Diploma in Statistical Analysis

Getting to know
the data
 2

Contents

3 Introduction

4 Lesson outcomes

5 Study design

# Data types

# Measures of central tendency

# Conclusion

# Additional resources

# References
3

Introduction
This lesson will be all about study design, data types and a sneak peek into summarizing data. Understanding the study
design and the type of scales that are used to measure data, is a crucial step in analysing the data accurately. Only after we
know the pros and cons of the way the data was gathered; can we start describing the data.

Study design
Study design is the framework of research each with its own advantages and disadvantages. We use the study
design to help us answer the question posed to us by the stakeholders. We can also describe study design as a
structure or set of procedures that we use to gather and analyse data.
Study designs can be divided into two groups:

1. Observational studies
2. Experimental studies

Observational study design

An observational study is a study where researchers observe the effect that, for instance, risk factors or diagnostic tests
have on, for instance, the patient, without trying to make changes to exposed patients.

Observational studies can be divided into four main study design classifications:
1. Descriptive or case-series studies
2. Case-controlled studies
3. Cross-sectional studies
4. Cohort studies

Descriptive or case-series studies

• When describing specific characteristics of a group of patients in a report, the study is referred to as a case-
series study.
4

• A case-series study is often over a shorter period.

• For example, a 1997 study introduced a group of 40 patients with information considering stroke, transient
ischemic attacks, or carotid bruits. The researcher wanted to differentiate between 2 methods to see which
is better at predicting peak systolic velocity. This study therefore had no control group or hypothesis.

• Descriptive studies are often meant as a precursor to other studies. Descriptive studies are one of the
simplest study designs and is often a precursor to other studies that would then subsequently be
investigated in a case-controlled, cross-sectional or cohort study.

Case-controlled studies

• Case-controlled studies, unlike case-series studies, begin with the lack of an outcome and look backwards
to try and recognize possible causes or risk factors.

• Case-controlled studies can be used for, identification of causes of a specific disease. Individuals
participating in the study could be chosen based on having the disease. The control group would be
individuals that do not have the disease.

• The history of both groups is then evaluated to try and find the risk factors involved in the cases of the
patients that have the disease in comparison to those that do not.

• Case-controlled studies span over a longer period than descriptive studies.

Cross-sectional studies
the third type of observational study design we will delve into, is cross sectional studies. this study design is also
known as survey studies, prevalence studies and epidemiologic studies.
cross sectional studies analyse data on a group of subjects at a specific point in time rather than over a period of
time
for example, if the researcher wants to learn more about the relationship between demographic information and
trauma patients identification with elevated blood alcohol levels, they could choose a specific time of day, gender
and age group to investigate and create 3 models based on a variety of these factors.

Cohort studies
We use a cohort study when we have a group of patients that have characteristics in common and who remain part
of a group over an extended period of time. Because we are evaluating events that will happen after the study has
begun, this study is prospective in nature.
for example, the researcher might want to evaluate the cause of incidence of a disease. We want to know what will
happen to the group of patients.
The researcher could evaluate a group of patients with, for example, hypertensive cardiovascular disease and like
the 1970 study of gordon and kannel, then conduct a comprehensive 20 year follow up.
this could lead to identification of the risk factors for these patients. in fact, journal articles have been written about
this cohort study and some of the children of this study are now being followed as well.
5

Experimental studies
An experimental study is when an intervention is introduced to the group of patients and the effect is evaluated
that the intervention has on the group. Experimental or clinical studies that involve humans are referred to as
clinical studies. The purpose is to draw conclusions about a specific treatment or procedure the patients are
undergoing.

Experimental studies can be either:

1. Controlled studies
2. Uncontrolled studies

Controlled studies

• A controlled trial is when the researcher compares one procedure or drug to another. This could be a
placebo or a previously accepted procedure. The aim of the research is to determine whether the
intervention will be different to the accepted procedure.

• Controlled trials can be concurrently controlled, where one group is the experimental group and the other
the placebo or controlled group. These two groups are then treated the same except for the procedures’
difference itself.

• Controlled trials can also have sequential controls or external controls

Studies with no controls

Studies with no controls are not by exact definition experiments or trials, because they do not have interventions
with controls. Nonetheless, these studies are useful in specific instances.
For example, the 1997 study of Crook and Associates described the results of a trial of radiotherapy for prostate
carcinoma in which patients were followed for a minimum of 12 months and maximum of 70 months after the
procedure. The researcher’s aim was to determine how long patients had no repeating tumours as well as how long
the patients survived after the treatment. This study was uncontrolled because there were no comparisons with
patients treated in another manner.

