Chapter

9
Common quantitative methods
Linda Shields and Roger Watson

KEY TERMS LEARNING OUTCOMES

control After reading this chapter, you should be
experimental designs able to:
explanatory (independent) variable • describe the overall purpose of
quantitative designs
manipulation
• distinguish the differences between
observational designs experimental, quasi-experimental and
outcome (dependent) variable non-experimental designs
quasi-experimental designs • list the criteria necessary for inferring
randomisation cause-and-effect relationships
• apply critical review criteria to evaluate
the findings of selected quantitative
studies.

Schneider_1374_Chapter 9_main.indd 161 7/25/2012 6:11:08 PM

 NURSING AND MIDWIFERY RESEARCH 4E

INTRODUCTION
This chapter provides an overview of the meaning, purpose and issues related to quantitative
research designs, and presents the common approaches used to answer a variety of nursing and
midwifery questions. Related issues such as sampling, data collection, assessment of
measurement instruments and data analysis are discussed in Chapters 10–13. The focus is on
providing research consumers with the information to evaluate quantitative studies critically.
Quantitative research refers to studies where the variables of interest are measurable and the
results are quantifiable and coded as numerical data. Figure 9.1 (Bryman & Cramer 2005)
shows one way of looking at quantitative research and shows how it can be divided into two
branches: survey (also called observational and correlational) and experimental. Both branches
include a range of designs but both are capable of testing theory by asking research questions
through the specific identification of independent and dependent variables, the formulation of
hypotheses and their testing using statistical methods. However, these two branches are not
equally capable of discerning cause and effect between variables and there is, essentially, a
hierarchy in the ability to do this as illustrated in Table 9.1. There are three major categories of
designs along the quantitative continuum: observational, quasi-experimental and experimental.
Each category includes a range of designs. Choice of a design relates to the research question or
hypothesis, the amount of control a researcher can have and study feasibility.
Much of what we know as quantitative research methods have been designed by the
discipline of epidemiology. Interestingly, one of the first epidemiologists was Florence
Nightingale. Statistical tests which she developed to produce evidence about outcomes for
wounded and sick soldiers at the Crimea were recognised at the International Congress of
Statistics in 1860 (McDonald 2001). Modern epidemiology (from Greek, meaning ‘the study
of people’) is defined by the Australasian Epidemiological Association (2010) as the study of
diseases in populations. Epidemiology has three main aims: to describe disease patterns in
human populations; to identify the causes of diseases (also known as aetiology); and to provide
data essential for the management, evaluation and planning of services for the prevention,
control and treatment of disease. There are many excellent books on epidemiology written
expressly for nurses and midwives; a recent example is Epidemiology for Advanced Nursing
Practice by Kiran Macha and John McDonough, published in 2011 by Joyce and Bartlett
Learning in the USA. Any midwife or nurse undertaking quantitative research is well advised to
learn the basics of epidemiology.

CONCEPTS UNDERPINNING illustrates the focus on variables and relationships

for each type of quantitative design.
QUANTITATIVE RESEARCH The term ‘design’ implies the organisation
The empirical paradigm that informs quantitative of research components into a coherent and
research designs emphasises ‘objective’ systematic plan, and represents the major
observation, accuracy and control — elements distinctive approach chosen for the specific
originating from scientific investigations in purpose of answering an explicit research
laboratory settings. Medicine in particular, but question. The question or aims and objectives
also other disciplines such as nursing and influence and guide the choice of design. To
midwifery, have followed this tradition when make an informed choice about which design
undertaking quantitative research. Figure 9.2 will best answer the research question, and for

162

Schneider_1374_Chapter 9_main.indd 162 7/25/2012 6:11:08 PM

 9 • Common quantitative methods

Theory

Hypothesis

Operationalisation
of concepts

Selection of
respondents or participants

Survey/correlational design Experimental design

Conduct interviews Create experimental

or administer and control groups
questionnaires

Carry out observations

and/or administer tests
or questionnaires

Collect data

Analyse data

Findings

Figure 9.1 The research process (From Bryman A and Cramer D 2005. Quantitative
data analysis with SPSS 12 and 13: a guide for social scientists. Taylor & Francis, London.)

TABLE 9.1
Continuum of quantitative research designs
DESIGN INCREASING
CONTROL AND ABILITY
TO ASSIGN CAUSALITY SUB-TYPES FEATURES
Observational Descriptive Describes variables
Correlational Examines relationships between variables
Cross-sectional Examines variables and relationships at one time
point
Retrospective Re-traces participants with an outcome backwards
to a possible exposure
Case-control Cases matched to controls
Cohort Follows participants from exposure to outcome
Longitudinal Repeated measurements of participants over time
Quasi-experimental Time-series Manipulation (intervention)
Non-equivalent Manipulation + Control
control group
Experimental After-only No measurement prior to intervention
experimental
Randomised Control & Manipulation + Random assignment to
controlled trial groups
Cluster-randomised Groups of participants (not individuals) are
controlled trial randomised

163

Schneider_1374_Chapter 9_main.indd 163 7/25/2012 6:11:08 PM

 NURSING AND MIDWIFERY RESEARCH 4E

Study design Variables and relationship less control over these elements, particularly
of interest three major concepts:
Descriptive 1) control
2) randomisation
Variable Variable
1 2 3) manipulation.
The strongest designs contain elements of all
Identifies variable/s; not able three of the above aspects.
to test relationships
Correlational Control
Control is the presence of constants in a study,
Variable Variable
1 2
including controlling for extraneous variables,
using comparison groups and implementing an
explicit study protocol so that all participants
Direct relationship between
identified variables;
are involved in the same way. In experimental
direction un-tested research the comparison group is the control
group (that receives the usual care or treatment),
Cohort/interventional
rather than the innovative experimental group
under investigation. In some longitudinal or
Variable Variable
1 2 crossover research designs, participants act as
their own controls.
The strength and rigour of a quantitative
Direct ‘cause-and-effect’
relationship tested
design relates to controlling the effects of any
extraneous variables that may cause bias (threats
Figure 9.2 Examining variables in quantitative to internal validity, such as selection, history and
designs maturation) and influence the study findings.
These variables can be either antecedent or
intervening. An antecedent (preceding) variable
occurs before the study commences but may
affect the outcome variable of interest and
influence findings (e.g. age, gender,
socioeconomic status or pre-existing health status
could affect outcomes such as recovery time and
a consumer to understand the implications and ability to integrate healthcare behaviours).
the use of research, a clear understanding of Similarly, an intervening (mediating) variable is
designs is important. On this basis alone, not part of the study design, but occurs during
clinicians can begin to determine whether a the course of the study and may influence the
paper describing a research study could be of outcome variable (e.g. a change in the model of
value in informing their practice (Greenhalgh clinical care during a longitudinal study). Threats
2010). to ‘internal validity’ are discussed further in
One of the most valuable places to find Chapters 11 and 12.
up-to-date information about quantitative
research is the website of the Cochrane
Collaboration: [Link] Randomisation
Important elements in quantitative research Randomisation (random assignment) to study
include objectivity in the conceptualisation of groups is designed to ensure that groups or
the problem, operational definition, accuracy, participants are similar on the variables of
feasibility, control of the experimental interest, so that differences in the outcome
environment, hypothesis testing, replication, variable can be attributed to the intervention. If
internal validity and external validity. Selection possible, participants are assigned to either the
of a particular design gives a researcher more or experimental or control group on a purely

164

Schneider_1374_Chapter 9_main.indd 164 7/25/2012 6:11:09 PM

 9 • Common quantitative methods

BOX 9.1 RESEARCH IN BRIEF

Excerpts from experimental
• Random assignment is the random
studies of random assignment allocation of participants (or whatever one is
to study groups trying to investigate) to the intervention or
control group of an experimental study.
‘This is a randomised controlled trial. 120 sets • Random sampling is a process of selecting a
of computer-generated random numbers were representative group of participants from
used, and patients who fitted the criteria were the population of interest.
randomised to the study or control group.’ • Randomisation need not be complicated.
(Wong et al. 2010 p 270) For small numbers, series of coin tosses
(heads into one group, tails into another) or
‘Using random stratified sampling from the
pulling names out of a hat (1 — into control
total population of hospital staff …’ (Shields
group, 2 — into experimental group, 3
et al. 2011 p 156)
— into control group, and so on) are valid
‘… participants were randomly allocated to ways of random allocation.
one of the three groups: … Randomisation • An important point for randomisation
was performed in blocks separately for each techniques: if at all possible, the people who
hospital, using labelled cards in numbered are doing the research should not know how
closed envelopes prepared by a statistician the randomisation falls, or who goes into
not involved in the study.’ (Goedendorp et al. what groups. ‘Blinding’ or ‘concealment’ to
2010 p 1124) the researchers makes for a much more
rigorous study, as lack of these means that
‘Eligible participants were asked whether they
the researchers might inadvertently influence
had ever had a class in which they practiced
the responses that the participants give. If
CPR (yes/no) and then assigned by coin flip to
this is not possible, it is important that
receive or not receive a cell telephone for
non-blinding or non-concealment is noted in
manikin resuscitation.’ (Merchant et al. 2010
the limitations section of the published
p 539)
study.

