SPOTLIGHT ON RESEARCH

KEYWORDS Research / Quantitative research / Reliability / Rigour

Provenance and Peer review: Unsolicited contribution: Peer reviewed; Accepted for publication April 2015.

Part three: The

quantitative approach
by J Quick and S Hall
Correspondence address: J Quick, Seacole Building, Birmingham City University, Westbourne Rd, Edgbaston, Birmingham, B15 3TN. Email: [email protected]

The previous articles in this series have explored the need for perioperative practitioners to
select a research approach that supports the aims of the study to be undertaken (Quick &
Hall 2015a,b). The quantitative approach is considered by some authors to be ‘the epitome of
the scientific approach’ (Meadows 2003) and this article, the third in a series, aims to give an
overview of this model.

Introduction sinus disease. This work may be either possible, on high quality primary research
retrospective, where data is collected from sources such as randomised controlled
Whereas qualitative methodology allows episodes of care or treatment that have trials and is therefore ranked highest
for the experiences of small numbers already been delivered prior to the formal in the evidence hierarchy (Figure 1). A
of subjects to be explored, quantitative establishment of a study, or prospective systematic review enables the researcher
research is grounded in the positivist where data is collated at the time of to pool numerical data from other studies
position - the concept that scientific treatment (Reeves 2008). and synthesise new conclusions as to the
truths exist and can be studied (Topping effectiveness of a particular treatment
2010). The principles of this model Quantitative researchers use a variety
(Hoe & Hoare 2012). One of the best
support the concept that data is capable of methods within these approaches.
known series of systematic reviews is that
of being analysed statistically (Research Systematic reviews are defined in
published by The Cochrane Collaborative
Methodology 2014). Researchers using Gerrish and Lacey (2010) as ‘a rigorous
(Nightingale 2009).
positivist concepts focus on the collection and systematic literature search using a
and evaluation of numerical data. Unlike well-defined question and strict criteria The randomised controlled trial (RCT) is
qualitative theories the quantitative for study inclusion and evaluation’. They acknowledged as the ‘gold standard’ of
researcher distances themselves from enable the researcher to present an research tools (Wright 2009). This model
research participants, the argument being overview of research previously published is used to examine the effectiveness of an
that this encourages objectivity. on a particular subject. It draws, whenever intervention which compares two groups

Quantitative approaches
BIAS
Quantitative research is defined by Gerrish Selection bias Participants of a trial are significantly different from those not selected. This can be caused by:
and Lacey (2010) as: ‘the broad term used
• participants being lost to follow up
to denote research designs and methods
• high numbers of participants failing to respond to elements of the study
that yield numerical data’. Meadows (2003)
discussed the two primary research designs • poor sample choice
for those planning to use the quantitative • inappropriate choice of comparison groups.
model. The experimental design usually Assessment or observer bias Knowledge of the treatment may influence the researcher’s interpretation
involves the introduction and manipulation of data
of a variable, for example, the effects of a Information bias Resulting from incorrectly recorded measurements
particular treatment versus non-treatment.
Attrition bias Due to differences in the groups being compared because of a loss of participants
A perioperative example of this might be
assessment of the costs of laparoscopic Performance bias Some participants receive extra attention, which influences their behaviour/compliance/
outcome
suturing of the cystic duct and artery
versus clip occlusion of these vessels in a The so-called ‘placebo’ or ‘Hawthorne’ effect of increased attention being given to research participants
laparoscopic cholecystectomy. was identified in the 1920s. Mayo (1933) commented that ‘the sympathy and interest of the observers’ in
the Hawthorne experiments made employees in the study feel better (Draper 2008).
The non-experimental approach uses
Publication bias is a significant factor: Articles which demonstrate results that lack statistical significance
data gathered from pre-existing patient or may be less likely to be published.
societal groups and allows the researcher
to explore relationships between variables, Petrie and Sabin (2005), Humadi and Stanfield (2006), Nelson et al ( 2010), Hoare and Hoe (2013).
for example regional wound healing rates
in patients undergoing surgery for pilonidal Box 1 Potential causes of bias in quantitative studies

