Qualitative data analysis: a practical example

Helen Noble,1 Joanna Smith2

The aim of this paper is to equip readers with an under-standing of the principles of qualitative data analysisand offer a practical example of how analysis might beundertaken in an interview-based study.

What is qualitative data analysis?Qualitative research is a generic term that refers to agroup of methods, and ways of collecting and analysingdata that are interpretative or explanatory in nature andfocus on meaning. Data collection is undertaken in thenatural setting, such as a clinic, hospital or a partici-pants home because qualitative methods seek todescribe, explore and understand phenomena from theperspective of the individual or group. Reality is cocon-structed by the research participants and the researcher,with the depth of data collected more important thanrecruiting large samples. The individual interviewmethod is the most widely used method of data collec-tion in qualitative research and a range of data can becollected including field notes, audio and video record-ings, images or documents. Qualitative researchersusually work with text when analysing data; data canbe transcribed in entirety or focus on selected sections.However, focusing on selected sections of the data maynot capture the nuances of observations or participantsdescriptions and may fragment the data. The challengefor qualitative researchers is to present a cohesive repre-sentation of the data, which can be vast and messy,1

and needs to make sense of diverse viewpoints orcomplex issues. The process of data analysis is to assem-ble or reconstruct the data in a meaningful or compre-hensible fashion, in a way that is transparent, rigorousand thorough, while remaining true to participantsaccounts.

What are the approaches in undertakingqualitative data analysis?Although qualitative data analysis is inductive andfocuses on meaning, approaches in analysing data arediverse with different purposes and ontological (con-cerned with the nature of being) and epistemological(knowledge and understanding) underpinnings.2

Identifying an appropriate approach in analysing

qualitative data analysis to meet the aim of a study canbe challenging. One way to understand qualitative dataanalysis is to consider the processes involved.3

Approaches can be divided into four broad groups: qua-sistatistical approaches such as content analysis; the useof frameworks or matrices such as a frameworkapproach and thematic analysis; interpretativeapproaches that include interpretative phenomenologicalanalysis and grounded theory; and sociolinguisticapproaches such as discourse analysis and conversationanalysis. However, there are commonalities acrossapproaches. Data analysis is an interactive process,where data are systematically searched and analysed inorder to provide an illuminating description of phenom-ena; for example, the experience of carers supportingdying patients with renal disease4 or student nursesexperiences following assignment referral.5 Data ana-lysis is an iterative or recurring process, essential to thecreativity of the analysis, development of ideas, clarify-ing meaning and the reworking of concepts as newinsights emerge or are identified in the data.

Do you need data software packages whenanalysing qualitative data?Qualitative data software packages are not a prerequisitefor undertaking qualitative analysis but a range of pro-grammes are available that can assist the qualitativeresearcher. Software programmes vary in design andapplication but can be divided into text retrievers, codeand retrieve packages and theory builders.6 NVivo andNUD*IST are widely used because they have sophisti-cated code and retrieve functions and modelling cap-abilities, which speed up the process of managing largedata sets and data retrieval. Repetitions within data canbe quantified and memos and hyperlinks attached todata. Analytical processes can be mapped and trackedand linkages across data visualised leading to theorydevelopment.6 Disadvantages of using qualitative datasoftware packages include the complexity of the soft-ware and some programmes are not compatible withstandard text format. Extensive coding and categorisingcan result in data becoming unmanageable and

Table 1 Data extract containing units of data and line-by-line coding

Data extract (carer) units of data (in vivo codes highlighted)Early descriptive codes/line-by-linecoding

He (the doctor) said there was nothing more he could do for her. I said to him,cant you give her dialysis? And he said, no because it would kill her. I supposeits too late in the day. I dont know. Thats the reason he gave me, it would killher.So I dont really know, but I thought well, why wait till theres only 20% functionleft before you tell me in the first place. So shouldnt he have told me when shecould have had dialysis? Shouldnt someone then have said to me, well look, shecan have dialysis before it got to the stage where she suddenly has 20% offunction and she cant have it. Couldnt someone have mentioned it earlier? Youknow what Im trying to say?

Nothing more they could do WantingdialysisTreatment would killToo late to treatTreatment would killNot being told early about prognosis/reduced kidney functionNot being involved in treatment decision/confusion*Missed treatment opportunity?

*This early description can be tracked through the following tables, essential in demonstrating transparency.

Editors choiceScan to access more

free content

10.1136/eb-2013-101603

1School of Nursing andMidwifery, Queenss UniversityBelfast, Belfast, UK2Department of Health Sciences,University of Huddersfield,Huddersfield, UK

Correspondence to:Dr Helen NobleSchool of Nursing andMidwifery, Queens UniversityBelfast, Medical Biology Centre,97 Lisburn Road, Belfast BT97BL, UK;helen.noble@qub.ac.uk

2 Evid Based Nurs January 2014 | volume 17 | number 1 |

Research made simple on 22 A

pril 2018 by guest. Protected by copyright.

http://ebn.bmj.com

/E

vid Based N

urs: first published as 10.1136/eb-2013-101603 on 20 Novem

ber 2013. Dow

nloaded from

http://ebn.bmj.com/

researchers may find visualising data on screen inhibitsconceptualisation of the data.