Manipulation
Manipulation is relevant only in interventional
(experimental or quasi-experimental) studies, not
observational designs. A researcher manipulates
the causal or ‘explanatory (independent) variable’
random, or chance, basis. Each participant by introducing a ‘treatment’ or ‘intervention’.
therefore has an equal and known probability of The intervention (treatment) group receives
being assigned to any group. If randomisation is manipulation of the explanatory variable, while
not possible, the design is quasi-experimental. the control group does not. The intervention
Random assignment may be done individually might be a treatment, a teaching plan or a
or by groups, with a variety of approaches medication. The effect of this manipulation
available — from a very simple coin toss or is measured to determine the effect of the
‘draw from a hat’, to more sophisticated intervention. Studies using true experimental
techniques such as a table of random numbers, designs in clinical settings are called randomised
or computerised random number generation controlled trials (RCTs), reflecting the high level
(see Box 9.1). of control imposed and the importance of
Randomisation assumes that important randomisation in gaining evidence for causation
intervening variables are then equally distributed (discussed later in this chapter). A study protocol
between the groups. Studies should therefore or procedure ensures that all participants in the
report how groups actually compared on intervention group receive similar treatment, and
important variables prior to any intervention (see assists the reader in understanding the nature of
‘Research in brief ’). the experimental treatment.

165

Schneider_1374_Chapter 9_main.indd 165 7/25/2012 6:11:09 PM

 NURSING AND MIDWIFERY RESEARCH 4E

The three broad categories of observational, powder — not the researchers, nor the treating
quasi-experimental and experimental designs are nurses and doctors, nor the patients themselves.
discussed below, after the section on ‘blinding’. Only once the trial was finished, and the data
collection complete, did the statistician and
computer programmer, who had set up the
Blinding coding, open the code to know which patient
Many studies are described as ‘blind’, or ‘double had received which intervention, and the effects
blind’, a technique used most commonly in analysed. The herbal medicine was found to be
clinical trials of a drug or some sort of an effective treatment for constipation.
intervention. Blinding means that some
pertinent information is hidden or withheld
from those conducting the research and, RESEARCH IN BRIEF
oftentimes, from the participants as well. In this A randomised controlled trial of an
way, there can be no influence, or intervention of home visits by nurses on
‘contamination’ from those applying the research, 12-year-old children of economically
or the treatment, or by the participant creating a disadvantaged women (n = 613) were
‘placebo effect’. A crude (and imaginary) example compared with a control group (Kitzman et al.
is a study of an ointment for wound healing. 2010). The families were similar in both
The ointment containing the relevant medication experimental and control group, but only the
is being contrasted with another one which does experimental group had the home visits. The
characteristics under examination included use
not contain the medication (a placebo). If the
of cigarettes, alcohol and marijuana in the 30
researchers and their assistants, or the nurses days before interview, a range of educational
working with the patients, know which ointment tests in reading and mathematics and reports
is being applied to the wound, they may of behaviour. The authors found that the
influence the results by, perhaps, allowing the children having the home visits were less likely
medicated ointment to stay on longer, or to smoke, drink alcohol or use drugs, and
cleaning the wound better for application of were less likely to have psychological
the medicated ointment rather than the non- disorders. When the mothers were visited by
medicated one. By blinding the ointment so nurses, the children scored higher
no-one knows which one is being applied, and academically than those in the control group,
then contrasting the results, a true indication of but there was no difference in the incidence of
behavioural problems.
the effectiveness of the ointment will be found.
The authors used appropriate statistical
Of course, it would be good if the patients on tests to examine the equivalence of the
whom the ointment is being tested don’t know groups for categorical variables.
either. This level of blinding, where the people
doing the applying and those receiving the
application, do not know which one is being
used, is called ‘double blinding’. To give a real
OBSERVATIONAL DESIGNS
example (rather than the imaginary one I have Observational (non-experimental) designs are
just given), this process is commonly used in used when a researcher wishes to construct a
drug trials. Only by ensuring that no-one knows picture of a phenomenon or explore events,
which patients receive the drug, and which ones people or situations as they naturally occur in
received the placebo, can true results, free of the environment. The aim is therefore to observe
contamination, be obtained. and identify variables of interest and explore
A Chinese study (Huang et al. 2011) of relationships between those variables. These
the use of a herbal medicine, CCH1, for designs are used when:
constipation was carried out in three long-term • there is little information or research about the
care units. A powder containing the medicine or topic of interest
a placebo (contained no medication) was given • the phenomena of interest are not amenable to
to patients randomly assigned to either the experimental designs (e.g. when ethical
control or experimental (intervention) group. considerations do not allow a ‘treatment’ to be
No-one knew who was receiving which type of denied to a control group).

166

Schneider_1374_Chapter 9_main.indd 166 7/25/2012 6:11:09 PM

 9 • Common quantitative methods

Observational designs work from a clear, involves only limited or no interaction with
concise problem statement based within an participants. Observers can use a pre-identified
appropriate theoretical framework. Although classification system to categorise observations or
a researcher does not actively manipulate an document free-text ‘field notes’ (see Chapter 12
intervention variable in this design, the concepts for further discussion of participant observation).
of control and rigour are still important and A survey is another common type of
should be evident to the reader. Demographic observational design (using either a questionnaire
data are collected to describe the sample of or interview) that enables collection of
participants, such as age, gender, income level, information about the characteristics of
ethnicity, occupation and educational level. particular individuals, groups, institutions or
Reporting this information assists a reader in situations, or about the frequency of a
considering the generalisability of the study and phenomenon’s occurrence, particularly when
the potential application of the findings to their little is known within the positivist paradigm.
own practice setting. The common types of Variables of interest commonly investigate
observational designs are discussed below, noting participants’ knowledge, beliefs or attitudes
their advantages and disadvantages, and using about a particular topic or concept. The terms
examples from published clinical research to ‘exploratory’, ‘descriptive’ and ‘survey’ are used to
illustrate features of specific designs. describe this study design. Note that the study
purpose is only to relate one variable to another.
Relationships between variables or the direction
RESEARCH IN BRIEF of an effect cannot be tested. There are both
In a study of attitudes to working with children
advantages and disadvantages to consider when
and their parents in a tertiary children’s undertaking survey research (see Table 9.2).
hospital, staff (nurses, doctors, allied health Importantly, descriptive studies cannot make
and ancillary staff) were randomly selected comparisons between groups or determine
from the total population of a hospital (Shields causality.
et al. 2011). They were asked to complete a A study of lesbian, gay, transgender or bisexual
questionnaire with two questions from which (LGBT) parents seeking healthcare for their
scores were derived for working with children children included a survey of attitudes to
and working with their parents, and the two homosexuality held by nursing and medical
were compared across the groups. Staff gave students (Chapman et al. 2011). While there is
a significantly more positive score for working
with children than working with their parents.
This was the first study of its kind to examine
such aspects of family-centred care, and TABLE 9.2
the randomised sample facilitated the Advantages and disadvantages of survey studies
generalisability of the findings. Further work
ADVANTAGES DISADVANTAGES
on this study will include subset analyses to
see if different occupational groups give A lot of information Large-scale surveys can
different scores, if the length of time working can be obtained be time-consuming and
in paediatrics has an influence, and if those from a large costly.
participants who are parents themselves give a population in an
different score for working with children and economical manner.
parents, than those who do not have children. Accurate if sample Information tends to be
is representative of brief and superficial, with
the population of breadth rather than depth
interest. of data the outcome.
Descriptive/exploratory studies
Requires expertise in a
Descriptive studies use a variety of approaches variety of research areas
to measure the variable/s of interest, including (e.g. sampling techniques,
observation or questioning of participants via a questionnaire construction,
questionnaire or interview, or health service data interviewing, data analysis)
sets, for example. Direct observation of study to produce a reliable and
participants from a quantitative perspective valid study.