192 October 2015 / Volume 25 / Issue 10 / ISBN 1750-4589

JPP Oct 2015.indd 192 17/09/2015 16:12

 SPOTLIGHT ON RESEARCH

Cluster sampling takes advantage of groups that already exist e.g.

cohorts of students, or patients attending a specific outpatient
clinic where registers can be used to generate the sample

of individuals, one undergoing treatment because they can last for many years, Principles of
and the other acting as a ‘control’ i.e. participants may leave the study for a quantitative research
not receiving treatment, or undergoing a variety of reasons. This has an impact on
different therapeutic regime. findings (Nelson et al 2010). Sampling
Gerrish and Lacey (2010) consider that In case control studies, participants with The applicability of quantitative study
elements outside the parameters of a the condition of interest are matched with a results to the wider community is supported
research project may constitute bias. This group who do not have it. Efforts are made by probability sampling, where each
can take many forms (see Box 1) and can to ensure that the two groups have similar individual in a target population has an
influence the findings of any research study. characteristics and that the control group equal chance of selection. It is however
RCTs use a variety of strategies to reduce members are free of the disease being dependent on the availability of an up-to-
bias. For example, by randomising the studied (Petrie & Sabin 2005).The data date database from which the sample can
groups to which participants are allocated collected on the two groups is compared be drawn. An example of this would be
researchers are able to reduce the impact so that researchers can explore the impact an electoral roll (Procter et al 2010). This
of selection bias. This helps to ensure the of difference on the condition under study method allows researchers to calculate,
validity of results (Nelson et al 2010). There (Hoe & Hoare 2012). and allow for, any bias inherent in the
are various techniques for randomising study (Procter et al 2010). Simple random
Cross sectional studies are used to collect sampling allows each member of the
participants, the simplest being a coin toss,
data on individuals at a single point in time potential study group an equal chance of
while the most sophisticated use computer
– for example breast screening imaging or being chosen. Stratified random sampling
programmes (Humadi & Stanfield 2006).
self-administered colon cancer detection sees the target group subdivided using pre-
Bias resulting from the drop out of kits. Whilst useful, this method does not established criteria with random samples
participants can be avoided by analysing allow for the evaluation of causative factors taken from each of the subgroups.
the data of all subjects. This is known of the disease in the population (Petrie &
as ‘intention to treat’ (ITT). Scrutiny of Sabin 2005, Hoe & Hoare 2012). An alternative method is systematic
information recorded only from those sampling, where the numbers of the target
Although low in the hierarchy of research group are divided by the number of the
patients who both experienced the
methodologies, case series and case sample desired.
intervention and subsequently followed the
studies are popular with healthcare
post intervention protocol is known as ‘per e.g. target group = 1000, required sample = 200
professionals since they allow the
protocol’ analysis (Nelson et al 2010). 1000
researcher to place the work within the
200 = 5 therefore,every fifth member of the
A strategy used to avoid the impact of clinical context whilst drawing on their target group is selected
measurement bias is termed ‘blinding’. own experience. This type of study may be (Meadows 2003).
The intention of this strategy is to limit either qualitative or quantitative (Clarke &
the awareness of participants, healthcare Reed 2010) and is particularly of use when Cluster sampling takes advantage of groups
professionals and others associated with there is little or no published evidence on a that already exist e.g. cohorts of students,
the study as to which group the individual chosen topic (Hoe &Hoare 2012). or patients attending a specific outpatient
has been allocated to. This approach is clinic where registers can be used to
useful when dealing with studies that generate the sample.
contain a subjective element. Blinding may Whilst researchers may preferentially
be ‘single’ or ‘double’. In single blinding, use probability sampling in quantitative
Systematic
it is usually the patient who is unaware of research, Procter et al (2010) stated that
review
which group they have been allocated to; non-probability sampling, particularly where
in double blinded studies, both the patient Randomised there are difficulties in identifying the study
and their clinical carers are ignorant of their controlled trial population, is popular in nursing research.
group (Nelson et al 2010). Convenience sampling uses a cohort of
Cohort studies are observational and Cohort studies potential participants who are readily
explore the progress of a group of available to the researcher. In snowball
participants with a shared disease or sampling, subjects are encouraged to
situation over time. The collection recruit others known to them who meet
Case control studies
of data takes place at pre- the inclusion criteria established by
determined intervals during the research team. This is a useful
the life of the study (Hoe & strategy in groups which are
Cross sectional studies difficult to identify or who may
Hoare 2012). These studies
may be retrospective or engage in illegal activities
prospective. One of (Meadows 2003). Division
Case studies / case series of the target group
the difficulties with
studies of this into subgroups
type is that, Figure 1 The evidence hierarchy (adapted from Gerrish & Lacey 2010) along the lines

