the origin of
variation in these characteristics, that is, whether the variation of predicted
characteristics results from the variation of other characteristics and the
identification of these characteristics. It is also desirable to know the relationships
of the characteristics that are the object of prediction (effects) with the
characteristics that produce their variation (causes). This knowledge is important
because knowledge of causes increases the ability to predict and control effects.
This is the purpose of explanatory research (Silva, 2022c).

Explanatory scientific research aims to derive inferences about causal

relationships between two subsets of characteristics of the units of a target
population where, supposedly, one of these subsets is the cause and the other is
the effect. The first subset of characteristics constitutes the class of explanatory
characteristics, and the second subset is a subset of the class of response
characteristics, respectively:

- Response characteristics express the performance or behavior of the

units, and explanatory characteristics are supposed to affect that
performance or behavior.

Explanatory research aims to provide information for the control or prediction of

response characteristics based on the alteration or knowledge of explanatory
characteristics. The difficulty of this process is the presence of natural variability
of the third class of characteristics of the units of the target population that also
affect the response characteristics, which are called extraneous characteristics.

70
 Scientific Research: An Introduction
Scientific Research Methods

The response and explanatory characteristics constitute the objective of the

research and are designated based on substantive scientific theories and
empirical knowledge. The identification of these characteristics requires
knowledge and vision of the research area. The class of the extraneous
characteristics is defined by exclusion, that is, the characteristics of the units
excluded the class of response characteristics and the class of explanatory
characteristics. The distinction and classification of the characteristics of the units
of the target population in these three classes is a crucial step in planning
explanatory research (Silva, 2022c).

In general, only a subset of the response characteristics is of interest in a

particular research. The characteristics of this subset and the explanatory
characteristics must be identified and defined explicitly. The extraneous
characteristics constitute an extremely numerous set and it is normally necessary
to identify their aggregates and individualize the most relevant characteristics.

The units of the sample also include the three subsets of the characteristics of
the units of the target population, that is, responses, explanatory and extraneous
characteristics. Conceptually, these characteristics of the sample are the same
as those considered in the target population. The levels of the response
characteristics in the sample are the same as in the target population. The levels
of the explanatory characteristics in the sample are the levels of the target
population themselves or subsets of these appropriately chosen. Ideally, the
levels of the extraneous characteristics of the sample are a representative
sample of the corresponding levels in the target population (Silva, 2022c).

These concepts are illustrated by the examples that follow:

a) Research on the control of the incidence of Gibberella in wheat crops in a

tricultural region.

The set of response characteristics comprises the characteristics that express

the performance of the crops, that is, the characteristics of the grain and the
plant. However, in this research, only the response characteristics relevant to its
objective are of interest, that is, control of the incidence of Gibberella, a fungus
that causes damage to wheat production in this region. These characteristics are
the weight of grain production and the degree of infection of Gibberella (Silva,
2022c).

To achieve the objective of this research, it is decided to consider the effect of

fungicides, and three fungicides are chosen: Mancozeb, Cyproconazole and
Propiconazole, and the absence of fungicide to verify the need for fungicide
application. As the effect of fungicide may depend on the cultivar, it resolved to
consider three cultivars: BR 18, BR 23 and BRS 177, respectively of early,
medium and long cycles. Thus, two explanatory characteristics are defined:
fungicide and cultivar; the first with four levels: fungicides Mancozeb,
Cyproconazole and Propiconazole and absence of fungicide, and the second
with three levels: cultivars BR 18, BR 23 and BRS 177. As the effects of

71
 Scientific Research: An Introduction
Scientific Research Methods

fungicides may depend on environmental conditions, which vary in space and

time, it is decided to conduct the research in four sites in the tricultural region in
the next three years; thus, location and year are also considered as explanatory
characteristics (Silva, 2022c).

The levels in the sample of the characteristics of fungicide and cultivar are the
same as in the target population. However, the levels of the local and year
explanatory characteristics in the sample are, respectively, the locations and
years of conducting the research; therefore, subsets of the respective levels
defined for the target population, are appropriately chosen to represent the
corresponding populations of levels.

In each location of each year, the research is conducted on a plot of land divided
into 48 plots. The 48 combinations of the four levels of fungicide and the three
cultivars are assigned to these plots so that each combination results in four
plots. Then, the extraneous characteristics are the characteristics of the seed
(genetic, except those related to the explanatory characteristic cultivar, and
phenotypic: size, vigor, purity, sanity, etc.), environment (soil, climate, incidences
of pests, diseases, weeds and predators, etc., except the characteristics inherent
to location and year, defined as explanatory characteristics), cultivation
techniques (soil preparation, planting, application of pesticides, except for
fungicide for control of the Gibberela, which are explanatory characteristics, etc.),
and measurement processes (Silva, 2022c).

b) Research on the control of intestinal worms in lamb production units.

The important responses characteristic for the objectives of the research are
body weight and number of parasites in the animal's viscera at slaughter (154
days of age), and body weight and number of eggs in specific moments of the
research period (70-154 days).

The researcher decides to consider the anthelmintic Ranizole and establishes

the objective of determining the optimal dose of this anthelmintic in the range of 0
to 40 mg/kg live weight for male and female animals. So, for the response
characteristics measured at slaughter, there are two explanatory characteristics:
anthelmintic and sex. For the response characteristics measured at several
moments of the research period, there is an additional explanatory characteristic:
the age of the animal. The levels of the explanatory characteristic anthelminthic
in the target population are the doses in the interval [0; 40 mg/kg of live weight];
the levels of sex are the two sexes – male and female and the levels of age are
the ages in the interval [70; 154 days] (Silva, 2022c).

The levels of anthelmintic in the sample are the two extremes and the middle
levels of the interval of doses in the target population, i.e. 0, 20 and 40 mg/kg
mg/kg. The levels of sex in the sample are the same levels in the target
population themselves, i.e., male and female, and the levels of age are 70, 84,
98, 112, 126, 140 and 154 days.

72
 Scientific Research: An Introduction
Scientific Research Methods

The research is carried out in a facility with 18 pairs of pens, each pen with
capacity for one animal, and each pair of pens with a common feeder and
drinker, and 36 Ideal breed animals, 18 males and 18 females, with an
approximate age of 70 days, with research period ending with the slaughter of
animals at five months of age. Then, the extraneous characteristics of the sample
are the characteristics related to the animal (genetic, age, age, weight, sanity
etc., except sex, which is explanatory characteristic), the environment
(installation, pasture, climate, incidences of diseases, parasites and predator),
management (lamb preparation, application of antibiotics, vaccines and
parasiticides, except for the application of anthelmintic, which is explanatory
characteristic) and measurement processes (Silva, 2022c).

c) Research on the effect of incidence of winding virus on the quantity and

quality of grape production of vines of the Italy cultivar, in a wine-growing
region.