167

Schneider_1374_Chapter 9_main.indd 167 7/25/2012 6:11:09 PM

 NURSING AND MIDWIFERY RESEARCH 4E

a large literature around differing sexualities, TABLE 9.3

few have examined the use of health services Advantages and disadvantages of
by LGBT parents. Students become health correlational studies
practitioners, and so this study shed light on if ADVANTAGES DISADVANTAGES
and where education should be provided to Increased flexibility to Unable to determine
ensure that once the students graduated, they investigate relationships a causal relationship
were able to communicate with LGBT parents in among variables. between variables
an open and sensitive manner. because of the lack of
manipulation, control
and randomisation.
Correlational studies Efficient and effective No random sampling
An extension of descriptive research is to explore method of collecting a possible with pre-
the relationships between variables that provide a large amount of data existing groups — the
about an issue of ability to generalise is
deeper insight into the phenomenon of interest.
interest. therefore decreased.
Correlational studies enable examination of the
relationship between pairs of variables, as well Provides a framework Unable to manipulate
for exploring the the variables of
as a comparison between groups. A researcher
relationship between interest.
uses this design to quantify the strength of the variables that are not
relationship between variables (i.e. as one able to be manipulated.
variable changes, does a related change occur in
A foundation for
the other variable?). For example, Shields et al. potential, future
(2006, 2010) investigated the correlation interventional studies.
between breastfeeding up to age 3 months and
obesity at 14 and 21 years of age. Using a
longitudinal data set of over 7000 people, Cross-sectional studies
logistic regression was used to determine Surveys that measure data at one point in time
models which would demonstrate possible (i.e. data are collected on only one occasion with
correlations. At age 21, whether or not people the same participants rather than on the same
were breastfed had no effect on whether or not participants at several points in time) are called
they were overweight or obese, but at age 14, cross-sectional studies. These studies are
there was a trend that indicated a possible effect. categorised as either ‘descriptive’ or ‘analytic’
However, maternal education level and BMI (Schoenbach 1999). Importantly, cross-sectional
were found to be far more influential on long- analytic (CSA) studies use inferential statistics
term outcomes for overweight and obesity. See to infer a relationship between two or more
Chapter 14 for explanation of correlation data variables.
analysis. To examine the prevalence of obstructive sleep
Correlation designs cannot, however, test a apnoea (OSA) and resulting quality of life (QoL)
‘cause-and-effect’ relationship (i.e. whether a in cardiac patients in India, 50 congestive heart
change in one variable causes a measurable failure patients were matched with 50 healthy
change in another variable). When reading a controls and, using several validated
correlational study, identify the variables and the questionnaires, a cross-sectional study was carried
relationship being tested, and consider whether out (Patidar et al. 2011). Of the cardiac patients,
the implied relationship is consistent with the 18% had OSA, as did 8% of the control group.
conceptual framework and research question Excessive daytime sleepiness was significantly
being asked. A common misuse of a correlational associated with OSA in the cardiac patients (P =
design is a researcher’s attempt to conclude that a 0.02), while clinical severity and duration of
causal relationship exists between the study illness were not. Increased body mass index and
variables. Despite this limitation, the design is neck circumference significantly increased the
useful — clinical practice applicability relates to risk of OSA; QoL was poor, and was worse in
the study question and feasibility. The advantages those cardiac patients with OSA.
and disadvantages of correlational studies are This design was appropriate to measure the
listed in Table 9.3. variables of interest in a broad sample at one

168

Schneider_1374_Chapter 9_main.indd 168 7/25/2012 6:11:09 PM

 9 • Common quantitative methods

point in time. It used a convenience sample, (e.g. lung cancer, heart disease). The study
so there was a potential for selection bias, and direction is retrospective, from the study
the questionnaires used were self-report, so outcome backwards to the cause or the exposure.
potentially were not objective. The use of Individuals with the outcome of interest are the
inferential statistics was appropriate and ‘cases’. Cases are matched with ‘controls’, a
demonstrated links between certain physical sample of people who come from the same
factors and OSA in this sample. ‘study base’ (similar to the cases for
A limitation of this design is that the outcome characteristics such as age, gender or clinical
of interest and the causal factor are measured procedure), but do not have the outcome of
simultaneously. The lack of time as a factor interest. There are commonly multiple controls
therefore provides only weak evidence of matched to each case (2–3 : 1). Detailed past
causality. In contrast, a longitudinal study which histories, particularly in relation to possible risk
followed a group of people for a period of time factors (exposure), are examined from both
during a series of studies would provide more groups to see if the cases have been more
strength of causality (see later in this section). exposed to the suspected causative factor variable
than the controls.
Retrospective studies
The aim of a retrospective study is to link present RESEARCH IN BRIEF
outcomes to some past events. The design is
also called ‘ex-post facto’ (favoured by social A case control study in the USA (McHugh
et al. 2011) was designed to examine patient
scientists), or comparative, and is similar to a
and hospital characteristics associated with
case-control study (see below). In a study of severe poor glycaemic control, an indication of
records kept by child health nurses at an severe complications of diabetes. All patients
Aboriginal community in Australia, the growth (cases) (n = 261) with poor glycaemic control
of children of 13 families was retrospectively admitted to California acute care hospitals
tracked over three generations (Alsop-Shields from 2005 to 2006, and 261 controls, were
1997). There were statistically significant matched using data for age, sex, major
increases in weights across the generations at diagnostic category and severity of illness. The
birth, one year and five years of age, indicating statistical test used was adjusted odds ratio
that the children’s health had improved over the (OR) for experiencing poor glycaemic control.
45 years of the records. In a Taiwanese study of Mortality (16% vs 9%, p = 0.01) and total costs
($26 125 vs $18 233, p = 0.026) were higher
peripherally inserted central catheters (PICCs)
among cases with poor glycaemic control.
(Ting-Kai 2011), failures of the devices were When the type of hospital was tested, levels of
evaluated by comparing two periods of nurse staffing was not significantly influential in
retrospective case studies, before and after teaching hospitals (OR: 0.98, 95% CI: 0.88–
implementing changes in nursing practice for 1.11), but in non-teaching hospitals each
PICCs. Wound oozing, infection, phlebitis, additional nursing hour per adjusted patient
occlusion and leaking rates improved significantly day significantly reduced the odds of poor
between the two periods, providing evidence that glycaemic control by 16% (OR: 0.84, 95% CI:
the changes in nursing care were effective. Such 0.73–0.96). The authors ascribed these results
a study would provide more rigorous results if about registered nurse hours in teaching
done as a RCT, but the investigators reported hospitals to possible increased complexities of
patients’ conditions (though this should have
that such a trial was not possible for them and
been accounted for by the matching of control
that this study provided them with a good basis and experimental participants), but the authors
from which they could design further studies. tell us that while their matching procedures
were as rigorous as possible, there were
Case-control studies several clinical factors that could not be
matched, a common problem with this sort of
A case-control study is an epidemiological
study. Despite this, case control studies are
approach which examines participants on the valuable ways of comparing different
basis of a study outcome (clinical characteristic, interventions or experiments.
condition or disease) that is or is not present