October 2015 / Volume 25 / Issue 10 / ISBN 1750-4589 193

JPP Oct 2015.indd 193 17/09/2015 16:12

 SPOTLIGHT ON RESEARCH

Part three: The quantitative approach

Continued

of age or gender using known estimates

of local or national figures allows the Formalisation of surgical skills assessment for SCPs/SAs
researcher using quota sampling to select a Please select one answer for each statement:
percentage of the subgroup for evaluation
(Meadows 2003). Q. Formal assessment of surgical skills should be a regular occurrence
Strongly agree Agree Neither agree nor disagree Disagree Strongly disagree
Quantitative studies often sample groups
far larger than those used in qualitative
projects, however, a decision on the size Box 2 An example from a questionnaire used in a quantitative study (from Hall 2011)
of the sample to be surveyed is important
when planning a quantitative research
project as this may have an impact on the access (Dillman et al 2009). It is recognised Scales may be used for subjective questions
duration of the project, costs associated that respondents who identify with the in questionnaires as they allow respondents
with the study and also affect the ability research topic are more likely to respond to to express opinions that can be quantified
of the researcher to detect a difference of questionnaires (Fowler 2009). by the researcher (Box 2). The criteria
statistical significance between the study established for the popular Likert scaling
Questionnaire design
groups. In order to help the investigator are as follows:
Questionnaire structure should follow
decide the number of participants • The scale contains several items
established guidelines, with early sections
necessary to ensure an appropriate sample
gathering information using simple closed • Response levels are arranged
size, tables are available in a variety
questions to enable respondents to engage horizontally
of formats (Meadows 2003, Procter et
with the format (Jackson & Furnham 2000). • Response levels are anchored with
al 2010).
Closed questions often require a ‘yes/ more-or- less evenly spaced graduations
Data collection in quantitative studies no/don’t know’ answer, and this can then
• Verbal labels are bivalent and
The data collected in quantitative research be entered numerically. Further closed
symmetrical about a neutral middle
is numerical (Meadows 2003), with a questions then collect information about the
topic of interest. If necessary these can be • The scale always measures attitude
variety of methods used to acquire this
followed by an open, subjective component, in terms of level of agreement/
information. Writing in 2000, Wilson
using a linear rating scale to capture data disagreement about a target statement
argued that the collection of data, inherent
(Box 2). (Uebersax 2006).
in the quantitative model may, for some
investigators, become a substitute for It has been argued by some authors that The structure of the question format using
consideration of the research problem. If the wording of questions has an impact a five or seven point choice of responses
you are planning to use this approach it is on response rate (Comley 2006 cited allows the expression of ‘degrees of
therefore important to keep your focus on by Dillman et al 2009). The concept opinion’ by the respondent whilst enabling
the research aim and to select the most underpinning this was formulated by Eric the researcher to ‘measure attitude’ (Jones
appropriate research collection tool to Berne (1920-1970) during his development & Rattray 2010). This method of recording
achieve this. Common tools are surveys of the theory of transactional analysis. attitudes may be used in other research
and questionnaires where the participant is Berne’s argument is that individuals occupy paradigms. Some researchers use ‘Likert
encouraged to express an opinion. Tests are one of three positions within relationships: like’ formats which, rather than provide the
also used; however, they focus on assessing that of ‘parent’, ‘adult’ or ‘child’. In order respondent with the opportunity to express
some aspect of the individual, whilst scales to ensure equality in the researcher/ their opinion, seek to ‘capture feelings,
can evaluate the subject against a pre- participant relationship it is important to actions and pragmatic opinion…. to
determined standard (Meadows 2003). give careful consideration to the way in ascertain the participant’s collective degree
Other methods that may be used are: which questions are worded. of agreement around an issue’ (Joshi et
analysis of data, structured observation and al 2015).
classical experimental design. This latter In contrast to qualitative researchers
who use open-ended questions to collate Regardless of the method used to collect
approach enables the investigator to test a
data that is rich in description, Jackson data, it must be recorded in such a way
hypothesis or research question (Nelson et
and Furnham (2000) consider that as to allow the information gained to
al 2010).
this technique in quantitative research be transferred onto a database, usually
Questionnaires as a data collection tool demonstrates poor insight into the review electronic. Walters and Freeman (2010)
Questionnaires are a popular tool topic. This is supported by other workers discuss the use of two readily available
for collecting data, with electronic in the field, who deem closed questions statistical packages: Excel™ and the
questionnaires affording a modern format in a self-administered survey to be more Statistical Package for the IBM Social
that allows respondents equal access and effective (Fowler 2009). Upon completion Sciences (IBM SPSS™), considering that the
anonymity (Jackson & Furnham 2000). of a draft questionnaire its effectiveness former is suitable for basic analyses and
It is of particular use when studying should, whenever possible, be evaluated the latter useful in situations where more
professional groups who have computer using a small test group (Dillman et ‘flexibility’ is required.
al 2009).