The important response characteristics are bunch weight, number of bunches,

average bunch weight, and sugar and acidity grape content.

The objective of this research define the incidence of winding virus as an

explanatory characteristic. It is decided to consider three degrees of infection: no
infection, medium infection and strong infection; so, these are the three levels of
this explanatory characteristic in the target population. Location and year are also
considered explanatory characteristics since the quantity and quality of grape
production may depend on environmental conditions.

The research is conducted in a vine orchard with eight plants chosen with each
of the three degrees of infection. Thus, the levels of winding virus infection in the
sample are the same as in the target population. The locations and years of the
sample are chosen so that the proper representation of the environmental
variation of the target population is achieved. The extraneous characteristics of
the sample are the characteristics of the plant (genetics, vigor, sanity, except the
incidence of winding virus, which is an explanatory characteristic), the
environment (soil, climate, incidences of pests, diseases, predators, etc., except
the properties inherent to location and year, which are explanatory
characteristics), cultivation techniques (fertilization, application of insecticides,
herbicides, fungicides, etc.), and measurement processes (Silva, 2022c).

The objective of the research defines five explanatory characteristics: cultivar,

fertilization, insecticide, herbicide and fungicide. The first of these characteristics
has an unspecified number of levels; the researcher decides to consider two
levels for each of the other characteristics – adopts and does not adopt. Location
and year are also considered as explanatory characteristics since vine
productivity depends on the environment.

The levels of explanatory characteristics fertilization, insecticide, herbicide and

fungicide in the sample are the same as the target population, that is, it adopts
and does not adopt; the levels in the sample of explanatory characteristics

73
 Scientific Research: An Introduction
Scientific Research Methods

cultivar, year and place are subsets of the levels of the target population, that is,
the cultivars, sites and years, respectively, that are present in the units chosen to
constitute the sample (Silva, 2022c).

The extraneous characteristics of the sample are the characteristics related to

seed (except cultivar, which is an explanatory characteristic), environment
(except the properties inherent to site and year), cultivation techniques (except
fertilization, insecticide, herbicide and fungicide, which are explanatory
characteristics) and measurement processes.

3.4.2 Sample Control

The variation of the response characteristics in the units of the sample is due
both to the effects of the explanatory characteristics and the effects of the
extraneous characteristics. Consequently, there is a confounding of the effects of
explanatory characteristics with the effects of extraneous characteristics (Silva,
2022c):

- The effects of the explanatory and extraneous characteristics on a

response characteristic are confounded when they cannot be
distinguished from each other.

Confounding hinders and may make it impossible to derive unbiased inferences

related to causal relationships between response characteristics and explanatory
characteristics:

- The effect of extraneous characteristics on a response characteristic is

called research error.

The term "research error" comes from the fact that inferences about the casual
effects of explanatory characteristics on response characteristics are subject to
errors arising from the effects of extraneous characteristics.

The basis of scientific research for inference about a causal relationship between
a response characteristic and an explanatory characteristic is the following
principle of causality: if the level of an explanatory characteristic is altered and
the extraneous characteristics remain constant, any variation manifested by the
response characteristic is due to the change in this explanatory characteristic.

This ideal is approximately achieved in certain laboratory experiments in physics

and chemistry, where essentially identical units can be built so that the research
error is reduced to an irrelevant proportion. However, in research in natural
sciences, particularly in biology, units can be considerably heterogeneous,
mainly due to variations in environmental characteristics. In these circumstances,
the control of this heterogeneity is generally necessary. On the other hand, the
exaggerated reduction of research error may imply conducting the research
under very special conditions, not representative of populations of natural units,
usually with considerable variability (Silva, 2022c).

74
 Scientific Research: An Introduction
Scientific Research Methods

The solution to this essential problem of explanatory research is the control of the
sample to avoid or decrease the confounding effects of explanatory and
extraneous characteristics, to the point where it is possible and appropriate, and
to make the remaining confounding unbiased.

The following procedures can be considered for sample control with a view to
reducing the research error and its bias: control of explanatory characteristics,
control of research techniques, pairing or local control, statistical control, and
randomization (Silva, 2022c).

[Link] Control of explanatory characteristics

This control is exercised by choosing the levels of one or more explanatory

characteristics for the sample and assigning those levels to the sample units
according to a plan.

Due to the relevance of the implications for the validity of inferences, two classes
of explanatory characteristics must be distinguished according to the control that
the characteristic is subjected in the sample: extrinsic characteristic (treatment
characteristic) and intrinsic characteristic (Silva, 2022c):

- An explanatory characteristic whose levels chosen for the sample are

randomly assigned to the units of the sample and whose effects are not
biasedly confounded with effects of extraneous characteristics is an
extrinsic characteristic or treatment characteristic. An explanatory
characteristic that is inherent to the units of the sample or whose effects
are biasedly confounded with the effects of extraneous characteristics is
an intrinsic characteristic.

The levels of a treatment characteristic are called treatments; more generally, the
levels of an explanatory characteristic (extrinsic or intrinsic) are called conditions.
The distinction between extrinsic characteristics and intrinsic characteristics is
illustrated by the following examples (Silva, 2022c):

a) In the research on control of the incidence of Gibberela in wheat crops,

the levels of fungicide and the levels of cultivar are randomly assigned to
the units of the sample. If control of extraneous characteristics ensures no
biased confounding of their effects with effects of fungicide and cultivar,
these two explanatory characteristics are treatment characteristics.
However, the sites and years are inherent to the units of the sample;
therefore, location and year are intrinsic characteristics. The levels of
fungicide and cultivar in the sample are treatments; the levels of fungicide,
cultivar, sites and years in the sample are conditions.
b) In the research on control of intestinal worms in sheep, doses 0, 20 and
40 mg/kg of the anthelmintic are randomly assigned to the pairs of boxes
and the sexes male and female are assigned at random to the two boxes
of each of those pairs. Thus, if the control of extraneous characteristics
guarantees the absence of biased confounding of effects of these

75
 Scientific Research: An Introduction
Scientific Research Methods

extraneous characteristics with anthelmintic effects, anthelmintic is a

treatment characteristic and its doses are treatments. However, sex is an
inherent characteristic of the animal; therefore, the effect of sex is
confounded with extraneous characteristics of the animal. If the
differences in the extraneous characteristics between the two sexes are
irrelevant, sex can be considered a treatment characteristic. This is the
case if the animals of the two sexes come from the same herd and proper
control of extraneous characteristics is exerted. Otherwise, sex should be
considered an intrinsic characteristic.
c) In the research on the effect of the incidence of winding virus on the
quantity and quality of grape production of vines, the degree of infection of
the winding virus is inherent to the plant and is manifested outside the
control of the researcher; therefore, winding virus infection is an intrinsic
explanatory characteristic. Age is also an intrinsic explanatory
characteristic because its effect is confounded with effects of the
extraneous characteristics (of the environment and cultivation techniques)
that vary between moments of mensuration. Thus, this research does not
comprise treatment characteristics.