169

Schneider_1374_Chapter 9_main.indd 169 7/25/2012 6:11:09 PM

 NURSING AND MIDWIFERY RESEARCH 4E

Selection bias may exist if the cases are not exposure and the subsequent study outcome.
from a well-defined study base (in time and This may be partly overcome in an historical
place). A critical reader therefore needs to cohort study if some of the data of individuals
cautiously evaluate the conclusions drawn in exposed in the past are available for use.
relation to measurement error. Advantages are
similar to those of a correlational design, but a
higher level of control is possible. Disadvantages Longitudinal studies
include an inability to draw a causal linkage In contrast to a cross-sectional design,
between the two variables (an alternative longitudinal studies collect data from the same
hypothesis may cause the relationship), and group of participants at different points in
finding naturally occurring groups of participants time (repeated measures). By collecting data
who are similar in all respects except for their from each participant at certain intervals, a
exposure to the variable of interest is very longitudinal perspective of the outcome variable
difficult. is possible. An example of a longitudinal cohort
study is the classic Nurses’ Health Study, one
of the largest cohort studies of risk factors for
Cohort studies major chronic diseases in women. In 1976 over
A cohort study is an epidemiological approach 120 000 married registered nurses aged 30–55
where the direction is from the exposure to were initially enrolled in the study (Belanger
the outcome, or cause to presumed effect. et al. 1978) and the cohort continues to be
Commonly, a researcher studies the development followed (Baer 2011), examining a wide range of
of a particular health outcome or disease state. issues including diet, menopause and breast
Participants are selected from a population cancer. At the time of writing, contemporary
known to be free of the health outcome under articles continue to be published from follow-up
study, and then classified according to whether assessments; over 1200 articles have been
they have one or more explanatory variables published from this data set from 1978–2010
hypothetically related to the outcome. These (see the Nurses’ Health Study website in
participants (referred to as the cohort) are then ‘Additional resources’ at the end of this chapter).
studied over a time period ranging from days to Similarly, the Framingham heart study examined
months to years, to determine who develops the the effect of blood pressure, cholesterol levels,
outcome of interest. The incidence of the study smoking, exercise and other variables on the
outcome is observed in relation to exposure to development of coronary artery disease in a
possible causes/risk factors (Grimes & Schulz cohort of healthy men, at specified intervals over
2002). a period of years (e.g. Margolis et al. 1974; Ho
Cohort studies can be cross sectional, or et al. 2010).
longitudinal. Cross-sectional studies examine There are three types of longitudinal design,
cohorts at one point in time, or perhaps each with their advantages and disadvantages:
sequences of time points, while longitudinal 1) longitudinal trend studies
cohort studies investigate cohorts over long time 2) longitudinal cohort studies
periods. In 2010 Sanada et al. used cross-
3) longitudinal panel studies
sectional cohorts of patients with pressure ulcers
to test the efficacy and cost effectiveness of a and each of these will be considered in turn.
system of pressure ulcer management by skilled
nurses, over a three-week period. Longitudinal trend studies
Cohort designs can be used to examine This type of design is good for looking at general
relationships both retrospectively and trends in populations; however, it is not suitable
prospectively. Prospective (longitudinal) studies for following individuals. Trend studies are used,
explore hypothesised causes, differences or for example, by newspapers in the run up to a
relationships, and move forward in time to the general election and are carried out by taking
presumed effect. A disadvantage of this design is regular samples from the population to find
the time period and high costs involved, as there out what likely voting patterns are going to be.
may be a considerable time lag between time of Therefore, trend studies use the same population

170

Schneider_1374_Chapter 9_main.indd 170 7/25/2012 6:11:09 PM

 9 • Common quantitative methods

but different samples from it over time. Trend TABLE 9.4

studies are unique in longitudinal designs in Advantages and disadvantages of
being immune to attrition, or dropout, from the longitudinal studies
study as they make no effort to re-sample the ADVANTAGES DISADVANTAGES
same people (although they may do this Each participant is Long duration of data
incidentally) and sampling at each stage simply followed separately collection — costly in
proceeds until sufficient participants have and therefore acts as terms of time, effort
responded. her or his own control. and resources.
Both relationships and Threats to internal
Longitudinal cohort studies differences can be validity include
This type of design uses a sample—called a explored between ‘testing’ and ‘mortality’
cohort—and then runs over time at regular variables. (loss to follow-up),
intervals by sampling from that cohort. For and the influence of
example, the cohort may be all the students confounding variables.
graduating in one year in nursing from a large Changes in the Social desirability bias
university nursing program. A university may be variables of interest are is possible (participants
assessed over time; respond in a way they
interested in following up this cohort over the
early trends in data believe is congruent
years after graduation and they can do this by can be investigated with the researchers’
sending out questionnaires to students in the at a subsequent expectations) —
group. The sampling can be done at random measurement. ‘Hawthorne effect’.
from the cohort or the whole cohort may be
sampled. However, over time, attrition will take
place due to people becoming impossible to trace
and some people may not respond at one time RESEARCH IN BRIEF
point but may reappear the next. Therefore, in
this design, there is no effort to sample exactly The Mater University Study of Pregnancy
(MUSP) ([Link]
the same people, although many of the same
musp) has for nearly 25 years studied a cohort
people will respond; rather it is information of about 8000 children, and has published
about the whole group that is being sought over 200 papers about the growth and
through those sampled at each stage. Clearly, this development of children. Shields et al. (2006,
design is more vulnerable to attrition than the 2010) used the MUSP data set to examine the
trend design. effect of breastfeeding on children at aged 14
and 21 years respectively. From a total of 7776
Longitudinal panel studies children originally recruited at the beginning
This type of design is similar to the cohort of the study, in the 14 year study, data were
design as it follows a defined group of people; available for 3698 children and 2553 at 21
however, unlike the cohort design, a panel study years. Such drop off is a characteristic of
longitudinal studies with a long timeframe, and
is designed to follow individuals over time and
statistical analysis of the loss to follow-up is an
makes every effort to re-recruit and, thereby, inherent part of subsequent studies. In the
re-sample precisely the same people at each stage 14-year-old children, breastfeeding for longer
in the study. This is a very powerful way of than 6 months was protective of obesity (OR
studying individual change and individual 0.6, 95% CI 0.4, 0.96), but not of overweight.
differences over time but provides less However, when confounding variables were
information about a whole group than a cohort considered the effect size diminished and lost
study and little information about the statistical significance (OR 0.8, 95% CI 0.5, 1.3).
population. However, panel studies are very In short, breastfeeding for less than 6 months
vulnerable to attrition as, once someone is lost had no effect on either obesity or overweight,
to the study, they rarely reappear and, while though a trend was found for increased
prevalence of overweight at 14 years with
statistical methods exist to compensate, their
shorter periods of breastfeeding. By age 21,
data are essentially lost to the study. The there was no influence on either overweight or
advantages and disadvantages of a longitudinal obesity from breastfeeding.
design are listed in Table 9.4.

171

Schneider_1374_Chapter 9_main.indd 171 7/25/2012 6:11:09 PM

 NURSING AND MIDWIFERY RESEARCH 4E

QUASI-EXPERIMENTAL
Point to ponder DESIGNS
When assessing the appropriateness of a Quasi-experiments are designs where a researcher
cross-sectional study compared to a manipulates an experimental treatment
longitudinal study, first examine the (intervention; explanatory variable) but some
purpose of the study in relation to the characteristic of a ‘true’ experiment is lacking
design. — either control or randomisation. Both designs
There are many longitudinal data sets test cause-and-effect relationships, but the lack of
available for study around the world, and control and/or randomisation in quasi-
the custodians of these are keen for other
investigators to use them. Consider
experimental designs threatens the study’s
approaching authors if you are interested in internal validity and weakens any causal
doing this sort of work. It is an interesting inference. Many different quasi-experimental
approach to quantitative research that can designs exist; the common types are discussed
yield not only worthwhile results, but also here (see Figure 9.3).
publications in high ranking journals. The Suppose a researcher is interested in the effects
following is a list of such datasets: of a new cardiac education program on the
Mater University Study of Pregnancy physical and psycho-social outcome of patients
(MUSP): ([Link] newly diagnosed with acute coronary syndrome.
[Link]/musp)
A researcher could randomly assign participants
Avon Longitudinal Study of Parents and
to either the group receiving the new program
Children (ALSPAC): [Link]
[Link]/alspac/ or the group receiving the usual program, but
for any number of practical reasons, that design
Raine Study: [Link]
is not possible. One difficulty is when patients
Peel Child Health Study: [Link]
on the same clinical unit have to receive both
[Link]/
experimental and control interventions, without
The Dutch Famine Birth Cohort Study:
[Link]
any overlap. The potential for participant
[Link] ‘contamination’ needs to be considered by the
researcher — how to stop participants receiving
or adopting the ‘wrong’ intervention. The
choices are to abandon the experiment, or
Causality in observational designs conduct a quasi-experimental study. With the
As noted in this section, observational designs latter, a researcher can find a similar clinical unit
are often limited in their ability to determine that has not introduced the new educational
‘causality’ — where one variable influences program and study the newly diagnosed cardiac
another in a cause-and-effect relationship. patients who are admitted as the comparison
Historically, only experimental research has been (control) group.
able to support the concept of causality. There
are, however, many instances in clinical research
where experimental studies cannot be conducted
because of ethical or practical reasons. To
overcome this limitation, several statistical RESEARCH IN BRIEF
analytical techniques are available to explain ‘Study contamination’ occurs when participants
relationships among variables and establish causal allocated to a specific study group receive the
links. These ‘causal modelling’ analyses (also alternative intervention. It can also occur when
termed ‘causal analysis’, ‘path analysis’, ‘Linear participants in the experimental group
Structural Relation Analysis [LISREL]’ and become aware of the effects of the
intervention from an external source; perhaps
‘structural equation modelling [SEM]’) are
people in the two groups compare effects and
introduced in Chapter 13 and discussed further so each group may express outcomes that are
in other specialty statistics texts (see ‘Additional not due to the intervention under study.
resources’ at the end of this chapter).

