Health Care Field Research Draft

22 Volume 22 Number 1 April 2019 JARNA

Understanding the research process Murray J Fisher* RN, DipAppSc, BHSc, MHPEd, ITU Cert, PhD Susan Wakil School of Nursing and Midwifery, Faculty of Medicine and Health, University of Sydney

Sydney, NSW, Australia

Royal Rehab, Sydney, NSW, Australia

Email [email protected]

Jacqueline Bloomfield RN, PhD, MN, PGDip (HealthCare Ed), PGDip (Midwifery), BN Susan Wakil School of Nursing and Midwifery, Faculty of Medicine and Health, University of Sydney

Sydney, NSW, Australia

*Corresponding author

Keywords Research; research process; scientific method; quantitative research; qualitative research.

For referencing Fisher MJ & Bloomfield J. Understanding the research process. JARNA 2019; 22(1):22-27.

DOI https://doi.org/10.33235/jarna.22.1.22-27

Research methods

Introduction

This brief paper is the second in a series of articles aimed at

informing readers of JARNA about research methods, and aims

to explain, in simple terms, the research process. The research

process, also referred to as the scientific method, is presented as

a series of sequential steps that researchers follow when planning

and conducting research and communicating the findings;

however, the steps are often revisited during the process. Most

journals that publish reports of research studies require that each

step of the research process is clearly described. It is, therefore,

important that nurses conducting or reading about research

understand this process. In this article, each step of the research

process is explained.

The research process

The research process has been described as “a series of steps

or stages that the researcher should progress through when

planning and conducting research. Researchers use the research

process to determine how to move from an idea about a problem

in practice to generating research findings that may contribute to

improving that practice” (Gelling, 2015, p. 44).

Most nursing textbooks that focus on research, such as that

edited by Schneider, Whitehead, Lobiondo-Wood, and Harber

(2016), identify 10 distinct steps that make up the research

process. Although these are typically explained in a linear manner,

in reality, the researcher may go back and forth between some

of the steps, until the research study has been refined (Gerrish

& Lathlean, 2015). The 10 steps that make up the research

process include:

1. Identification of the research problem

2. Searching and evaluating the literature

3. Developing a researchable question, hypothesis or aim

4. Identifying an appropriate research design or methodology

5. Addressing ethical considerations

6. Sampling and participant recruitment

7. Data collection

8. Data analysis

9. Findings and drawing conclusions

10. Dissemination of research findings

1. Identification of the research problem

All research starts with the identification of a problem or issue

that needs to be addressed. This is the first step in the research

process. The nature of the problem, and the paradigm through

which the researcher views the problem, will determine the

research question and, subsequently, the type of research and

study design needed to answer that question (Davies & Fisher,

2018). Before conducting a study, researchers need to clearly

articulate a problem statement that identifies the concept or

phenomena of interest (or the variables under investigation),

specify the population being studied and demonstrate the focus

of the study (that is, exploration of an experience, exploration of

variables and relationships, or empirical testing).

23JARNA Volume 22 Number 1 April 2019

2. Searching and evaluating the literature

The second step in the research process requires the researcher

to undertake a literature review. Literature relevant to the

problem or phenomenon of interest should be searched for from

multiple sources and then reviewed with the aim of generating

a comprehensive picture of what is known or not known about

the issue (Coughlan & Cronin, 2017; Hart, 2018). Importantly,

this will allow the researcher to develop an understanding of any

gaps in the existing body of knowledge.

The literature review requires more than simply describing the

content of each study included in the review. Instead, it should

provide a synthesised appraisal of the research undertaken in

the topic area. Where applicable, the literature review should

also highlight the theoretical frameworks that underpin the

research and describe the variables/phenomena and how best

to measure them. The literature review should also articulate the

significance of the problem and justify the need for the research

study (Hart, 2018). In nursing research this may, for example,

involve explaining the nature of the issue and its implications for

patient care.