[Link] Control of research techniques

- The control of research techniques comprises the execution of

intervention actions on the sample to make the manifestation of
extraneous characteristics constant or irrelevant.

The control of research techniques is exercised from the initial choice of the
sample to the measurement process. In the initial constitution of the sample, it
consists of the selection or construction of homogeneous units regarding
extraneous characteristics. During the conduct of the research, it is carried out by
the controlled execution of research techniques with the purpose of making
manifestations of extraneous characteristics constant or uniform. Control of
research techniques is also used in measurement processes to avoid the
introduction of extraneous characteristics that may result in loss of precision or
bias due to poorly calibrated instruments, and measurement errors from other
sources. The measurement of some response characteristics is performed by
judges. This is the case, for example, of measurements by organoleptic
evaluation by tasters. These assessments may require subjective judgment and
experienced judges with specialized training (Silva, 2022c).

The use of research technique control is illustrated by the examples below.

a) In each of the four locations and three years of running the research on
control of the incidence of wheat Gibberella with fungicides, the control of
research techniques should include use of homogeneous seeds regarding
germination, vigor, purity, health, etc., and the adoption of uniform
cultivation techniques (soil preparation, planting, control of diseases,
pests, invaders and predators, harvesting) and measurement procedures.

76
 Scientific Research: An Introduction
Scientific Research Methods

b) In each of the three sites, the research on the control of intestinal sheep
worms with anthelmintics should be conducted with homogeneous
animals regarding genetic characteristics, sanity, age, and weight, with
uniform management techniques (food and water supply, applications of
antibiotics, vaccines and parasiticides) slaughter and measurement
processes.
c) The research on the incidence of vine winding virus should use plants
chosen from each of the defined three degrees of incidence of the virus
that are uniform in terms of vigor and sanity, except for the incidence of
the winding virus, and adopt uniform cultivation techniques and
measurement procedures.

The extraneous characteristics submitted to the control of research techniques

are manifested in the sample constantly or, more often, in the level of uniformity
achieved by the control exercised. Thus, the confounding of the effects of the
controlled fractions of these characteristics that are confounded with effects of
explanatory characteristics is eliminated or reduced (Silva, 2022c).

The use of this sample control technique should be used with care in
technological research, where it is required that the sample appropriately
represents the target population. It should be used without compromising the
representativeness of the sample.

[Link] Local control or pairing

- Local control or pairing (matching) consists of classifying the sample units

into groups of units more homogeneous in relation to a relevant
extraneous characteristic than the complete set of sample units so that the
variation in response between groups is eliminated from sampling error
and from differences in responses between levels of explanatory
characteristics.

Illustrations of the use of local control or pairing are provided by researches

described in previous examples: a) For the control of soil characteristics (fertility,
texture, structure, moisture), which are usually more similar in nearby than
distant plots, in each of the four sites and the three years, the 48 plots are
classification into four blocks of 12 nearby plots. Then, a complete collection of
the 12 combinations of the four fungicide levels and the three cultivars is
assigned to the 12 plots of each block. Thus, the variation between blocks can be
eliminated from the differences between fungicides and cultivars, and from the
research error. b) The 18 pairs of pens are classified into six groups of three
homogeneous pairs for the characteristics of the environment, and the 36 lambs
into six groups of three males and three females, according to six dates on which
they become available at the specified age of 12 weeks. The six groups of three
pairs of pens are associated with the six groups of animals to constitute six
homogeneous blocks of pairs of pens and animals. Then, the three doses of the
anthelmintic are assigned to the three pairs of pens of each of the six blocks and,
on each of the dates on which animals become available, three pairs of male and

77
 Scientific Research: An Introduction
Scientific Research Methods

female animals are allocated to the three pairs of a blocks. Thus, the differences
between blocks, due to pairs of pens and times of entry of the animals are
eliminated from the differences between anthelmintics and the research error. c)
If the plants of each of the three degrees of incidence of the virus are
heterogeneous, they should be ordered as to the level of vigor and then
classified for matching according to the vigor level, each of the groups consisting
of plants with the three degrees of winding virus. Thus, the variation due to the
vigor level can be eliminated from the differences between degrees of infection
and the research error (Silva, 2022c).

[Link] Statistical control

- The statistical control comprises recording data for a variable expressing a

relevant extraneous characteristic and using it to adjust the observed
values of the response characteristic, to eliminate variation attributable to
the controlled extraneous characteristic.

With an appropriate statistical analysis procedure, called covariate analysis, the

variation due to the controlled extraneous characteristic (expressed by a variable
called covariate or concomitant variable) is eliminated from the research error
and the differences in response between the levels of explanatory
characteristics.

For illustration of statistical control consider again the research on intestinal

worm control of sheep with anthelmintic. If the initial variation in the animal's
weight is relevant, it may be appropriate to exercise statistical control to adjust
the observed values of response variables to eliminate the differences
attributable to this extraneous characteristic.

Statistical control has two assumptions: the existence of a linear or otherwise

simple relationship between the response variable and the covariate, and the
absence of effect of the explanatory characteristics on the covariate.

Local control or pairing and statistical control do not interfere in the constitution of
the sample. The manifestation of controlled extraneous characteristics is
unchanged. However, the variation of the response attributable to controlled
characteristics is eliminated from differences between the levels of explanatory
characteristics and the research error. Thus, these two procedures can allow the
control of relevant extraneous characteristics, eliminating the relevant
confounding of their effects with effects of explanatory characteristics, without
preventing their manifestation in the sample (Silva, 2022c).