172

Schneider_1374_Chapter 9_main.indd 172 7/25/2012 6:11:10 PM

 9 • Common quantitative methods

A Non-equivalent control group design

Experimental pre-test experimental treatment post-test

group

Control pre-test post-test

group

B After-only non-equivalent control group design

Experimental experimental treatment post-test

group

Control post-test
group

C One group pre-test post-test design

Experimental pre-test experimental treatment post-test

group

D Time series design

Experimental pre-test pre-test experimental treatment post-test post-test

group

Control pre-test pre-test post-test post-test

group

Figure 9.3 Comparison of quasi-experimental designs

There is no random assignment of patients TABLE 9.5

by the researcher in this situation; therefore Advantages and disadvantages of quasi-
a quasi-experimental design. There is, however, experimental studies
a range of potential threats to validity when ADVANTAGES DISADVANTAGES
interpreting the results of quasi-experimental Practical, feasible Need to rule out any
studies (see Table 9.5). and able to be plausible alternative
generalised. explanations for the
Non-equivalent control findings — control by
design or statistical
group studies analysis.
This design is the quasi-experimental version of a May be the only Unable to make clear
‘true experimental’ design, except participants are design to evaluate cause-and-effect
unable to be randomised to study groups because some hypotheses, inferences.
of practical or feasibility issues (Figure 9.3, A). particularly in
Despite the lack of randomisation, this design clinical settings.
is commonly used in clinical research as it is More adaptable
relatively robust to threats to internal validity. to the real-world
Collection of data at ‘pre-test’ (measurement practice setting
than controlled
prior to the intervention) allows comparison
experimental
of the two groups before the intervention is designs.
introduced, and any differences between groups
can also be controlled during data analysis. This
feature strengthens the influence of the
intervention, and minimises any effects of
extraneous variables.

173

Schneider_1374_Chapter 9_main.indd 173 7/25/2012 6:11:10 PM

 NURSING AND MIDWIFERY RESEARCH 4E

Under the next heading there is an example of

RESEARCH IN BRIEF a study of an intervention of a teaching leaflet
A quasi-experimental design was used to about pressure ulcers for older people (Hartigan
evaluate three nursing interventions — et al. 2011). Measuring the participants’
occlusive wrap, chemical mattress and knowledge after the education leaflet would not
regulation of delivery room temperature— tell us whether their knowledge levels differed
singly and in combination on thermoregulation before they received the leaflet, or whether the
in six groups of low birth weight infants
education motivated individuals to learn more
(Lewis et al. 2011). 133 infants under 1500
grams were included, and the interventions
about their health problem. Therefore a
tested on different groups of infants over 3 researcher’s conclusion that the teaching program
years. A control group of 295 infants on which improved physical status and psycho-social
retrospective chart data were available over an outcome would be subject to the alternative
earlier 3-year period was used. For each of the conclusion that the results were an effect of
interventions, the percentage with a normal pre-existing knowledge (selection effect) in
temperature in NICU in each intervention combination with greater learning in those more
group was larger than the percentage of the motivated (selection–maturation interaction). See
control group, but it was not a significant Chapters 11 and 12 for further description of
difference. Each intervention — occlusive wrap these threats to study validity.
alone, occlusive wrap in addition to chemical
mattress and occlusive wrap in addition to
chemical mattress and increased delivery room One group pre-test–post-
temperature — had a positive influence on
thermoregulation.
test studies
This design is used when only one group is
available for studying the effects of an
After-only non-equivalent control intervention (Figure 9.3, C). The lack of
group studies randomisation and a control group limit the
internal validity and generalisability of any
Suppose a clinical researcher did not measure findings. A one group pre-test–post-test design
specific characteristics of participants before the was used in an Irish study about older people
introduction of a new intervention, but later and pressure ulcers (Hartigan et al. 2011). In
decided that it would be useful to have data Ireland, a quasi-experimental study of a patient
demonstrating the effect of the program. Perhaps education leaflet about pressure ulcers for older
a health service manager asked for these data to people was conducted (Hartigan et al. 2011).
determine whether the extra cost of a new teaching Uncontrolled, it used a pre- and post-test to
program was justified. The study that could be assess 75 participants’ level of knowledge about
conducted would be the after-only non-equivalent preventative strategies for pressure ulcers. A tool
control group design, shown in Figure 9.3, B. This to test knowledge was developed and then
design minimises ‘testing effects’ (completing a applied to both pre- and post-test groups.
pre-test can affect the post-test scores), but Post-intervention, the participants’ knowledge
assumes that the two groups are equivalent before increased about: what a pressure ulcer was (68%
the intervention is introduced. pre-test — 91% post-intervention); causes
(77%–89%); prevention when sitting in a chair
(79%–91%) and a figure (no numbers or
percentages were given for knowledge about
Points to ponder signs and symptoms) showed a marked
Check that two non-randomly assigned improvement of knowledge post-intervention,
groups are comparable on important and knowledge about common sites was high
characteristics at the beginning of the study.
and not changed by the leaflet. While the
Examine the concept of ‘loss to follow-up’ statistics in this study were very simple
and how it can influence statistical results:
[Link]
percentages only, it is still valid. More
resources/glossary/e---l/loss-to-follow-up/ complicated statistics may have yielded more
definitive information about the changes, but for

174

Schneider_1374_Chapter 9_main.indd 174 7/25/2012 6:11:10 PM

 9 • Common quantitative methods

the purposes of a study such as this, simple 0.63–0.81). After the intervention was stopped,
results can be just as informative as the most the caesarean rate increased again to 31.8% (RR
complicated statistical analyses. In fact, they 1.19 CI 1.09–1.32), but it was still a decrease of
are often more informative as they are easy to 5% from the first, basal period (RR 0.86 CI
follow and understand, and are therefore more 0.76–0.97). The authors concluded that their
readily critiqued and eventually applied in new system was an effective and safe strategy to
practice. reduce the rate of caesarean section. They suggest
Attempting to include a control group in this that a strength of the study is that it was
environment would have led to study designed in steps, with the first period allowing
contamination, but limited generalisability. There them to identify the main groups contributing to
was minimal description of the training program. the overall caesarean rate, creating ‘groups of
The measuring instrument was previously interest’ that represent more than 50% of the
developed by the researcher; acceptable reliability caesarean sections in their maternity ward.
was reported but validity was not discussed. The However, they suggest that the Hawthorne effect
sample was small and in one site only; these (behaviour change because of knowing one is
limitations were noted by the author. part of a study) may have influenced the
findings, they had no control group with which
Before–after design to compare and, to quote the authors, a major
limitation pertinent to time period studies was
Sometimes researchers use a ‘before–after’ design that ‘an interrupted time series methodology has
to test the effects of an introduced intervention the potential bias that it does not control all
when participants are different individuals before variables, including the fact that seasonal
the intervention than after the intervention. The variations can influence the results’ (Scarella
Irish study (Hartigan et al. 2011) again provides et al. 2011 p 139).
an example. Even with the absence of a control group, the
number of data collection points minimises
threats to validity such as ‘history’ effects.

1 Tutorial Trigger
Given the information in the ‘before–
after design’ section, describe the
reasons why an RCT design was not
However, ‘testing’ is a threat to validity with this
design because measures are repeated so many
times. Why alternative explanations for the
findings in a quasi-experimental study are not
possible in the Irish study (Hartigan plausible should be discussed by the researcher.
et al. 2011).
In some cases, clinical or practical knowledge of
the problem and patient population can suggest
that a particular explanation is not plausible.
Time-series studies Nonetheless it is important to replicate studies to
Another approach when only one group is support any causal inferences developed through
available is to study that group over a longer quasi-experimental designs.
period, using a time-series design (illustrated in
Figure 9.3, D). To rule out alternative
explanations for the findings of a one-group EXPERIMENTAL DESIGNS
pre-test–post-test design, the variable/s of interest An experiment is a scientific investigation
are measured over a longer period while that involves observation and data collection
introducing the intervention at some time according to explicit criteria and protocol.
during the data collection period. Experimental designs have all three identifying
Investigators from Chile (Scarella et al. 2011) properties: randomisation, control and
used three time points to study the effects of a manipulation. These designs are used to test
new classification system on caesarean section ‘cause-and-effect relationships’ between an
rates, over a 21-month period. At the first time intervention (treatment) and an outcome, and
point (3 months after implementation), the rate minimise or control any alternative explanations
had reduced from 36.8% to 26.5%, and again in (threats to validity) for the study findings
the second period (6 months; RR 0.71 CI (Thompson 2004). To infer causality requires:

175

Schneider_1374_Chapter 9_main.indd 175 7/25/2012 6:11:10 PM

 NURSING AND MIDWIFERY RESEARCH 4E

• the explanatory (causal) and outcome (effect) There are several different experimental
variables must be associated with each other designs, each based on the classic ‘true
• the ‘cause’ to precede the ‘effect’ experiment’ presented in Figure 9.4, A.
• that the relationship is not able to be explained Participants are randomly assigned to the control
by another (extraneous) variable/s. or experimental group(s), with the treatment
Experiments can be classified by setting. given only to those in the experimental group.
While ‘field’ experiments such as an RCT and The outcome variable is measured in both groups
‘laboratory’ experiments share the same design as a pre-test and post-test (before and after the
characteristics and properties, they are conducted intervention is introduced).
in fundamentally different environments. When reviewing experimental (or quasi-
Laboratory experiments take place in an artificial experimental) studies, the prime focus is on
setting created specifically for the purpose of assessing the validity of the findings. Did the
research, enabling almost complete control by a intervention (explanatory variable) cause the
researcher. Field experiments, however, occur in desired effect on the outcome variable? Validity
a real, existing social setting such as a hospital of the findings and conclusion depend on how
unit, clinic or community where the well the researcher has controlled other variables
phenomenon of interest actually occurs. that may also explain any change in the outcome
Conducting an experimental study can be variables. Although random assignment and
challenging, as all relevant variables need to be control in the classic experimental design
identified and then controlled, manipulated or minimise the effects of many threats to internal
measured. However, they are the only way that validity, it is not perfect in practice as some
nurses and midwives can generate research relevant threats are difficult to control. In particular,
to their knowledge and evidence requirements. ‘mortality’ effects (a threat to internal validity)
Unlike the physical sciences, nursing and may be a problem for studies with long
midwifery researchers continue to identify follow-up periods, as participants tend to drop
important complex concepts and relationships out because of the burden over an extended
that are the province of our disciplines. Nurses and period of time (called ‘loss to follow-up’
midwives are ideally placed to use the broadest [CONSORT 2011]). There may be important
range of research methods available, pertinent to differences between participants who withdraw
the topics they wish to examine. Nursing and and those who complete the study; these
midwifery, with their emphasis on interactions differences may explain the study findings, not
with human beings, are often the best disciplines the intervention. Guidelines for reporting
from which thoroughly conceptualised and experimental studies have been developed by the
rigorously prepared human factor evidence and CONSORT (consolidated standards of reporting
knowledge can be generated. This puts nurses and trials) group for reporting these types of threats
midwives in a particularly privileged position, as (Moher et al. 2001; see ‘Additional resources’ at
they are best able to determine the appropriate the end of this chapter), and it is very important
design of a study to answer certain questions to examine and discuss any loss to follow-up in
relating to the human factors which are an studies as these may influence the study’s
inherent part of the two disciplines. findings (Hennekens et al. 1987).
While most experiments in nursing and

2
midwifery are field experiments and control is
such an important element, it is common that Tutorial Trigger
a researcher cannot control contamination. The group assignment for your
Conversely, studies conducted in the laboratory research subject is to critique an
are by nature ‘artificial’, and therefore not assigned quantitative study. To
proceed, you must first decide on the design
‘real-world’. Although laboratory experiments
to be used: you think it is an ex-post facto
have stronger internal validity than fieldwork design; the others in the group think it is an
studies, their weakness is with external validity. experimental design because it has several
When reading research reports, it is important to explicit hypotheses. How would you convince
consider the setting of the experiment and what them that you are correct?
impact this may have on the study findings.

176

Schneider_1374_Chapter 9_main.indd 176 7/25/2012 6:11:11 PM

 9 • Common quantitative methods

A True or classic experiment

Experimental pre-test experimental treatment post-test

group
Random
assignment
Control pre-test post-test
group

B Solomon four-group design

Experimental pre-test experimental treatment post-test

group

Control pre-test post-test

group
Random
assignment
Control experimental treatment post-test
group

Control post-test
group

C After-only experimental design

Experimental experimental treatment post-test

group
Random
assignment
Control post-test
group

Figure 9.4 Comparison of experimental designs

RANDOMISED CONTROLLED control group gets no experimental treatment

but is also measured for comparison with the
TRIALS experimental group. The difference between the
A randomised controlled trial (RCT) is the two groups at post-test reflects whether a causal
clinical equivalent of a true experiment, and is link actually exists between the explanatory and
the ‘gold standard’ for testing cause-and-effect outcome variables.
relationships in clinical research (Thompson In a study about breastfeeding rates in a
2004). As illustrated in Figure 9.4, A, health region in the UK (Jolly et al. 2011), 2724
participants are randomly assigned to the women gave birth following antenatal care in
experimental and control groups, so that any 66 clinics, so 33 clinics (1,267 women) were
pre-intervention differences (antecedent randomised to the intervention (a service given
variables) are measured and/or controlled. These by peer support workers [PSW] in both ante-
pre-test measures or observations provide a and postnatal periods) and 33 clinics (1457
baseline score for verifying similar characteristics women) to usual care. Of those, 848 women
between control and treatment groups, and consented to additional follow-up by
therefore the effect of the intervention. The questionnaire at 6 months. Data about
intervention is then introduced to the ‘treatment’ breastfeeding (any and exclusive) at 10–14 days,
group and the outcome variable is again 6 weeks and 6 months were obtained from
measured to see whether it has changed. The routine records. Loss to follow-up was 6% at

177

Schneider_1374_Chapter 9_main.indd 177 7/25/2012 6:11:11 PM

 NURSING AND MIDWIFERY RESEARCH 4E

10–14 days and 22% at 6 months. There was no in Figure 9.4, C. This design is composed of two
difference between intervention and usual care in randomly assigned groups, but neither group is
any breastfeeding at 10–14 days (OR 1.07, 95% measured at pre-test. The process of randomly
CI 0.87–1.31, p = 0.54), and the proportion of assigning participants to groups is assumed to be
women reporting any breastfeeding in the sufficient to ensure a lack of bias in determining
intervention group at 6 weeks was 62.7% and whether the treatment created significant
64.5% in the usual care group (OR 0.93, 95% differences between the two groups. This design
CI 0.64–1.35); and at 6 months was 34.3% and is particularly useful when testing effects are
38.9% respectively (OR 1.06, 95% CI 0.71– expected to be a major problem and the number
1.58). The authors concluded that this peer of available participants is too limited to use a
support service did not improve breastfeeding Solomon four-group design.
rates up to 6 months in this population.
However, the loss to follow-up of almost a
quarter of the original sample may have
influenced the findings — if all had been Point to ponder
included at 6 months, would the results have
Experimental studies are not the most
been different? commonly used designs in nursing or
midwifery research. Experimentation assumes
Solomon four-group design that all of the relevant variables involved in a
phenomenon have been identified and are
A more complex experimental design used to measurable. For many areas of nursing and
minimise the effect of repeated ‘testing’ (a threat midwifery this is not the case, and descriptive
to internal validity) is the Solomon four-group or exploratory studies need to be completed
design. This design has two groups identical to before experimental interventions can be
those used in the classic experimental design, but applied to a clinical issue.
two further groups, an experimental after–only
group and a control after–only group are also
included. As Figure 9.4, B, illustrates, all four Cluster-randomised controlled trials
groups have randomly assigned participants. A cluster-randomised controlled trial (c-RCT)
Addition of the last two groups helps to rule out uses coherent groups or clusters of individuals,
‘testing’ threats that the pre-test groups may such as nursing homes (Pettersson et al. 2011),
experience. This design enables evaluation of the classes of students (e.g. Chen & Chung 2011),
effect of the pre-test on the post-test. A midwifery clinics (e.g. Jolly et al. 2011) or
limitation, however, is the need for a larger hospitals (Shields et al. 2007) as the random
sample size when compared to a two-group allocation unit, not individuals. All individuals
design. In contemporary research, this design is in the group allocated to the intervention then
not common in nursing or midwifery. receive the experimental treatment, eliminating
the potential for study contamination. These
Evidence-based practice tip studies are, however, expensive to conduct as the
‘cluster’ is the unit of analysis and any variance
Solomon four-group designs have come under
methodological scrutiny. A systematic review
within the group works against any intervention
of studies about behaviour using this process effect (Christie et al. 2009) (see ‘Additional
found that no studies could be included due resources’ at the end of this chapter).
to significant methodological quality issues Chapter 15 discusses the critical review
(McCambridge et al. 2011). The statistics for process, which is directed towards evaluating the
this design are complex and so more study is appropriateness of the study design in relation to
needed into their feasibility and application. factors such as the research problem, theoretical
framework, hypothesis, methods and data
analysis and interpretation. The overall purpose
After-only design of reviewing studies is to assess the validity of
Another alternative design is the after-only findings and to determine whether these findings
design (or post-test-only control group), shown are worth incorporating into your professional