3. Developing a researchable question, hypothesis or aim

The third step in the research process involves developing a

researchable question, a hypothesis and/or an aim. The nature

of the research question is informed by the paradigm (world

view) from which the researcher views the problem (Davies &

Fisher, 2018). For this reason the research question is key to

everything else in the study as it will inform the study design

and the methods used to undertake the study. As such, it is

essential that the research question is clear. The preciseness of

the question will depend on existing knowledge about the topic

and the specific focus that the researcher wishes to take. This

emphasises the importance of undertaking a comprehensive

literature review. A research question must be “researchable”.

This means that it should be clear and unambiguous, focused,

ethically sound, and that conducting a study to find the answer

to the question is both realistic and feasible. Characteristics of

researchable questions are presented in Table 1. Typically the

research question follows the PICO (for quantitative studies)

or PEO (for qualitative studies) format, where (P) refers to

the population in which the study focuses on, (I) refers to the

intervention or issue of interest, (C) refers to the comparator with

which the intervention is to be tested, and (O) is the outcome

measure. The PEO format, which is used for qualitative research,

refers to Population, Exposure (or the phenomena of interest)

and Outcome (or experience).

Table 1: Characteristics of researchable questions

– Clear

– Unambiguous

– Focused

– Any related concepts should be clearly defined

– Feasible

– Ethically sound

– Not focused on too many issues

4. Identifying an appropriate research design or methodology

The research design (in the case of quantitative research) or

methodology (in the case of qualitative research) describes

the procedures that will be followed to undertake the research.

To use a simple analogy, it is the “recipe” or process that the

researcher will follow to conduct the study, and is referred to

as the method. Identifying an appropriate research design is the

fourth step in the research process and this will depend on two

key issues. These are: the research paradigm and the research

question (Davies & Fisher, 2018).

Quantitative research is predominantly informed by the positivist

or post-positivist paradigms (Davies & Fisher, 2019). This type

of research involves the study of variables that can be quantified

or measured. There are a number of quantitative research

designs, including, for example, randomised controlled trials

(RCTs), quasi experiments such as pre-test/post-tests and non-

experiments such as cross-sectional surveys. All have varying

degrees of control, which in quantitative research, refers to

the actions taken by the researchers to reduce the possibility

of an erroneous finding caused by the influence of extraneous

variables on the outcome variable being measured. Control

in quantitative research can be achieved by using: precise

measurement methods; attaining a representative sample

through probability sampling, and, where appropriate, using

techniques to blind participants and/or researchers to the

assignment of participants to an intervention or control group

and outcome measures. Common quantitative research designs

used in nursing research, in order of hierarchy of control, from

least to highest, includes: descriptive, cohort, cross-sectional,

case control, quasi-experimental and experimental designs

(RCTs).

Qualitative research is predominantly informed by the interpretive

and critical research paradigms. Qualitative methodology not

only determines the methods of the research, but also provides

the philosophical and theoretical positioning of the research.

24 Volume 22 Number 1 April 2019 JARNA

Researchers must make clear how the methods used in the

study are congruent with the chosen methodology. Common

qualitative methodologies used in nursing research include

Phenomenology, Grounded Theory, Ethnography, Narrative, Life

History and Feminist methodologies.

Some research designs use both quantitative and qualitative

methods. These are known as mixed or multi-methods research

designs and, similar to both quantitative and qualitative designs,

also require clear identification of the methods that will be

followed during the conduct of the study. Common mixed

methods designs include convergent parallel design, explanatory

sequential design, embedded design, transformative design and

multiphase design (Creswell & Plano, 2011). Multiple methods

research designs may include case study research, action or

participatory research and evaluation research methods such as

realist evaluation.

5. Addressing ethical considerations

Conducting any type of research study with humans requires

ethical approval from the relevant research ethics committee

(National Health and Medical Research Council, 2018).