[Link] Randomization

Randomization is a procedure for controlling extraneous characteristics whose

manifestation is not subject to the control of research techniques, local control
and statistical control. It can be adopted in the association of levels of treatment

78
 Scientific Research: An Introduction
Scientific Research Methods

characteristics with sample units and in the execution of research techniques, to

prevent treatments from being benefited or harmed by bias.
- Randomization is the association of extraneous characteristics of the
sample units with levels of treatment characteristics by a random process
to avoid the confounding of their effects with the effects of the treatment
characteristics is biased.

Randomization is accomplished through the randomization of treatments and the

randomization of the order of the execution of research techniques.

- The randomization of the treatments consists in the assignment of the

treatments to the units of the sample by objective draw procedure that
assigns all units the same chance of receiving any of the treatments.
- The randomization of research techniques consists of the establishment of
a random order of execution of research techniques, determined by some
objective draw procedure that assigns to all units of the sample an equal
chance of being favored or disadvantaged.

In research in which matching or local control is performed, randomization must

be carried out block by block, separately and independently (Silva, 2022c).

Consider, again, the research described in previous examples. a) In the research

on control of incidence of giberela in wheat crops, the 12 combinations of the
levels of the explanatory characteristic fungicide and cultivars are randomly
assigned to the 12 plots of each block, separately and independently for each
block. b) In the research on control of intestinal worms in sheep, randomization is
carried out by the assigned of the anthelmintics to the pair of pens, separately
and independently for each block, and of sexes to the pens within pairs of pens,
separately and independently for each pair. c) Randomization is not feasible in
the research on the incidence of vine winding virus, since it does not include
explanatory characteristics of treatment.

Examples a) and b) can also illustrate randomization of research techniques: In

example a), if sowing or harvesting takes place over a considerable time interval,
randomizing the order of execution of this technique can avoid bias in its effect
on grain production; in example b), when the order of application of a
management technique, such as shearing, can affect body weight gain or wool
production, its randomization can avoid the bias of this effect.

If there are differences in extraneous characteristics that could favor or disfavor

responses to treatments, randomization ensures that all treatments have an
equal chance of being favored or disadvantaged. Thus, the confounding of the
effects of extraneous characteristics covered by randomization with treatment
effects becomes unbiased.

Randomization has no effect on the constitution of the sample and does not
avoid the confounding of effects of extraneous characteristics with explanatory
characteristics effects. However, it makes this confounding unbiased.

79
 Scientific Research: An Introduction
Scientific Research Methods

[Link] Classification of extraneous characteristics implied by sample

control

Sample control procedures are essential for deriving inferences from the sample
to the target population. Local control or pairing and statistical control contribute
to greater sensitivity to detect effects due to explanatory characteristics;
randomization allows the detection of such effects without bias.

The following classification of the extraneous characteristics of the sample may

be useful for a better understanding of the implications of the sample control
procedures adopted for the inferences derived for the target population:

controlled extraneous characteristics,

randomized extraneous characteristics and
potentially disturbing extraneous characteristics.

- The controlled extraneous characteristics are the extraneous

characteristics of the sample whose control is exercised through local
control (pairing) and statistical control; randomized extraneous
characteristics are the extraneous characteristics covered by
randomization; potentially disturbing extraneous characteristics are the
other extraneous characteristics, i.e., the uncontrolled and non-
randomized extraneous characteristics (Silva, 2022c).

The effects of potentially disturbing characteristics on response characteristics

are confounded with explanatory characteristic effects. Characteristics of this
class that manifest themselves in an irrelevant way behave as if they were
randomized, that is, their effects do not result in biased confounding with effects
of explanatory characteristics. Control of research techniques is the resource that
may be proper for this purpose. The effects of potentially disturbing
characteristics that manifest themselves in a relevant way are tendentiously
confounded with the effects of explanatory characteristics (Silva, 2022c):

- The relevant potentially disturbing extraneous characteristics whose

effects are biasedly confounded with the effects of explanatory
characteristics constitute the class of disturbing extraneous
characteristics.

The randomized characteristics and potentially disturbing characteristics

constitute the research error. For this reason, in the assumption of the absence
of disturbing extraneous characteristics, the research error is usually called
random error or casual error.

Matching or local control and statistical control are intended to reduce the effect
of bias that may be caused by disturbing extraneous characteristics and to
reduce random errors resulting from random extraneous characteristics.

80
 Scientific Research: An Introduction
Scientific Research Methods

However, it is not practical to control more than a few of the extraneous

characteristics. Most of them should be left uncontrolled. Such characteristics
should be placed in the class of random characteristics, through randomization,
or irrelevant behavior characteristics, by appropriate research techniques.

3.4.3 Explanatory Research Methods

The most common explanatory research methods are experiments and

observational studies.

[Link] Experiment

- The experiment, also called a controlled experiment, is the explanatory

research method in which there is intervention in the sample, with the
imposition of one or more treatment characteristics to its units by
randomization, and control of extraneous characteristics with experimental
techniques to avoid biased confounding of effects of treatment
characteristics with extraneous characteristics.

The sample is usually composed of units built especially for the research.
The experiment comprises at least one treatment characteristic, whose levels,
that is, treatments, are assigned to the sample units by a random process.
Typically, treatments are stimuli assigned to the units under the control of the
researcher in order to assess the corresponding effects on response
characteristics (Silva, 2022c).

The experiment is the appropriate method for the research described in

examples a) on control of Gibberella incidence in wheat crops with fungicide and
b) on control of intestinal worms in lambs with antihelmintic.

The experiment is the research method that adopts the strongest sample control.
The control of research techniques, called control of experimental techniques, is
intensively used. The grouping or classification of units to control extraneous
characteristics, which in the experiment is called local control, and statistical
control, are also very often adopted. Moreover, the experiment is the only
research method that adopts randomization in the assignment of the levels of
explanatory characteristics to the sample units.

Due to these properties, the experiment is the research method that supplies
greater reliability for inferences about causal relationships of characteristics. For
this reason, it is considered the method of explanatory research par excellence
(Silva, 2022c).

[Link] Observational study

- The observational study is the explanatory research method in which the

sample choice is limited to accessible target population units; the
explanatory characteristics are inherent to the sample units, so that they

81
 Scientific Research: An Introduction
Scientific Research Methods

manifest themselves without interference from the researcher; explanatory

and response characteristic data are recorded in the present or in the
future or are taken from records constructed in the past.

Therefore, in the observational study there is no treatment characteristic;

explanatory characteristics are all intrinsic characteristics. On the other hand, the
choice of the sample is limited: the sample consists of units selected from the
target population that are accessible (Silva, 2022c).

The explanatory observational studies comprise three main classes:

prospective observational study,

retrospective observational study and
uncontrolled observational study.

