Unit Author: Professor Nick J Fox

Learning objectives

Having successfully completed the work in this unit, you will be able to:

  • Evaluate the advantages and disadvantages of qualitative methods of research.
  • Compare and contrast naturalistic and positivistic approaches to research.
  • Identify the main methods of qualitative data collection.
  • Compare and contrast issues of validity and reliability in quantitative and qualitative research.


While often it is possible to ‘quantify’ the findings of a study: to count the number of events, or measure the extent to which a variable is present, absent or correlates with other variables, sometimes such quantification is either not possible or is not appropriate. In such situations, the quality of data is more important than quantities. This is particularly the case in the human and social sciences, where what is being explored are human attributes such as attitudes, beliefs or judgements. This course unit looks at how we can make such explorations rigorous and systematic

Qualitative methods are sometimes criticised (usually by researchers committed to quantitative approaches to research) as ‘woolly’, ‘unscientific’, ‘journalistic’ or simply a waste of time and money. Poorly done, I would agree that qualitative research could be all these things (as could any research).

As we will see in this unit, the need to specify a research question, to refine that question to make it feasible, to identify concepts and indicators, and to find ways to rigorously and systematically collect valid and reliable data are all part of a qualitative methodology. In their detail they may be quite different from quantitative approaches (for example, inferential statistics play no part in qualitative data analysis). The choice between quantitative and qualitative approaches needs to be made by asking these questions:

  • Which is the most appropriate method to answer my research question?
  • Which method is the most adequate to provide the data that I need to collect?

In this unit we will look at four methods of collecting data:

  • participant observation (including ethnography)
  • interview-based research (including focus groups and consensus techniques such as Delphi methodology)
  • documentary and material culture approaches (using public or private documents, or objects and other material artefacts that reveal aspects of a culture or sub-culture);
  • action research (although not necessarily a qualitative methodology, it can be best considered here).

1. The nature of qualitative design

The purpose of qualitative methodology is to provide rigorous designs to enable use to gather and interpret this data on quality. Each of the methodologies we will consider approaches this task in a different way, but the objective is the same: to provide us with ways to understand the world which add more than a purely quantitative approach can achieve.

Three points are worth making:

First, it is a premise that qualitative methodology is not intrinsically better or worse than a quantitative alternative. The appropriateness of any methodology can be judged only in relation to the research question that needs to be answered. No methodology, from a randomised controlled trial to an ethnographic study, should be selected for its own sake: the merit of a methodology rests only upon its capacity to generate answers to a specific research question.

Second, the need to specify a research question, to refine that question to make it feasible, to identify concepts and indicators, to find ways to rigorously and systematically collect valid and reliable data, to analysis that data, and to draw inferences or conclusions are all part of a qualitative methodology. In their detail they will be quite different from quantitative approaches (for example, inferential statistics play no part in qualitative data analysis). But they are based upon epistemological propositions concerning how to gain knowledge of the world, and the designs have been developed to generate ‘evidence’ just as have quantitative methodologies. There is continuous methodological innovation in qualitative research, and entire journals are devoted to discussion and application of such techniques (for example, the journals Qualitative Research and Qualitative Inquiry).

Finally, we may note that quantitative and qualitative are used increasingly in tandem, providing a richness of data that could not be achieved by uses of one alone. For instance, the detailed description and inductive reasoning provided by a qualitative study (for example, an observation of a specific care setting) can be a stepping stone to a quantitative study that gathers detailed measurements of phenomena in a range of settings. On the other hand, a quantitative survey (for instance, of sickness among ethnic groups in the UK) may pose questions that require the in-depth interpretive work that an interview or focus group can contribute subsequently.

2. Qualitative research questions and methods

We will now turn to the four main designs in qualitative research.

Below are some of the kinds of research questions that each has been developed to answer, although of course the range is vast.

Table 8.1 Research questions and qualitative methodologies

Types of research question to be answered Methodology
  • What happens in a setting?
  • What are the organisational characteristics of a setting?
  • How do people’s attitudes and beliefs affect what they do?
Participant observation
  • What do people think about an aspect of their lives?
  • What is the range of views on a subject among interested parties?
  • Is there a consensus view about a subject?
Interviews,focus groups andconsensus techniques
  • What has been written/recorded about a subject over a specified period of time?
  • What are the cultural products of a group, and how do these relate to their values, attitudes and beliefs?
Documentary and material culture studies
  • What happens when we try to improve the quality of a setting?
  • How may we enable people in a setting to improve their situations?
  • How can we challenge established power structures in a setting?
Action research

SAQ 8.1 Match the methodology to the study

Here are a series of ideas generated by social researchers.Suggest appropriate qualitative methods for each of them. (There may be more than one per question).