Addressing the often many ethical considerations associated

with research represents the fifth step in the research process.

This requires the researchers to consider not only whom they will

recruit to the study as participants, but also how they will recruit

them. Detailed information must be provided to all potential

participants that addresses issues such as: the purpose of the

study, possible benefits and risks, what their involvement will

require, the types of data that will be collected and how this will be

done, how their privacy and confidentiality will be protected, and

how the study findings will be used and disseminated. Upholding

the four main ethical principles of autonomy, beneficence, non-

malificence and justice are crucial to ensure that participants

provide voluntary consent without coercion. Researchers must

also obtain appropriate permission to access the study site for

recruitment and data collection purposes. For example, this

could be a hospital ward or educational institution. It is essential

that ethical approval is obtained prior to researchers advertising

the study, recruiting participants and collecting data.

6. Sampling and participant recruitment

The sixth step in the research process requires decisions to

be made about the sample, including the sample size and how

participants will be recruited to the study. As with previous steps

in the research process, determination of the type and size of the

sample, and participant recruitment will differ according to the

study design or methodology.

The purpose of quantitative research, irrespective of the study

design, is to select a sample from the population, measure

or test the variable/s of interest and use the findings to make

inferences about the larger population. For the inferences to be

accurate and relevant, the study sample must be representative

of the population. Selecting a sample that is representative

of the population requires quantitative researchers to use a

probability (random) sampling technique. This ensures that every

individual in the population has an equal chance (probability) of

selection for inclusion in the study (Fisher & Fethney, 2016a).

Probability sample selection techniques include: simple random,

stratified random; cluster random; and systematic sampling. The

assumption of using a probability sampling technique is that

everyone in the population can be identified. In reality, this is not

always the case and may not be possible. However, if a probability

sampling technique was not used, the researcher is unlikely to be

able to demonstrate that their sample is representative of the

population, and, therefore, the findings cannot be generalised

(Fisher & Fethney, 2016a).

Unlike quantitative research that aims to make generalisations

of the findings from a study to the population, the purpose of

qualitative research is to gain an in-depth understanding of a

phenomenon. This is achieved by collecting narratives from

individuals who have experienced the phenomena and then using

these data to inform the understanding of the phenomena. As

such, in qualitative research, a probability sample is not needed.

Typically, qualitative research uses non-probability sampling

techniques (Fisher & Fethney, 2016a). These include: purposive

sampling, snowball sampling, quota sampling and convenience

sampling. Details about each of these sampling techniques are

presented in Table 2.

Specific sample selection criteria, including both inclusion and

exclusion criteria, should also be reported. The research question

should clearly identify the population of interest and the selection

criteria should be consistent with this. Often the sample is

narrowed by specific criteria that is used to include or exclude a

subset of the population. For example, some studies may exclude

individuals over a particular age. It is important to realise that, if

this is the case, the study results cannot be generalised to these

sub-populations.

Sample size is another very important issue that must be

considered as part of the research process. In quantitative

research, it is vital that the sample is of an adequate size, and that

this is pre-determined prior the start of the study. A general rule

of thumb is to recruit the largest sample size possible; however,

25JARNA Volume 22 Number 1 April 2019

whenever possible a power analysis should be undertaken to

calculate the sample size (Fisher & Fethney, 2016a). This may

be done with the assistance of a statistician and involves a

mathematical calculation. If the sample size is too small the effect

of the independent variable (the variable being studied) on the

dependent variable (outcome variable) may not be detected. This

is known as a Type II error. Alternatively, although less common,

using a sample size that is too large may result in the detection

of an effect that is not really there and inaccurate conclusions are

made. This is described as a “false positive” and is known as a

Type I error (Fisher & Fethney, 2016b).

In qualitative studies, the sample size is typically much smaller

and is not predetermined before the study commences. Usually

in qualitative research, the researcher will continue to recruit

participants until data saturation has occurred. This is when the

researcher decides that no new insights about the phenomenon

are to be gained through further sampling.