[Link] Prospective observational study

- In a prospective observational study, a group of units with the appropriate

characteristics for the purposes of the research is chosen; data of
explanatory characteristics of these units are recorded; and then these
units are followed by a time interval, during which information about the
response characteristics is recorded, sometimes corresponding to the
occurrence of an event of interest, such as death.

Thus, the explanatory characteristics are characteristic of the units present at the
beginning of the research and the response characteristics are characteristics
manifested in the future.

The prospective observational study method is illustrated by the research

example c) on the effect of the incidence of winding virus on the grape production
of vines of the cultivar Italia, in a wine-growing region.

In epidemiology, the prospective observational study method called cohort study

is used for inferences regarding the relationship between disease incidence and
risk factors. Two comparable groups of individuals are selected, one of whom
was exposed to a risk factor, to verify whether the two groups develop the
disease in different proportions. For example, in a study on the relationship
between smoking and lung cancer incidence, conducted in England between
1951 and 1961, data were collected from more than 2/3 of 60,000 physicians
who were sent a questionnaire regarding smoking. After 10 years, 135 deaths
associated with lung cancer were recorded, of which only three were
nonsmokers. The research concluded that the risk of death from lung cancer of
smokers is almost 32 times higher than the risk of death for nonsmokers (Silva,
2022c).

[Link] Retrospective observational study

- In a retrospective observational study, a group of units with the

appropriate characteristics for the purposes of the research is chosen;

82
 Scientific Research: An Introduction
Scientific Research Methods

data on response characteristics are recorded from these units; and then
historical information from these units is examined for the identification of
relevant explanatory characteristics.

In these circumstances, the response characteristics are characteristic of the

units present at the beginning of the research and the explanatory characteristics
are characteristics of the past.

In epidemiology, the retrospective observational study method called case-

control study is used to research the relationship between risk factors and the
incidence of a disease. Two comparable groups of individuals are chosen, one of
these groups (which is called a case) consisting of individuals who manifest the
disease under consideration and the other (called control) by individuals who do
not manifest it, to verify whether the two groups differ in relation to exposure to
risk factors under consideration (Silva, 2022c).

One example follows: In a study on risk factors in breast cancer conducted at the
Clinical Hospital of the Federal University of Minas Gerais, between 1978 and
1987, a group of breast cancer patients aged between 25 and 75 years (cases)
and a group of patients without indication of breast pathologies, but similar in
terms of age distributions and date of hospitalization in the hospital (control) was
considered. The research concluded that the presence of a family history of
cancer increases the risk of this pathology by 8.84 times.

These observational study methods use more restricted sample control than the
experiment. They do not adopt randomization and do not adopt the control of
research techniques or adopt it considerably less pronounced than the
experiment. Therefore, they need to resort more markedly to the use of the
control of extraneous characteristics through pairing and statistical control.

In these circumstances, prospective observational studies and retrospective

observational studies are less reliable explanatory research methods than the
experiment. However, they are research methods that the researcher must resort
to when the intervention in the sample required by the experiment is impractical
or inconvenient, often for ethical reasons. This last situation occurs in much
research in human health and in sociology (Silva, 2022c).

These methods of observational study are sometimes called quasi-experiments

and pseudo experiments because they have similar purposes as the experiment,
but the lack of control over the manifestation of explanatory characteristics and
the control over extraneous characteristics is markedly limited.

[Link] Uncontrolled observational study

- The uncontrolled observational study uses existing data from previously

collected records for some purpose other than current research. The
researcher has no control over the choice of the sample or data collection.

83
 Scientific Research: An Introduction
Scientific Research Methods

This research method receives names in some areas of research, such as pure
observational study and archive research in social sciences.

Unlike prospective observational study and retrospective observational study, in

the uncontrolled observational study, there is no data collection; data already
collected and available are used. In fact, a huge amount of data with varied and
rich information is available in archives of research institutions, public offices,
organizations specialized in data collection, hospitals, clinics, etc. For example,
national and regional agricultural research units have available data generated
by regional experiments of plant breeding, usually planned for the
recommendation of cultivars based on the comparison of averages. This data
can be exploited to generate relevant information for various purposes, by
research that requires few resources, such as: a) studies of adaptability and
stability of genotypes to the environment; b) regionalization of plant breeding
programs; c) regionalization of recommendation of cultivars; d) location of
regional experiments and determination of the appropriate amounts of sites and
years of execution of experiments (Silva, 2022c).

The uncontrolled observational study is the explanatory research method that

usually requires fewer financial resources. However, it has quite restrictive
disadvantages. The main ones are a) absence of research planning, particularly
with regard to the choice of sample, design of the research and data collection
process; b) lack of control over data generation and, for this reason, little or no
information on data quality; c) precarious representativeness, resulting from the
absence of control of the researcher over the choice of the sample, unless,
perhaps, exclusion of units according to information that may be available; d)
often, data available in files were collected and expressed in inappropriate
variables for particular research; e) very often there is no information that can
clarify doubts, especially regarding the quality of the data (Silva, 2022c).

Because of these restrictions, this research method is not recommended for the
derivation of inferences about causal relationships of characteristics. Results of
explanatory research that employs it should be used with great caution and be
more properly considered as hypotheses to be verified by reliable explanatory
research.

[Link] Choosing the explanatory search method

From an appreciation of the properties of explanatory research methods, it can

be inferred that they present important common and uncommon properties. The
common property is the non-random process of choice of the sample; in the
experiment, units are usually constructed during its execution; in observational
studies, the units are chosen among the available units of the target population
with characteristics appropriate to the objectives of the research. The non-
common property concerns the level of interference and control exercised by the
researcher: in the experiment, the researcher has control over the choice of the
levels of one or more explanatory characteristics for the sample and their
manifestation in the sample units, which can be determined by a random

84
 Scientific Research: An Introduction
Scientific Research Methods

process. This difference between experiment and observational studies is

relevant for its implication for the two desirable properties of explanatory
research: a) reliability of inferences related to causal relationships and b) validity
of the inferences for the target population. The reliability of inferences related to
causal relationships between response characteristics and explanatory
characteristics distinguishes the experiment from observational studies. As for
this property, the experiment has three important advantages (Silva, 2022c):

a) Control over the choice and variation of explanatory characteristics

clarifies the direction and nature of the causal relationship between
response and explanatory characteristics,
b) Randomization of extraneous characteristics prevents bias that can result
from the manifestation of potentially disturbing extraneous characteristics,
c) The flexibility of experimental designs allows efficient and powerful
statistical manipulation, while the analytical treatment of observational
studies presents statistical difficulties.