Data types
Qualitative vs Quantitative
Data can be either qualitative or quantitative. Quantitative data is typically measures of values and denoted with
numerical values whereas qualitative data is measured by different types of data such as names or symbols.

Quantitative data

Quantitative data, which we will work with in this course, is numerical and is more objective than qualitative data. It is
drawn from questions that are typically more close-ended. We use quantitative data to validate our hypothesis.
6

Qualitative data

Qualitative data is non numerical data that can be collected through interviews like focus groups. Typically qualitative
data is gathered from open ended questions and is subjective. We use this type of data to generate the hypothesis and
answer the “why” to our question.

Scales of measurement
Scales of measurement refer to ways in which variables/numbers are defined and categorized. Each scale of measurement
has certain properties which in turn determines the appropriateness for use of certain analyses.

In this course, we subdivide scales of measurement into three common categories: nominal, ordinal and numerical.

Nominal

• Nominal scale of measurement is seen as the simplest level of measurement.

• Data is nominal when the Values are words describing categories
• There values can also be referred to as the naming or labelling variables
• This data type needs a Minimum of 2 categories.
• This type of data is Not ordered or ranked.
• This type of data is best represented by a Pie chart or a bar graph
• Examples of this type of data include: Gender (male or female or other), marital status (married, unmarried,
divorced, partnered, single, and so forth).

Ordinal

• An inherent order exists amongst categories of ordinal data.

• Data that is ordinal is very similar to nominal data, but the category items can be ordered or ranked to be less
than or greater than the other. it is best represented by a box plot.
• An example of this type of data includes rankings for athletes. Athletes may be ranked as first, second or third.
However, we do not know how far the first athlete won nearly that they won.

Numerical

• The difference between the observations carry meaning on a numerical scale.

• This scale is sometimes referred to as quantitative observations.
• The two types of numerical scales are continuous and discrete.
• Continuous scales measures numbers on a continuum and a discreet scale measures numbers equal to integers,
for example as number of fractures.

Summarising data
Measures of central tendency
Summarizing data helps us make sense of the mass amount of information that we have collected. some useful
information starts emerging when we indicate how the centre of the distribution looks. These numbers are collectively
7

referred to as measures of central tendency. Three measures of central tendency include the mean, median, and the
mode. These measures are all used for numerical data.

Mean
∑𝑿
• ̅=
𝑿 , ∑ means to add or to sum, X represents the observations and n the number of observations.
𝒏
• The mean is also known as the arithmetic mean of the observations.
• The mean is not used for an ordinal scale and is sensitive to outliers or extreme observations.

Median

• The median is known as the middle observation or the 50th percentile. The median is the point where half of the
data will be smaller than the observation and half of the data will be larger than the observation.
• There is no standard symbol for the median, but some sources refer to it by M or 𝑀𝑑 . We arrange the observations
in the sample from smallest to largest and choose the observation in the middle as the median value.
• The median is less sensitive to outliers than the mean.

Mode

• The mode represents the numbers that appear the most frequently or the observations that are the most
commonly occurring.
• The mode is also known as the modal observation and when a set of observations has two modes, it is called
“bimodal”. We more commonly use the mode when dealing with qualitative information.
• The mode is not affected by extreme outliers.

Additional resources
Red wine data set link: [Link]
8

References
• Dawson, B. & Trapp, R.G., 2004, Basic & Clinical Biostatistics, 4th ed., United States of America,

McGraw-Hill.

• Dillard, J., 2017, The data analysis process: 5 steps to better decision making, Big Sky Associates,

[Link]

Decision-Making

• [Link]

• [Link]

• Institute for work and health, 2016, Observational vs. experimental studies, Institute for work &

health,

• Ranganathan, P., & Aggarwal, R. (2018). Study designs: Part 1 - An overview and classification.

Perspectives in clinical research, 9(4), 184–186. [Link]

• Research Guides, 2019, Data Module #1: What is research data?, Research Guides,

[Link]

• Toves, J., 2017, SAS University Edition, PAT Research,

[Link]

edition/#:~:text=SAS%20University%20Edition%20is%20a,%2C%20medical%2C%20health%20and

%20engineering.