7. Data collection

Data collection is the seventh step of the research process and

all research should include a comprehensive description of the

data collection methods. Whilst many methods of data collection

can be used in quantitative and qualitative research, it is how the

data are derived that distinguishes the difference. In quantitative

research the data will either be in numerical form when collected,

such as in measures, scales or frequencies, or will be able to

be transformed into numerical values. In qualitative research the

data is usually in narrative (story) or word form.

Data in quantitative research may consist of direct clinical

measures, surveys and scales. Regardless of the data collection

method used, the researcher must consider the validity and

reliability of the measure. In simple terms, this refers to the

accuracy of the measure and its consistency. When deciding

on quantitative data collection tools, the researcher needs to

consider the level of measurement and the validity and reliability

of the measures. The level of measurement will determine the

type of statistics used to determine difference between groups

or association between variables. This will be further discussed

under data analysis.

Data collection methods used in qualitative research often

include participant interviews and observations, documents,

photographs and art work. A clear description of how, where and

when the data were collected should be included. The researcher

should provide a detailed explanation of the processes used

to collect the data so that a decision trail can be followed.

When conducting interviews, the researcher should provide a

description of the interview process, including where and how it

was conducted and the questions that were asked to guide the

interview. In the case of participant observation, the researcher

should describe the purposes for which observation is used, the

stances or roles of the observer, and information regarding the

when, what, and how to observe (Kawulich, 2005).

8. Data analysis

The eighth step in the research process is data analysis. In

quantitative research the methods used to analyse the data will

consist of descriptive and inferential statistics. The appropriate

use of the statistic is determined by the level at which a variable

is measured and the degree to which particular assumptions are

met (Fisher & Fethney, 2016b). Variables may be measured at

Table 2: Sampling techniques

Non-probability sampling technique

Description

Purposive Sampling technique whereby individuals who are known to have had the specific experience or characteristic of interest are selected

Snowball Sampling technique whereby recruited participants identify others who may be suitable for selection to the study

Quota Sampling technique whereby individuals with specific characteristics are recruited to a study in proportion to the presence of these in the population of interest

Convenience Sampling technique whereby the sample is selected on the basis of convenience of accessibility to the researcher

Probability sampling technique

Simple Sampling technique whereby each member of the population has an equal and independent chance of being selected. A simple method such as use of a random number generator is used

Stratified Sampling technique whereby the population is divided into sub-groups (strata) and a number of participants are randomly selected on the basis of their proportion in the strata

Cluster Sampling technique whereby the population is divided into clusters and a random number of clusters is selected for the sample

Systematic Sampling technique whereby, starting from the first individual every kth person is selected for the sample

For example, selecting every 10th individual from a population of 3000 to select a total of 300.

26 Volume 22 Number 1 April 2019 JARNA

nominal (categorical), ordinal (ordered categorical) and interval/

ratio (continuous scale) levels. Examples of nominal level data

include: religion, gender, and country of birth. In nominal level

data the categories must be exhaustive, that is, there must

be enough categories for each individual to identify with one,

and the categories must be exclusive, that is, individuals must

only identify with one category. Ordinal level data is defined as

being ordered or hierarchical categories. Examples of ordinal

level data include the Likert scale, pain scales and the Glasgow

Coma Scale. Like nominal level, the ordinal categories must

be exhaustive and exclusive. Interval and ratio level data are

measured on a continuous scale where the points on the scale

are equal distance. The difference between interval and ratio

is that interval level data does not have an absolute zero. For

example, in relation to temperature, which is interval level data,

a measurement of zero does not mean there is no temperature.

Examples of ratio data includes weight in grams, height in

centimetres, and pressure in mmHg. Ratio data has a meaningful

zero, that is, a zero score means non-existence.