178

Schneider_1374_Chapter 9_main.indd 178 7/25/2012 6:11:11 PM

 9 • Common quantitative methods

BOX 9.2 to the validity of the findings. You must then

Criteria for evaluating quantitative decide if the author’s explanations are clear and
designs logical.

a) What design was used in the study

(observational, quasi-experimental or
RESEARCH IN BRIEF
experimental)?
b) Which specific type of the above design The study above about a peer support
was used in the study, and was it intervention to improve breastfeeding rates
appropriate? used cluster randomisation (Jolly et al. 2011).
This design was used for practical reasons in
c) Was the design appropriate for the
preference to an RCT, but equally it helped to
research problem and data collection
avoid contamination across study groups
methods?
within individual clinics.
d) Was the problem examining a cause-and-
effect relationship?
e) Were the common threats to the validity of
findings for this design addressed? Cluster randomisation could be used to study
f) Was the design suited to the study setting? a model of care that is in use across many
g) Were the limitations of the design
institutions. Family-centred care is widely
adequately discussed? used in children’s hospitals, but is untested
(Shields et al. 2007). An RCT using an intensive
h) Were there other limitations related to the
design that you identified but they were
family-centred care model involving the whole
not discussed? hospital could be conducted, with one group of
hospitals receiving the family-centred care
i) For observational designs, did the report
go beyond the parameters of this design,
intervention, while the other group used their
and infer cause-and-effect relationships normal everyday practice. Each hospital would
between variables? constitute a cluster. However, there is probably
j) For experimental designs, how were only one country where this could be successfully
randomisation, control and manipulation carried out. The United Kingdom has 33
applied? children’s hospitals all run along similar lines
k) What were the plausible alternative under the National Health Service, so that
explanations for findings, and were they would remove variables that might influence the
discussed and discounted? outcomes such as different ways of running the
l) Are the findings able to be generalised to hospital. Also they are geographically far enough
other practice settings and the larger apart to prevent contamination between the
population of interest? clusters. However, this would be extremely
expensive to conduct, and so other ways to test
the model called ‘family-centred care’ have to be
devised.

practice and/or as the evidence base for current SUMMARY

institutional clinical practices. Criteria for the Quantitative research methods enable a
review of quantitative designs relate to how well researcher to manipulate numerical data to
any potential biases in the methods have been answer specific research questions and draw
addressed (see Box 9.2). inferences from their findings. Observational
The most important question to ask as you (non-experimental) designs are used to construct
read experimental studies is, ‘What else could a picture or description of events as they
have happened to explain the findings?’ This naturally occur, but cannot establish cause-and-
question of potential alternative explanations effect relationships between variables. Quasi-
will be addressed by a well-written report, experimental designs are frequently used in
which will systematically review potential threats clinical research to explore cause-and-effect

179

Schneider_1374_Chapter 9_main.indd 179 7/25/2012 6:11:11 PM

 NURSING AND MIDWIFERY RESEARCH 4E

relationships, as there are times when variable and to randomly assign participants to
experimental designs which can test this study groups.
relationship are impractical or unethical to The following four chapters discuss,
conduct within that setting. True experiments respectively, sampling, data collection, measuring
are characterised by control of extraneous instrument assessment and data analysis issues
variables, manipulation of the explanatory for quantitative research.

KEY POINTS
• Three main strengths of quantitative approaches are the objectivity, precision and control
afforded through design, sampling strategies and analytical tests.
• Experimental and other quantitative designs allow nurses and midwives to justify in a
scientific manner the outcomes of their actions and provide the basis for effective high
quality and evidence-based clinical practice.
• Extraneous and confounding variables represent a major influence on the interpretation of
any quantitative study as they limit the validity of the study and generalisability of the
findings. This is particularly evident with observational designs.
• When reviewing experimental studies, check that the author has addressed all of the
relevant issues, such as evidence of a representative sample, and minimal and unbiased loss
to follow-up.

Learning activities
1. Quantitative research uses numbers, 5. Non-experimental correlation studies are
counts and statistical tests. used frequently in nursing and midwifery
True research so that the findings from this
design can be generalised to larger
False
populations.
2. A study using questionnaires with
True
numerical scores to examine feelings,
perceptions and social effects is a False
qualitative study. 6. A time series design has two groups
True identical to the true experimental design
plus an experimental after-group and a
False
control after-group.
3. Administration of an intervention to one
True
group of participants and not another is
an experimental study. False
True 7. A true experiment design includes three
properties — randomisation, control and
False
manipulation.
4. Bias is introduced when factors extraneous
True
to the study influence the findings.
False
True
False

180

Schneider_1374_Chapter 9_main.indd 180 7/25/2012 6:11:11 PM

 9 • Common quantitative methods

8. Control is implemented in the design of a 13. Identify whether the following studies are
study by manipulation of the outcome experimental or quasi-experimental. Use
variable. the abbreviations E for experimental and
True Q for quasi-experimental.
False a) 100 elderly people with leg ulcers are
randomly assigned into an experimental
9. A randomised controlled trial is the most
wound support group and a regular
powerful type of design for examining
support group. Before the program and
cause-and-effect relationships.
at the end of the 6-month program,
True wound healing patterns are compared
False between the two groups.
10. When data are collected multiple times b) Parents on two separate neonatal units
before and after the introduction of the are given a stress score questionnaire to
intervention, the study is a non-equivalent complete at the end of their infant’s
control group design. first hospital day and on the day of
True discharge. The infants on one unit
False receive care directed by a nurse case
manager, and those on the other unit
11. In quasi-experimental designs, one of the receive care from the usual rotation of
characteristics of a true experimental nurses. Parent stress scores are
design is lacking. compared.
True c) Student midwives are randomly
False assigned to two groups. One group
12. Loss to follow-up is one of the potential receives an experimental independent
disadvantages of longitudinal studies. study program and the other receives
True the usual classroom instruction. Both
groups receive the same post-test to
False evaluate learning.
d) A study was conducted to compare the
effectiveness of shortened fasting times
for adults pre-anaesthetic on post-
operative recovery. Participants were
randomly assigned into groups and
post-operative nausea and vomiting
were measured in the first day
post-operatively.

Additional resources Critical Appraisal Skills Program (CASP), UK National

Health Service Critical Appraisal Tools —
Avon Longitudinal Study of Parents and Children randomised controlled trials; cohort studies; case
(ALSPAC) Online. Available: http:// control studies. Online. Available: http://
[Link]/alspac/ [Link]/casp
Cochrane Collaboration 2011. Online. Available: Dutch Famine Birth Cohort Study. Online. Available:
[Link] [Link]
Cook T D, Campbell D T 1979 Quasi-experimentation: Fletcher R H, Fletcher S W 2005 Clinical
Design and Analysis Issues for Field Settings. Epidemiology: the Essentials, 4th edn. Lippincott
Rand-McNally, Chicago, USA Williams & Wilkins, Baltimore, USA