In fact, in the experiment, specific levels of explanatory characteristics are

chosen and applied to the units by an objective process of randomization under
the control of the researcher. In the ideal experiment, all extraneous
characteristics are controlled by experimental techniques, local control or
statistical control or randomized. Thus, with no potentially disturbing
characteristics, the bias that may come from these characteristics is eliminated.
In these circumstances, it can be logically inferred that a considerable response
variation in units with different treatments is a consequence of differences in the
effects of these treatments.

However, in the experiment, as well as in the observational study, the non-

random choice of the sample may imply low representativeness and bias of
these inferences to the target population. Consequently, in research with these
methods, in general, there is a disparity between the target population and the
sampled population that should be considered subjectively in inferences. In fact,
conceptually, sample inferences are valid for the sampled population; and will be
valid for the target population to the extent that the disparity between these two
populations, i.e., the sampling error is irrelevant. This judgment is necessarily
subjective (Silva, 2022c).

The advantages of the experiment as an explanatory research method are well

known. However, in many situations, the experiment is not feasible and, even
when feasible, presents difficulties that must be circumvented or tolerated. The
first of these is that the choice and control operation of explanatory
characteristics and extraneous characteristics can be difficult. Therefore,
delineating an "ideal" experiment can be difficult or impossible. Thus, the
advantages of the experiment related to sample control are often only relative,
not absolute. Second, it is often difficult to design experiments with
representative samples of important populations. In fact, the problem of sample
representativeness has been ignored in experimental research. Thirdly, the
elaboration of the sample with the desired realism of the conditions present in the

85
 Scientific Research: An Introduction
Scientific Research Methods

target population is generally not feasible and is often not considered in

experimental research (Silva, 2022c).

In general, in any scientific research, only part of the conditions for reliable and
valid inferences can be controlled objectively and firmly. Another part remains
uncontrolled; their implications should be evaluated by subjective judgment. The
researcher should use the available resources to maximize the objective
component and, consequently, minimize the subjective, and strive to exercise
skilled and judicious judgment. He should make the strategic choice of the
research method, considering the objectives of the research, the means for its
execution and its costs, and taking into account the following three desirable
criteria of scientific research: representativeness, realism and reliability.

The emphasis on reliability or realism tends to result, respectively, in experiments

or observational studies. As previously pointed out, the experiment is powerful in
the control of explanatory characteristics and extraneous characteristics, but it is
usually weak regarding the representation of the target population and often also
about realism. Observational studies are usually weak in control and
representativeness; their advantage is the realism of the manifestation of the
characteristics of the target population in the sample.

Usually, these three criteria cannot be adequately met in scientific research, and
often not even two of them, due to unfeasibility or lack of resources. Very
commonly, the researcher emphasizes a criterion for reasons of cost and
convenience, or because it may seem, convincingly or hopefully, the most
justifiable. On the other hand, none of the three criteria is superior to the others
for all situations and, sometimes, the three can be satisfied in complementary
research of a research program. Thus, for each situation, the researcher must
decide on a compromise between the desirable and the feasible and choose the
research strategy that best fits the available resources (Silva, 2022c).

Furthermore, each of these explanatory research methods can be improved by

efforts to overcome their main weaknesses. Surveys can be improved by using
auxiliary information to control for potentially disturbing characteristics. On the
other hand, sometimes experiments and observational studies can be improved
regarding representativeness by some compromise between the desired
amplitude for the target population and the amplitude that can be reached for the
sampled population.

3.5 USE OF SCIENTIFIC RESEARCH METHODS

A summary of the essential characteristics of the research methods described in

the previous sections may be useful for a better understanding of their
applicability in scientific research.

Exploratory research methods (case study, prototype study and mathematical

modeling) are systemic research methods, that is, research methods that focus
on units globally. The fundamental distinction between case study and prototype

86
 Scientific Research: An Introduction
Scientific Research Methods

study is that the first focuses on the existing units without the intervention of the
researcher, while in the latter the units are assembled or constructed based on
information about the units in the target population, information generated by
scientific research and opinions. Mathematical modeling performs the conceptual
representation of existing units and new units that insert recent results generated
by scientific research and opinions. Descriptive research methods (census and
sample survey) and explanatory research methods (experiment and
observational study) are analytical research methods, which focus on subsets of
characteristics of the target population (Silva, 2022c).

In the initial synthesis phase of a cycle of the scientific method, the researcher
proceeds to the overall assessment and inspection of the units of the target
population to detect the relevant characteristics, particularly the most influential
ones on the functioning and performance of the units, and the relationship of
these characteristics, with the purpose of identifying important problems and
obtaining suggestions for fertile hypotheses. This is an exploratory research
phase. Case studies and mathematical modeling are useful in this phase of
synthesis.

In the analysis phase, the descriptive survey is useful to provide a description of

the relevant characteristics and their relationships and to identify research
problems and hypotheses. Descriptive research and exploratory research are
particularly relevant for the establishment of the most effective explanatory
research paths. Then, explanatory research methods, namely experiment and
observational study, are now applicable. The new knowledge generated by this
research is characteristically sparse since it is derived from a set of research that
covers several parts of the units (Silva, 2022c).

In the final synthesis phase, this new knowledge should be inserted into the
existing body of knowledge, through new research that focuses on systems
globally. The study of prototypes and mathematical modeling are the research
methods proper for this purpose.

Briefly, this research approach requires the execution of exploratory, descriptive

and explanatory research, which requires the appropriate and strategic use of the
various research methods described in sections 3.2, 3.3, and 3.4 (Fig. 3.1).

Fig. 3.1. Use of scientific research methods in the three phases of a cycle
of the scientific method

87
 Scientific Research: An Introduction
Scientific Research Methods

To illustrate the use and interrelationship of these research methods, consider

the situation of agricultural research.

Experiments and sample surveys are traditional research methods in agricultural

research. The observational study has been little used. However, this research
method can be used to great advantage, especially the uncontrolled
observational study that may be useful for the use of the large mass of data
generated by research already performed.

Case studies and prototype studies have occasionally been used in agricultural
research. In agriculture, a case can be, for example, an agricultural property, a
farm, a crop, an orchard, or a farm facility for which the producer provides
reliable information and gives ease of access. A prototype can be one of the
same types of units installed by the research institution, on its own physical basis
or on producer property, to realistically simulate a viable production system,
incorporating results indicated by research (Silva, 2022c).