[table id=44 /]

3. Participant observation and ethnography

The terms participant observation and ethnography are used interchangeably to describe a type of qualitative research that uses extensive fieldwork observation to generate rich data about a setting.  Participant observation is ‘about engaging in a social scene, experiencing it and seeking to understand and explain it. … By listening and experiencing, impressions are formed and theories considered, reflected upon, developed and modified’ (May, 2011: 189).

Ethnography (meaning ‘writing a culture’) is a term that derives from anthropology, and traditionally this term was reserved for studies of ‘other cultures’, typically of developing world or indigenous peoples by Europeans. More recently, this term has come to be used to also document ‘own culture’ studies, partly to ensure that some principles established in anthropological methodology are applied: for instance, the effort to ‘make strange’ that which is familiar, and to resist making assumptions about what one is observing.

Ethnography also implies a particular kind of research report which seeks to fully document a setting (Bryman, 2012: 432). ‘An ethnography’ thus refers to this written output.

To describe the specific methodology of field observation, we can use the term participant observation, to encompass both the anthropological tradition and the body of qualitative research that emerged from sociology in the 1920s and 1930s under the influence of the ‘Chicago School’ (May, 2011: 161). Studies on crime and deviance, race relations and urban life were classic studies using this approach, but it has been applied widely and some of the most significant studies in this genre are in health and social care.

As the name suggests, two things are involved in participant observation: observation and participation. The first is comparatively straightforward: this is a methodology that is based on direct observation of a setting by a researcher. The second is more significant: it recognises that the researcher is a participant in this process, immersed in the setting rather than being separate from what is going on. Proponents of participant observation are adamant that few kinds of observation are non-participatory. Only what has been described as the complete observer can completely avoid involvement in what happens in a field, through the use of a one-way mirror or closed-circuit TV.

Typically participant observation takes place over an extended period, during which a researcher familiarises her/himself with the setting and the people who populate it.  During this period of fieldwork the objective is to gain a ‘rich picture’ of what is going on, document this by field notes and other recording means, and then analyse and interpret the findings to establish some generalisations about what is going on in a setting.

Participation in the field can take a variety of forms, depending on the role that the researcher adopts. Tim May (2011: 172-3) describes three models of participation:

The complete participant

A researcher may become a complete participant. S/he covertly becomes part of the field, taking on a role. For instance, a social scientist might gain employment as a hospital porter in order to undertake covert observation of what goes on and around a surgical operating theatre, or a researcher might take a job as a waiter to observe the restaurant business from the inside. It is argued that this reduces the likelihood that a researcher’s presence will affect the behaviour s/he is observing. Needless to say, it requires total commitment by the researcher for an extensive period of time, and may be emotionally draining if one tries to sustain a deception when interacting with co-workers who may even become friends. There are also ethical questions about this kind of participant observation.

Participant as observer

Here the researcher is open about her/his position and does not attempt to disguise what s/he is intending as an outside observer. The researcher seeks to establish good relationships with subjects, so that the latter serve both as informants and respondents. Typically, a ‘key informant’ will emerge, who provides detailed information that could not be gathered simply by observation, either because it requires background knowledge, or because it is not immediately comprehensible to an outsider (for instance, technical processes or culturally-specific behaviour such as rituals or conventions)

The researcher may adopt a role rather like a ‘fan’ of a celebrity or prestigious institution, seeking to learn as much as possible from people in the setting, while standing back from becoming a member of the group. Thus, a researcher might seek to learn from a consultant surgeon why they perform certain tasks in particular ways. The researcher has to guard against ‘going native’ in this role, as this will endanger the capacity to stand back from unquestioning adoption of setting norms of behaviour.

Observer as participant

Here there is very limited contact with the field, and may be a one-off engagement, possibly in the context of undertaking an interview. There is little effort to fully understand the norms, values and beliefs of the people one is observing. An example would be to take a tour of an operating theatre/room where a surgeon is operating as part of a visit to a hospital to interview administrators about management of waiting lists. Any observation is formalised, and only provides context to other data sources.

3.1 Online ethnography

With the emergence of the internet and more recently so-called ‘Web 2.0 media such as blogs, discussion fora, Facebook and Twitter that enable people to interact online, social researchers have developed methods to use the material available through these social media as research data.  Some researchers have regarded online fora and social media as like physical spaces where people interact.  So they have mimicked ethnographic techniques of participant observation within these ‘virtual spaces’.

Taking the opportunities afforded by these fora (for instance, those associated with vegetarianism or self-help groups for people with illnesses or disabilities), the simplest approach has been to use the Internet to ‘lurk’ in these public or semi-public interactive spaces, noting the textual and other kinds of interactions that take place, and using these interactions as their raw data.  Other researchers consider this kind of covert observation to be unethical, and have typically made themselves either complete participants in the community, or revealed their identity as a researcher and asked permission to observe and record interactions, subject to the usual guarantees of anonymity and confidentiality.  The latter option does run the risk of refusal, but that is of course the same for any other kind of participant observation.