Descriptive statistics simply describe the sample for a given

measure and includes measures of central tendency (the midpoint

of a distribution of scores) and measures of dispersion (the spread

of the scores in a distribution). Measures of central tendency

include the mode (most frequent score), median (middle point in

an ordered distribution of scores) and the mean (mathematical

average). Measures of dispersion includes the frequency of each

score, range and standard deviation (the average distance each

score falls around the mean). The appropriate use of descriptive

statistics is dependent on the level of measurement, for example,

it is not possible to calculate an average of categories therefore

for nominal level data measures such as mode and frequencies

would be used to describe the sample. For ordinal level data,

median and range would be used to describe the sample. For

interval and ratio data, the mean and standard deviation are used

to describe the sample if the sample is normally distributed.

Inferential statistics are used to test difference between samples

or relationships (association) between variables. The specific

type of statistical test used will depend on what the researcher

wants to find out (difference versus association), the level of

measurement, the number of samples or variables and whether

specific assumptions for a test are met (Fisher & Fethney,

2016b). Statistical procedures test the probability that samples

belong to the same population and therefore are the same, or

the probability variables are associated (correlated) (Fisher

& Fethney, 2016b). Inferential statistics attempt to quantify

the degree to which chance is accounted for in the results by

calculating a probability (p value).The lower the p value, the less

likely the observed result is due to chance alone.

Methods of qualitative data analysis include a variety of

techniques, and the method selected will depend on the research

methodology, the type of data that have been collected and the

question that the researcher is asking. Data analysis in qualitative

research begins as soon as data collection commences.

Qualitative researchers continuously move between data

collection and analysis in a cycle of questioning the data and

verifying interpretation. Some examples of the most common

methods of qualitative data analysis include: thematic analysis,

content analysis, constant comparative analysis, framework

analysis, discourse analysis and grounded theory approach. A

brief description of each of these are presented in Table 3.

Table 3: Qualitative data analysis methods

Qualitative data analysis method

Description

Thematic analysis Analysing for themes, patterns and common threads within data

Content analysis Analysing for categories, constructs and domains; researchers can quantify and analyse the presence, meanings and relationships of such categories

Constant comparative analysis

Analysis and comparison of a piece of data with the other pieces of data in order to develop conceptualisations of the possible relations between various pieces of data

Framework analysis Analysis that involves identifying a framework and categorising and mapping data to this

Discourse analysis Analysis of language (the systematic account of structures, strategies and processes of text) to reveal the socio-psychological characteristics of individuals

Grounded Theory Analysis that involves fracturing the data (open coding) and then gluing the data into concepts or categories (Theoretical coding). Axial coding may also be used to analyse relationships between categories and subcategories to inform properties and dimensions.

9. Findings and drawing conclusions

The ninth step in the research process is reporting the study

findings and drawing conclusions. Study finding should be

reported in a way that is clear, concise and consistent with the

research question, methodology and data analysis procedures

that were undertaken. Importantly, the findings and conclusions

should directly align with the research question or study

hypotheses.

27JARNA Volume 22 Number 1 April 2019

As part of the findings, the characteristics of the sample should

be reported, including sample size, response rate (in the case

of survey research), and a clear description of the demographic

characteristics of the sample. In research that compares different

samples, a description of each sample should be reported with

statistics comparing characteristics between the groups. This

information is often presented in tables.

The aims of quantitative research is to either describe a sample

for specified characteristics or to test the difference between

samples for a test variable or determine relationships between

variables. The research question or study hypotheses determines

what is reported in the findings. In quantitative research, the

researcher should report sufficient and appropriate descriptive

and inferential statistics.

Like quantitative research, in qualitative research the findings must

directly relate to the research question. In qualitative research,

the findings report the researcher’s interpretations of the data

and should include sufficient raw data as exemplars to allow

the reader to establish that the conclusions and interpretations

arose from participants. The way findings are presented will be

determined by the qualitative research methodology but are

commonly presented in themes.