Case study and prototype study are of great potential for use to the benefit,
respectively, the better knowledge of the systems in use by producers and to test
the integration of technologies generated by research in real systems. They are
also instruments of high value for the integration of research activity with the
extension and approximation with producers. Mathematical modeling can also
fulfill these functions, with the advantage over those two empirical methods of
providing more flexibility to simulate real production systems from a much wider
range. In fact, physical models are usually restricted to a basic formula in a single
environmental situation, or to a few formulas in a few environmental situations,
due to high costs and practical restrictions. Well-developed conceptual models
are flexible, which allows their use to verify the different behaviors of systems
under various technological alternatives and various environmental conditions.
However, they need to be tested empirically, and for these tests the physical
models are relevant. Thus, these two forms of modeling are neither alternative
nor competitive: the physical model assists in the test of the conceptual model,
and this collaborates in the generalization of that model (Silva, 2022c).

Systemic research methods, particularly mathematical modeling, appeared and

developed mainly in industry applications. In agriculture, the systems are usually
more complex, which makes the construction of these models difficult and
laborious. There are few examples of success in this area and much remains to
be done. The main difficulties to overcome are, mainly, the lack of training of
researchers to understand the scientific methodology, particularly systemic
research methods, and the lack of specialized personnel in this methodology.

The use of the various methods of scientific research has reciprocal implications.
For a better understanding of the interrelated use of these research methods, it is
convenient to classify them into three groups: 1) mathematical modeling, 2) case
study and prototype study, and 3) experiment, survey and observational study.
The main interrelationships of the use of these three groups of research methods
are outlined in Fig. 3.2.

88
 Scientific Research: An Introduction
Scientific Research Methods

Fig. 3.2. Relationships between the three groups of research methods in

agricultural research
(1)→(2) Mathematical modelling suggests the assembly or modification of prototype study
and supplies recommendations for case study
(2)→(1) Case study and prototype study allow to validate and develop mathematical
modelling, and suggest new mathematical modelling
(1)→(3) Mathematical modelling generate problems and hypotheses for experiment, survey
and observational study
(3)→(1) Experimen, survey and observational study supply information for the formulation
and improvement of mathematical modelling
(2)→(3) Case study and prototype study suggest problems and hypotheses for experiment,
survey and observational study
(3)→(2) Experiment, survey and observational study suggest recommendations for
changing certain components of prototype study and case study)

It is also important to highlight the interrelationships between research within

each of the two empirical research groups: survey and observational study are
useful to suggest problems and hypotheses for experiment; case study can
supply information for use in prototype study. At the end of this chain of research
methods, prototype study generates recommendations for case study (Silva,
2022c).
___________________________________________________________________________________
© Copyright (2024): Author(s). The licensee is the publisher (B P International).
DISCLAIMER
This chapter is an extended version of the article published by the same author(s) in the following journal.
International Journal of Science and Research, 11(10), 2022. DOI: 10.21275/SR221004074851

89
 Scientific Research: An Introduction
Conclusion

CONCLUSION

Knowledge of nature has been an incessant desire of man since ancient times.
Its search spanned several stages throughout the evolution of civilization. This
search has been hampered by the complexity of nature, which comprises
complex interconnections of phenomena, and by the reach allowed by culture
and available resources. This complexity implies a variation in phenomena that
makes their control and prediction difficult. The scientific method, comprising a
set of objective, systematic, rational and critical procedures, has been the most
successful approach to generating trustful knowledge. These properties are
essential for the generation of a system of methodical knowledge about nature
that distinguishes itself from other sources of knowledge by requiring objective
empirical verification of every explanation regarding phenomena, which allows an
understanding of their nature and causes free from influences, desires and
prejudices of the observer. The special feature of science is the basic principle of
knowing the natural world through demonstrative logical arguments that require
empirical verification. Thus, a representation or interpretation of a phenomenon is
only scientific knowledge if it has been verified or demonstrated empirically.
________________________________________________________________
© Copyright (2024): Author(s). The licensee is the publisher (B P International).

90
 Scientific Research: An Introduction
Disclaimer (Artificial Intelligence)

DISCLAIMER (ARTIFICIAL INTELLIGENCE)

Author(s) hereby declare that NO generative AI technologies such as Large

Language Models (ChatGPT, COPILOT, etc) and text-to-image generators have
been used during writing or editing of manuscripts.
___________________________________________________________________________________
© Copyright (2024): Author(s). The licensee is the publisher (B P International).

91
 Scientific Research: An Introduction
Competing Interests

COMPETING INTERESTS

Author has declared that no competing interests exist.

___________________________________________________________________________________
© Copyright (2024): Author(s). The licensee is the publisher (B P International).

92
 Scientific Research: An Introduction
References

REFERENCES

Bernard, C. Introduction to the study of experimental medicine. Buenos Aires:

Losada, 1944.
Bertalanffy, L. von. Teoria geral dos sistemas. Petrópolis: Vozes, 1973.
Portuguese.
Bunge, M. La investigación científica, su estrategia y su filosofía. 4. ed.
Barcelona: Editorial Ariel, 1975. Spanish.
Bronckingtn, N. R. Sistemas, modelos y experimentos en agricultura. In: Scarci,
J.C. (editor). Enfoquen de Sistemas en la Investigación Ganadera.
Montevideo: Instituto Interamericano de Ciencias Agrícolas – OEA, 1972.
p.3-15. Spanish.
Carey, S. S. A beginner's guide to scientific method. 2. ed. Belmont, California:
Wadsworth, 1998.
Cervo, A. L.; Bervian, P. A. Metodologia científica. 3. ed. São Paulo: McGraw-
Hill, 1983. Portuguese.
Christensen, L. B. (1997). Experimental methodology. [Link]. Boston: Allyn and
Bacon.
Cochan, W. G. (1963). Sampling techniques. 2. ed. New York: John Willey.
Cochan, W. G. (1083). Planning & analysis of observational studies. New York:
John Willey.
Cox, D.R. (1958). lanning of experiments. New York: John Wiley.
Descartes, R. (1993). Discurso del método. Barcelona: Altaya. Spanish.
Federer, W. T. (1973). Statistics and society; data collection and interpretation.
New York: Marcel Dekker.
Fisher, R. A. (1962). The place of the design of experiments in the logic of
scientific inference. Colloques Internationaux du Centre National de la
Recherche Scientifique, Paris, n.110, p.13-19.
Gottschall, C. A. (2003). M. Do mito ao pensamento científico: A busca da
realidade, de Tales a Einstein. São Paulo: Ateneu. Portuguese.
Green, P. E.; Tull, D. S. (1970). Research for marketing decisions. 2. ed.
Englewood Cliffs, New Jersey: Prentice-Hall.
Heath, O. V. S. A (1981). Estatística na pesquisa científica. São Paulo:
Universidade de São Paulo. Portuguese.
Hinkelmann, K.; Kempthorne, O. (1994). Design and analysis of experiments.
New York: John Wiley.
Johann, J. R. (1997). Introdução ao método científico; conteúdo e forma do
conhecimento. 2. ed. Canoas: ULBRA. Portuguese.
Kempthorne, O. (2977). Why randomize? Journal of Statistical Planning and
Inference, v.1, p.1-25.
Kempthorne, O. (1979). Sampling inference, experimental inference and
observation inference. Sankhya, 40, Serie B, Parts 3 and 4, p.115-145.
Kish, L. (1985). Sample surveys versus experiments, controlled observations,
census, registers, and local studies. Australian Journal of Statistics, v.27,
n.2, p.111-122.
Kish, L. (1987). Statistical design for research. New York: John Wiley.
Lastrucci, C. L. (1963). The scientific approach, basic principles of the scientific
method. Cambridge, Massachusetts: Schenkman.