194 October 2015 / Volume 25 / Issue 10 / ISBN 1750-4589

JPP Oct 2015.indd 194 17/09/2015 16:12

 SPOTLIGHT ON RESEARCH

In all quantitative methodologies, it is important that researchers collect and

report data in a consistent and meaningful way, without demonstrating bias

Reeves (2008) argues that for retrospective

studies where patient care is delivered and
data collected prior to the establishment of
a study, consent can be difficult to obtain
from potential ‘participants’. He considers
that this should not prevent researchers
from writing a protocol, which takes account
of this fact.
Rigour
The words ‘reliability’ and ‘validity’ are
associated with quantitative research
(Lacey 2010). These words support the
concept of rigour, which can be defined as
‘the strength of a research design in terms
of adherence to procedures, accuracy
and consistency’ (Gerrish & Lacey 2010)
in ensuring that the research process has
been correctly conducted.
An initiative to improve the quality and
Figure 2 Example of a bar chart used in quantitative research (from Hall 2011)The use of tables may also
understanding of RCTs was instigated by
make it easier for the reader to interpret results (Hoare & Hoe 2013) as demonstrated below (Table 1) in a group of clinical trial methodologists,
this analysis of the question demonstrated in Figure 2 above. The number of respondents is commonly guideline developers, knowledge translation
denoted by the use of ‘n’. experts and journal editors in 1993;
with the most recent policy statement
Strongly Agree Total percentage in Neither agree Disagree Strongly published in 2010 (CONSORT Group 2010).
agree agreement nor disagree disagree
The document, known as the CONSORT
32.43% 45.95% 78.38% (n=29) 18.91% 2.7% 0% 2010 statement contains a checklist and
flowchart which aim to support investigators
Table 1 Attitudes to regular formal skills assessment (Hall 2011) in their reporting of how they have
conducted their RCT.
Presenting the results comparisons are being made between In all quantitative methodologies, it is
A significant part of the research process treatments, is the research proposal. This important that researchers collect and
is the presentation of the results (Holloway needs to be submitted to the necessary report data in a consistent and meaningful
& Wheeler 2010). Formats will vary committees and approval granted by an way, without demonstrating bias (Lacey
depending upon institutional requirements in-house research ethics committee. The 2010). This ensures that research findings
but whatever the design the results from Health Research Authority (2013) sets the are credible and can be applied to the wider
quantitative studies are numerically parameters of what constitutes research population.
represented. Graphs, histograms and pie requiring approval by a national body.
Conclusion
charts are popular methods of presenting In a discussion of consent and research
results, particularly for the novice participants, Dimond (2011) stressed the This third article in the series has provided
researcher, as computer software will need for potential participants to be fully an overview of the positivist approach.
generate the necessary display (Figure 2). informed about the nature of the study. Quantitative research projects should be
Patients with capacity should be given carefully planned. Attention to detail in the
Ethical principles
the opportunity to consent after they have data collection stage should be meticulous,
The main principles of ethics were ensuring that bias is avoided. To avoid
been provided with information regarding
discussed in the first article in this series mistaken emphasis on the collection of
the risks and benefits of the proposed
(Quick & Hall 2015a) and aim to protect data, the focus of a quantitative approach
intervention. Consent that is not given
the dignity, safety, rights and interests of should remain on the research aim. The
freely has no validity. Research involving
individuals involved in research studies. next and final article in the series Spotlight
minors should only be carried out if it is in
There are however additional moral on Research advises on the preparation
the child’s best interest (Herring 2012).The
considerations when undertaking a and presentation of a research dissertation.
Mental Capacity Act (2005) cited by Herring
quantitative study.
(2012) does not exclude research involving
Vital to the methodology of any research individuals who lack capacity however, it
project, but particularly so when does provide protection for those individuals
who are entered into trials.

October 2015 / Volume 25 / Issue 10 / ISBN 1750-4589 195

JPP Oct 2015.indd 195 17/09/2015 16:12

 CLINICAL FEATURE

Part three: The quantitative approach

Continued

References Joshi A, Kale S, Chandel S, Pal D 2015 Likert scale:

explored and explained British Journal of Applied
Science and Technology 7 (4) 396-403
About the authors
Clarke C, Reed J 2010 Case study research In: Gerrish
K, Lacey A (eds) The research process in nursing Lacey A 2010 The research process In: Gerrish K, Lacey Julie Quick
Oxford, Wiley-Blackwell 237-47 A (eds). The research process in nursing Oxford, MSc, NMP, SCP, RGN
Comley P 2006 The games we play: a psychoanalysis Wiley-Blackwell 13-26
of the relationships between panel owners and panel Senior Lecturer, Department of Allied and Public
Mayo E 1933 The human problems of an industrial
participants Amsterdam, ESOMAR 123-32 Health, Birmingham City University and Surgical Care
civilisation New York, MacMillan
Practitioner, Walsall Healthcare NHS Trust
CONSORT Group 2010 Consolidated standards of Meadows K 2003 So you want to do research? 4: An
reporting trials 2010 statement [online] Available introduction to quantitative methods British Journal Susan Hall
from: www.consort-statement.org/consort-2010 of Community Nursing 8 (11) 519-26 MSc (Clinical Sciences), RGN
[Accessed April 2015]
Nelson A, Dumville J, Torgerson D 2010 Experimental
Dillman D, Smyth J, Christian L 2009 Internet mail and research In: Gerrish K, Lacey A (eds) The research Senior Lecturer in Surgical Care Practice, Faculty of
mixed mode surveys. The tailored design method process in nursing Oxford, Wiley-Blackwell 199-215 Medical Science, Anglia Ruskin University / Retired
Hoboken, John Wiley and Sons Surgical Care Practitioner
Nightingale A 2009 A guide to systematic literature
Dimond B 2011 Legal aspects of nursing Harlow, reviews Surgery 27 (9) 381-84
Pearson Education Ltd No competing interests declared
Petrie A, Sabin C 2005 Medical statistics at a glance
Draper S 2008 The hawthorne, pygmalion, placebo Oxford, Blackwell
and other effects of expectation: some notes [online] Members can search all issues of the BJPN/JPP
Available from: www.psy.gla.ac.uk/~steve/hawth Procter S, Allan T, Lacey A 2010 Sampling In: Gerrish K, published since 1998 and download articles free of
[Accessed April 2015] Lacey A (eds) The research process in nursing Oxford, charge at www.afpp.org.uk.
Wiley-Blackwell 143-52 Access is also available to non-members who pay a small
Fowler F 2009 Survey research methods Thousand
Quick J, Hall S 2015a An introduction to the research fee for each article download.
Oaks(California), Sage Publications Inc
process Journal of Perioperative Practice 25 (4) 78-81
Gerrish K, Lacey A 2010 Glossary In: Gerrish K, Lacey A
(eds) The research process in nursing Oxford, Wiley- Quick J, Hall S 2015b Qualitative research Journal of
Blackwell p533 Perioperative Practice 25 (7&8) 129-133