Regardless of the type of research, the conclusions drawn by

the researcher should directly relate to the research question

and must be consistent with the findings. The researcher should

report the strengths and weaknesses of their study and provide

an explanation on how these impact on the generalisability or

transferability of the findings. In many instances, it is left up to

the reader of the research to critically appraise the research and

determine what conclusions should/could be made from the

research.

10. Dissemination of research findings

Dissemination of the study findings is a crucial step in the

research process, and represents the 10th and final step. There

are many ways that research findings can be shared. The more

common methods of disseminating nursing research include

research reports, journal articles, conference papers, and

books. As a reader of research, it is essential that nurses are

able to determine the authority of the publication. In the case

of journal articles, most manuscripts go through a peer-review

process, whereby it is scruntinised by other experts in the same

field to determines its accuracy and suitability for publication.

Typically, two expert reviewers are involved in the process.

Similarly, conference abstracts are often peer reviewed prior to

acceptance for presentation.

It is useful to be aware that, with regard to quantitative research,

there is generally a publication bias towards studies that report

a statistically significant result. In many cases, inadequately

powered studies in which findings do not support the research

question or are not statistically significant, are not published

(Koretz, 2019). Importantly, this can influence the nature of

evidence used to support changes in practice (Koretz, 2019).

Conclusion

This brief paper has described the research process, which can

be likened to a framework to help guide the researcher through

what may otherwise be a long and complicated process. The

research process is comprised of 10 key steps, and adequate

consideration and attention to each of these is essential in

order to ensure that the research study is undertaken with

rigour. Although in reality, many nurses will not be conducting

research studies, it is essential that the research process is

fully understood. This knowledge will enhance nurses’ research

literacy and the ability to understand and determine the quality

of published studies, in view of potentially translating research

findings into their nursing practice.

References Coughlan, M., & Cronin, P. (2017). Doing a literature review in nursing,

health and social care (2nd ed.). London: Sage.

Creswell, J., & Plano Clark, V. L. (2011), Designing & Conducting Mixed Methods Research (2nd ed.). Thousand Oaks: Sage.

Davies, C., & Fisher, M.J. (2018). Understanding research paradigms. Journal of the Australasian Rehabilitation Nurses’ Association, 21(3), 21–25.

Fisher, M. J., & Fethney, J. (2016a). Sampling in Quantitative Research. In Schneider, Z., Whitehead, D., Lobiondo-Wood, G., & Harber, J. (Ed), Nursing and Midwifery Research: Methods, Critical Appraisal and Utilisation (5th ed.). Sydney: Mosby.

Fisher, M. J., & Fethney, J. (2016b). Analysing Data in Quantitative Research. In Schneider, Z., Whitehead, D., Lobiondo-Wood, G., & Harber, J. (Ed), Nursing and Midwifery Research: Methods, Critical Appraisal and Utilisation (5th ed.). Sydney: Mosby.

Gelling, L. (2015). Stages in the research process. Nursing Standard, 29(27), 44–49.

Gerrish, K., & Lathlean, J. (2015), The research process in nursing (7th ed.). Oxford: Wiley-Blackwell.

Hart, C. (2018). Doing a literature review: Releasing the research imagination (2nd ed.). London: Sage.

Koretz, R. L. (2019). Assessing the evidence in evidence-based medicine. Nutrition in Clinical Practice, 34(1), 60–72.

National Health and Medical Research Council. (2018). National Statement on Ethical Conduct in Human Research 2007(Updated 2018). The National Health and Medical Research Council, the Australian Research Council and Universities Australia. Commonwealth of Australia, Canberra.

Copyright of Journal of the Australasian Rehabilitation Nurses' Association (JARNA) is the property of Australasian Rehabilitation Nurses' Association (ARNA) Inc. and its content may not be copied or emailed to multiple sites or posted to a listserv without the copyright holder's express written permission. However, users may print, download, or email articles for individual use.