93
 Scientific Research: An Introduction
References

Madden, E. H. (1960). The structure of scientific thought, an introduction to

philosophy of science. Boston: Houghton Mifflin.
Mitchell, M.; Jolley, J. (1996). Research design explained. 3. ed. Fort Worth,
Texas: Harcourt Brace.
Petrie, A.; Watson, P. (1999). Statistics for veterinary and animal science.
Londres: Blackwell Science.
Shadish, W. R.; Cook, T. D.; Campbell, D. T. (2002). Experimental and qusi-
experimental designs for generalized causal inference. Boston: Houghton
Mifflin.
Silva, J.G.C. (2022a). Science and scientific method. International Journal of
Science and Research, 11(4): 621 - 633. Access 25 May 2024. Available:
[Link]
Silva, J.G.C. (2022b). Scientific research. International Journal of Science and
Research, 11(9): 635 - 648. Access 25 May 2024. Available:
[Link]
Silva, J.G.C. (2022c). Scientific research methods. International Journal of
Science and Research: 11(10): 334 – 350. Access 25 May 2024.
Available: [Link]
Smith, T. M. F.; Sugden, R. A. Sampling and assignment mechanisms in
experiments, surveys and observational studies. International Statistical
Review, v.56, n.2, p.165-180, 1988.
Urquhart, N. S. The anatomy of a study. Hortscience, v.16, n.5, p.621-627, 1981.
Wilson, E. B. An introduction to scientific research. New York: McGraw-Hill,
1952.
___________________________________________________________________________________
© Copyright (2024): Author(s). The licensee is the publisher (B P International).

94
 Scientific Research: An Introduction
Index

INDEX

A
Antiquity, 3, 8–9
Aristotle's physics, 2

B
Belgian andreas vesalius, 9
Bias, 15, 38, 42–43, 67, 75–76, 79–80, 85

C
Census, 33, 56, 60, 87
Civilizations, 4, 9
Cluster random sampling, 64–65
Continuous variable, 40
Convenience sampling, 67–68
Copernicus, 9, 11
Copernicus' heliocentric idea, 11

D
Dalton's atomic theory, 12
Deductive inference, 43
Discrete variable, 40

E
Electromagnetism, 12
Empiricism, 7–8, 24
Euclid's geometry, 2
Exploratory research, 45–46, 49, 56–58, 60, 86–87
Extraneous characteristics, 46, 70–81, 83, 85–86

F
Fatalism, 8

G
Galileo, 10–11, 26, 28
Greek theory, 2
Gregor johann mendell, 13

H
Haphazardly sampling, 67, 69

I
Inductive inference, 15, 43–44
Inductive jump, 44
Inductive reasoning, 25, 44

J
Judeo-christian religion, 4
Judgement sampling, 67

95
 Scientific Research: An Introduction
Index

K
Knowledge generation, 1–3, 7, 35

L
Lavoisier's nomenclature, 12
Law of gravity, 11, 22
Law of nature, 8
Logical positivism, 8

M
Mathematical modeling, 56–57, 59, 86–88
Mathematical techniques, 9
Milky way, 11
Mythical knowledge, 4

N
Natural phenomena, 1–4, 8, 10, 21, 27, 29–30
Neopositivists, 8
Newtonian theory of motion, 22

P
Plato's theory of ideas, 2
Polytheism, 4
Prospective observational study, 82, 84

Q
Qualitative characteristics, 31
Quota sampling, 67–68

R
Religious experience, 4
Retrospective observational study, 82–84

S
Scholastic, 9
Science strategy, 2
Scientific research, 1, 3–90
Sources of knowledge, 2–3, 6, 90
Statistical analysis, 41, 54–55, 78
Stratified random sampling, 61–64, 68
___________________________________________________________________________________
© Copyright (2024): Author(s). The licensee is the publisher (B P International).

96
 Scientific Research: An Introduction
Biography of author(s)

Biography of author(s)

João Gilberto Corrêa da Silva (Retired Professor)

Federal University of Pelotas, Pelotas, RS, Brazil.

He holds the degrees of B. Sc. in Agricultural Engineering and [Link]. and Ph.D. in Statistics. He was a
Full Professor at the Federal University of Pelotas and Researcher and Head of the Department of
Quantitative Methods at the Brazilian Agricultural Research Corporation (EMBRAPA). He was twice
President of the Brazilian Region of the International Biometric Society. He was a member of editorial
boards and a reviewer for scientific journals, and a consultant at universities and research institutions.
He taught courses in Mathematics and Theoretical and Applied Statistics, and extension courses for
students and researchers. He has an interest in scientific research, particularly in the areas of biology
and agriculture. In these areas, he has guided students at the master's and doctoral levels. He
participated and presented communications and conferences at scientific meetings, published articles in
scientific journals and conference proceedings, and books, book chapters and textbooks. In 2019, he
was awarded the University Merit distinction from the Federal University of Pelotas.
___________________________________________________________________________________
© Copyright (2024): Author(s). The licensee is the publisher (B P International).

Peer-Review History:
This book was reviewed by following the Advanced Open Peer Review policy. This book was thoroughly checked to
prevent plagiarism. As per editorial policy, a minimum of two peer-reviewers reviewed the manuscript. After review and
revision of the manuscript, the Book Editor approved the manuscript for final publication. Peer review comments,
comments of the editor(s), etc. are available here: [Link]

97

View publication stats