Hall S 2011 Steadfast hands with experience Reeves B 2008 Principles of research: limitations of non
Unpublished MSc thesis randomised studies Surgery 26 120-24
Health Research Authority 2013 Research governance Research Methodology 2014 Positivism [online]
frameworks [online] Available from: http://www.hra. Available from: http://research-methodology.net/
nhs.uk/resources/research-legislation-and-governance/ research-philosophy/positivism/ [Accessed August
research-governance-frameworks/ [Accessed April 2015]
2015] Topping A 2010 The quantitative-qualitative continuum
Herring J 2012 Research in Medical Law and Ethics In: Gerrish K, LaceyA (eds). The research process in
pp598-632 Oxford, Oxford University Press nursing Oxford, Wiley-Blackwell 129-41
Hoare Z, Hoe J 2013 Understanding quantitative Uebersax J 2006 Likert scales: dispelling the confusion
research: part 2 Nursing Standard 27 (18) 48-55 [online] Available from: http://john-uebersax.com/
stat/likert.htm [Accessed April 2015]
Hoe J, Hoare Z 2012 Understanding quantitative
research Nursing Standard 27 (15-17) 52-57 Walters S, Freeman J 2010 Description of analysis of
quantitative data In: Gerrish K, Lacey A (eds) The
Holloway I, Wheeler S 2010 Qualitative research in research process in nursing Oxford, Wiley-Blackwell
nursing and healthcare West Sussex, Wiley-Blackwell 437-53
Humadi S, Stanfield N 2006 Statistics and computing in Wilson T 2000 Recent trends in user studies: action
surgery Surgery 24 (8) 279-88 research and qualitative methods [online] Available
Jackson CJ, Furnham A 2000 An overview of the from: www.informationr.net/ir/5-3/paper76.html
design In: Designing and analysing questionnaires [Accessed April 2015]
and surveys: a manual for health professionals and Wright J 2009 Non randomised and randomised trials
administrators London, Whurr Publishers Ltd 65-72 Surgery 27 371-74
Jones M, Rattray J 2010 Questionnaire design In: The
research process in nursing Gerrish K, Lacey A (eds)
Oxford, Wiley-Blackwell 369-80

196 October 2015 / Volume 25 / Issue 10 / ISBN 1750-4589

JPP Oct 2015.indd 196 17/09/2015 16:12

 Disclaimer The views expressed in articles published by the
Association for Perioperative Practice are those of the writers and do not
necessarily reflect the policy, opinions or beliefs of AfPP.

Manuscripts submitted to the editor for consideration must be the

original work of the author(s).

© 2015 The Association for Perioperative Practice

All legal and moral rights reserved.

The Association for Perioperative Practice

Daisy Ayris House
42 Freemans Way
Harrogate
HG3 1DH
United Kingdom

Email: [email protected]
Telephone: 01423 881300
Fax: 01423 880997
www.afpp.org.uk