181

Schneider_1374_Chapter 9_main.indd 181 7/25/2012 6:11:11 PM

 NURSING AND MIDWIFERY RESEARCH 4E

Hanson B P 2006 Designing, conducting and bisexual and transgender parents seeking health care
reporting clinical research: a step by step approach. for their children. Journal of Clinical Nursing doi:
Injury, International Journal of the Care of the 10.1111/j.1365-2702.2011.03892.x
Injured 37:583–94 Chen H S, Chung C H 2011 The learning effectiveness
Macha K, McDonough J 2011 Epidemiology for of nursing students using online testing as an
Advanced Nursing Practice. Joyce and Bartlett assistant tool: A cluster randomised controlled trial.
Learning, New York, USA Nurse Education Today. Early view, doi:10.1016/j.
Mater University Study of Pregnancy 2011 Mater nedt.2011.03.004
University Study of Pregnancy (MUSP). Online. Christie J, O’Halloran P, Stevenson M 2009 Planning a
Available: [Link] cluster randomised controlled trial: methodological
Medical Research Council (United Kingdom) 2008 issues. Nursing Research 58(2):128–34
Complex interventions guidance. Online. Available: CONSORT 2011. Loss to follow-up. CONSORT:
[Link] Transparent Reporting of Trials. Online. Available:
[Link]?d=MRC004871 [Link]
Moher D, Schulz K F, Altman D G, for the glossary/e---l/loss-to-follow-up/ [accessed 20
CONSORT Group 2001 The CONSORT November 2011]
statement: revised recommendations for improving Goedendorp M, Peters M E W J, Gielissen M F M,
the quality of reports of parallel-group randomised et al 2010 Is increasing physical activity necessary
trials. Lancet 357:1191–4 to diminish fatigue during cancer treatment?
Munro B H 2005 Statistical Methods for Health Care Comparing cognitive behavior therapy and a brief
Research, 5th edn. Lippincott Williams & Wilkins, nursing intervention with usual care in a
Philadelphia, USA multicenter randomised controlled trial. The
Oncologist 15(10):1122–32
Nurses’ Health Study. Online. Available: http://
[Link]/nhs/[Link] Greenhalgh T 2010 How to Read a Paper, 4th edn.
Wiley Blackwell, Oxford, UK
Peel Child Health Study. Online. Available: http://
[Link]/ Grimes D A, Schulz K F 2002 Cohort studies:
marching towards outcomes. Lancet 359:341–5
Polivka B J, Nickel J T 1992 Case-control design: an
appropriate strategy for nursing research. Nursing Hartigan I, Murphy S, Hickey M 2011 Older adults’
Research 41:250–3 knowledge of pressure ulcer prevention: a
prospective quasi-experimental study. International
Raine Study. Online. Available: [Link] Journal of Older People Nursing doi:
[Link]/ 10.1111/j.1748-3743.2011.00274.x
Weinert C, Burman M 1996 Nurturing longitudinal Hennekens C H, Buring J E, Mayens S L (ed) 1987
samples. West Journal of Nursing Research 18:360–4 Epidemiology in Medicine. Lippincott, Williams and
Wilkins, Philadelphia, UK
References Ho J E, Lee D S, Gona P, et al 2010 Abstract 15840:
Alsop-Shields L 1997 The growth of children for two Clinical predictors of heart failure with preserved
generations at an Australian Aboriginal Community. vs reduced ejection fraction. Data from the
Journal of Pediatric Nursing 11(6):402–8 Framingham Heart Study Circulation 122:
A15840
Australasian Epidemiological Association 2010 What is
epidemiology? Online. Available: http:// Huang C H, Su Y C, Li T C, et al 2011 Treatment of
[Link] [accessed 4 June 2012] constipation in long-term care with Chinese herbal
formula: a randomised, double blind placebo-
Baer H J, Glynn R J, Hu F B, et al 2011 Risk Factors controlled trial. Journal of Alternative and
for mortality in the Nurses’ Health Study: A Complementary Medicine 17(7):639–46
Competing Risks Analysis. American Journal of doi:10.1089/acm.2010.0150
Epidemiology 173(3):319–29 doi:10.1093/aje/
kwq368 Jolly K, Ingram L, Freemantle N, et al 2011 Effect of a
peer support service on breast-feeding continuation
Belanger C F, Hennekens C H, Rosner B, Speizer F E in the UK: a randomised controlled trial.
1978 The nurses’ health study. American Journal of Midwifery. Early view, doi:10.1016/j.
Nursing 78:1039–40 midw.2011.08.005
Bryman A, Cramer D 2005 Quantitative Data Analysis Kitzman H J, Olds D L, Cole R E, et al 2010
with SPSS 12 and 13: A Guide for Social Scientists. Enduring effects of prenatal and infancy home
Routledge, London, UK visiting by nurses on children: follow-up of a
Chapman R, Watkins R, Zappia T, Nicol P, Shields L randomised trial among children at age 12 years.
2011 Nursing and medical students’ attitudes, Archives of Pediatric and Adolescent Medicine
knowledge and beliefs regarding lesbian, gay, 164(5):412–18

182

Schneider_1374_Chapter 9_main.indd 182 7/25/2012 6:11:11 PM

 9 • Common quantitative methods

Lewis D A, Sanders L P, Brockopp D Y 2011 The Sanada H, Nakagami G, Mizokami Y, et al 2010

effect of three nursing interventions on Evaluating the effect of the new incentive system
thermoregulation in low birth weight infants. for high-risk pressure ulcer patients on wound
Neonatal Network: the Journal of Neonatal Nursing healing and cost-effectiveness: A cohort study.
30(3):160–4 International Journal of Nursing Studies 47(3):
Margolis J R, Gillum R F, Feinleib M, Brasch R C, 279–86 doi:10.1016/[Link].2009.08.001
Fabsitz R R 1974 Community surveillance for Scarella A, Chamy V, Sepulveda M, Belizan J 2011
coronary heart disease: the Framingham Cardio- Medical audit using the Ten Group Classification
vascular Disease Survey: methods and preliminary System and its impact on the cesarean section rate.
results. American Journal of Epidemiology 100: European Journal of Obstetrics & Gynecology and
425–36 Reproductive Biology 154(2):136–40 doi:10.1016/
Mater — University Study of Pregnancy 2011 Mater [Link].2010.09.005
— University Study of Pregnancy (MUSP). Online. Shields L, Mamun A, O’Callaghan M, Najman J,
Available: [Link] Williams G, Bor W 2010 Breastfeeding and obesity
[accessed 7 November 2011] at 21 years: a cohort study. Journal of Clinical
McCambridge J, Butor-Bhavsar K, Witton J, Nursing. 19:1612–17
Elbourne D 2011 Can research assessments doi:10.1111/j.13652702.2009.03015.x
themselves cause bias in behaviour change trials? Shields L, Mamun A, Pereira S, O’Nions P, Chaney G
A systematic review of evidence from Solomon 2011 Measuring family centred care: working with
4-group studies. PLoS One 6(10): e26223 children and their parents in a tertiary hospital. The
doi:10.1371/[Link].0025223 International Journal of Person Centered Medicine
McDonald L 2001 Florence Nightingale and the early 1(1):155–60
origins of evidence-based nursing. Evidence Based Shields L, O’Callaghan M, Williams G, Najman J,
Nursing 4:68–69 doi:10.1136/ebn.4.3.68 Bor W 2006 Breastfeeding and obesity at 14 years:
McHugh M D, Shang J, Sloane D M, Aiken L H a cohort study. Journal of Paediatrics and Child
2011 Risk factors for hospital-acquired ‘poor Health 42(5):289–96
glycemic control’: a case–control study. International Shields L, Pratt J, Davis L M, Hunter J 2007 Family-
Journal of Quality in Health Care 23(1):44–51. centred care for children in hospital. Cochrane
doi:10.1093/intqhc/mzq067 Database of Systematic Reviews 2007, Issue 1. Art.
Merchant R M, Abella B S, Abots E J, et al 2010 Cell No: CD004811. DOI: 10.1002/14651858.
phone cardiopulmonary resuscitation: audio CD004811.pub2
instructions when needed by lay rescuers: a Schoenbach V J 1999 Analytic Study Designs. Online.
randomised, controlled trial. Annals of Emergency Available: [Link]
Medicine 55(6):538–43.e1 [Link] [accessed 15 February
Moher D, Schulz K F, Altman D G, for the 2012]
CONSORT Group 2001 The CONSORT Thompson C 2004 Fortuitous phenomena: on
statement: revised recommendations for improving complexity, pragmatic randomised controlled
the quality of reports of parallel-group randomised trials, and knowledge for evidence-based
trials. Lancet 357:1191–4 practice. Worldviews on Evidence-Based Nursing
Patidar A B, Andrews G R, Seth S 2011 Prevalence of 1:9–17
obstructive sleep apnoea, associated risk factors, and Ting-Kai L, Chi-Ming L, Cheng-Jeng, C, Yueng-Ling
quality of life among Indian congestive heart failure L, Chia-Chin L 2011 A retrospective study on the
patients: a cross-sectional survey. Journal of long-term placement of peripherally inserted central
Cardiovascular Nursing 26(6):452–9 catheters and the importance of nursing care and
Pettersson E, Vernby Å, Mölstad S, Lundborg C S education. Cancer Nursing 34(1):E25–E30. doi:
2011 Can a multifaceted educational intervention 10.1097/NCC.0b013e3181f1ad6f
targeting both nurses and physicians change the Wong F K Y, Chow S K Y, Chan T M F 2010
prescribing of antibiotics to nursing home residents? Evaluation of a nurse-led disease management
A cluster randomised controlled trial. Journal of programme for chronic kidney disease: A
Antimicrobial Chemotherapy 66(11):2659–66 randomised controlled trial. International Journal of
doi:10.1093/jac/dkr312 Nursing Studies 47(3):268–78

For further content associated with this chapter visit

[Link]

183

Schneider_1374_Chapter 9_main.indd 183 7/25/2012 6:11:11 PM