How do you begin analysing qualitativedata?Despite the diversity of qualitative methods, the subse-quent analysis is based on a common set of principlesand for interview data includes: transcribing the inter-views; immersing oneself within the data to gaindetailed insights into the phenomena being explored;developing a data coding system; and linking codes orunits of data to form overarching themes/concepts,which may lead to the development of theory.2

Identifying recurring and significant themes, wherebydata are methodically searched to identify patterns inorder to provide an illuminating description of a phe-nomenon, is a central skill in undertaking qualitativedata analysis. Table 1 contains an extract of data takenfrom a research study which included interviews withcarers of people with end-stage renal disease managedwithout dialysis. The extract is taken from a carer who istrying to understand why her mother was not offereddialysis. The first stage of data analysis involves theprocess of initial coding, whereby each line of the datais considered to identify keywords or phrases; these aresometimes known as in vivo codes (highlighted) becausethey retain participants words.

When transcripts have been broken down into man-ageable sections, the researcher sorts and sifts them,searching for types, classes, sequences, processes, pat-terns or wholes. The next stage of data analysis involvesbringing similar categories together into broader themes.Table 2 provides an example of the early developmentof codes and categories and how these link to formbroad initial themes.

Table 3 presents an example of further categorydevelopment leading to final themes which link to anoverarching concept.

How do qualitative researchers ensure dataanalysis procedures are transparent androbust?In congruence with quantitative researchers, ensuringqualitative studies are methodologically robust is essen-tial. Qualitative researchers need to be explicit indescribing how and why they undertook the research.However, qualitative research is criticised for lacking

transparency in relation to the analytical processesemployed, which hinders the ability of the reader to crit-ically appraise study findings.7 In the three tables pre-sented the progress from units of data to coding totheme development is illustrated. Not involved in treat-ment decisions appears in each table and informs oneof the final themes. Documenting the movement fromunits of data to final themes allows for transparency ofdata analysis. Although other researchers may interpretthe data differently, appreciating and understandinghow the themes were developed is an essential part ofdemonstrating the robustness of the findings. Qualitativeresearchers must demonstrate rigour, associated withopenness, relevance to practice and congruence of themethodological approch.2 In summary qualitativeresearch is complex in that it produces large amounts ofdata and analysis is time consuming and complex.High-quality data analysis requires a researcher withexpertise, vision and veracity.

Competing interests None.

References1. Lee B A real life guide to accounting research: a behind the

scenes view of using qualitative research methods. Amsterdam:Elsevier, 2004.

2. Morse JM, Richards L. Read me first for a users guide toqualitative methods. London: Sage Publications, 2002.

3. Smith J, Cheater F, Bekker H. Theoretical versus pragmaticdesign challenges in qualitative research. Nurse Res2011;18:3951.

4. Noble H, Kelly D, Hudson P. Experiences of carers supportingdying renal patients, managed without dialysis. J Adv Nurs2013;69:182939.

5. Robshaw M, Smith J. Keeping afloat: student nursesexperiences following assignment referral. Nurse Educ Today2004;24:51120.

6. McLafferty E, Farley AH. Analysing qualitative data usingcomputer software. Nurs Times 2006;102:346.

7. Maggs-Rapport F. Best research practice: in pursuit ofmethodological rigour. J Adv Nurs 2001;35:37387.

Table 2 Development of initial themes from descriptivecodes

Early descriptive codes/categories Broad initial theme

Wanting dialysisNo benefit from treatmentsNot involved in treatment decisions*Poor understanding of diseasemanagementConfusion about treatmentsNot sure about which treatmentoptions to takeRequiring further knowledgeInadequate communication

The less informeddecision

*This early description can be tracked through the followingtables, essential in demonstrating transparency.

Table 3 Development of final themes and overarchingconcept

Categorydevelopment Final themes

Overarchingconcepts

Arduous nature ofdialysis

Informed andautonomousdecisions

The patientsdecision

Difficulties in gettingto hospital

Previous experienceof dialysis

Age as a reason notto start dialysis

Uncertainty abouttreatment options

Less informeddecisions

Not involved intreatment decisions*

Poor understandingof diseasemanagement

Confusion abouttreatments

*This early description can be tracked through the followingtables, essential in demonstrating transparency.

Evid Based Nurs January 2014 | volume 17 | number 1 | 3

Research made simple on 22 A

pril 2018 by guest. Protected by copyright.

http://ebn.bmj.com

/E

vid Based N

urs: first published as 10.1136/eb-2013-101603 on 20 Novem

ber 2013. Dow

nloaded from

http://ebn.bmj.com/