Participating in these virtual spaces has advantages of accessibility and convenience but there are issues of validity, as it can be harder to judge whether participants are truthful, honest and even who they say they are.  There is also a high risk of a researcher being rejected by a group, especially if it addresses a sub-culture, an exclusive activity or a persecuted minority (for instance, pro-ana supporters).  Online ethnographic approaches also pose some challenges in terms of how to analyse the heterogeneous data (text, audio, video) they generate.Many of the other issues described below are also relevant.

For an example of online ethnography, see Fox and Ward (2008) which describes a study of a vegetarian online forum.

3.2 Issues in participant observation and ethnography

This design raises various methodological issues:


Gaining access to a field may require prolonged negotiation. Even when invited to observe, it may take a long time to be fully accepted, as outsiders can be regarded with suspicion, especially where the people being observed are not powerful (for example, hotel porters). It will take a while for the subjects to relax and behave ‘normally’ when the researcher is present.

The research bargain

Observers may have to give something back in return for being accepted. This may simply be a matter of feeding back findings to participants, or can involve more complex and even financial contributions to key informants.

Emotional engagement and objectivity

Being part of a field setting can become the major part of a researcher’s life. It may be difficult to keep professional distance, and field notes need to be used conscientiously to ensure validity and reliability of data. Leaving the field may be hard for both researcher and subjects.


Participant observation turns ordinary people’s lives into the subject of academic study. Is it ethical to do this, and what should be the benefit to the people in return for this incursion?

3.3 Advantages and disadvantages of participant observation and ethnography

Advantages Disadvantages
  • Provides in-depth data about a setting (highly internally valid)
  • Good for identifying cultural norms, values and beliefs
  • Gives an insider view of a setting
  • Time consuming and emotionally difficult
  • Needs to ensure subjects are not exploited by making them ‘guinea pigs’
  • Hard to generalise (low external validity)
  • Hard to sustain objectivity (intra-observer reliability)

SAQ 8.2 Ethnography

Please read this paper, which reports an ethnography of religious ceremonies:Donnelly, C.M. and Wright, B.R.E. (2013) Goffman goes to church: face-saving and the maintenance of collective order in religious services.  Sociological Research Online, 18 (1) 18.This may be accessed online at http://www.socresonline.org.uk/18/1/18.html

Now answer the following questions

[table id=45 /]

4. Interview methodologies, including focus groups and consensus techniques

The interview is one of the most familiar methodologies used by qualitative researchers. In this section, we will look at the main approaches to interviewing, and then consider other methodologies that have adapted interviews to somewhat different ends.

The interview is based on a structured conversation between research and respondent. Normally the latter is asked a series of questions, which may or may not be pre-arranged and may or may not be delivered in a set order. The answers constitute the data that are to be analysed, possibly enhanced by notes on non-verbal behaviour on the part of a respondent during the research encounter.

The objective of an interview may be:

  • to elicit factual information about a person or a phenomenon of which the respondent has knowledge
  • to elicit data on experiences, attitudes and opinions
  • to enable a researcher to gain impressions of the emotional responses to certain subjects by analyse of verbal and non-verbal behaviour.

4.1 Types of interview

Different typologies of interviews have been developed (May, 2011; Bryman, 2012), but we will look at four main kinds of interview.

4.1.1 Structured or standardised interview

This approach has been considered in Unit 6.

4.1.2 Semi-structured interview

This falls between structured and unstructured interviews: questions are specified, but the interviewer can explore in detail particular topics that emerge during the interview. The purpose is to achieve clarification and elaboration of themes (May, 2011: 134). Interviewees may be encouraged to expand on issues as they are raised.

Questions in this kind of interview tend to be more general, enabling respondents more flexibility to give responses hat match their, as opposed to the researcher’s agenda. The order of questions is also flexible, enabling a more ‘natural’ flow to an interview.

The advantage of this approach is that it sustains some structure, in terms of the topics to be covered, and perhaps the order in which questions are asked, to enable rigour to be sustained, while at the same time permitting respondents to develop their answers, providing richer data.

4.1.3 Unstructured or focused interview

The third kind of interview, and one that epitomises the naturalistic frame of qualitative data collection, is open-ended in character, allowing the most flexibility to researchers to explore topics with interviewees, and for respondents to follow their own trains of thought and lines of reasoning.

The interview is guided by an aide memoire of topics to be covered, rather than any kind of formal schedule. This leads to an interview that may have more of the character of a conversation.

This approach is called focused because it enables the interview to focus on the respondents’ frame of reference and enables them to lead the interview into their own focus, rather than one pre-determined by a researcher.

The disadvantage of the unstructured/focused interview is that it reduces comparability between subjects. Its advantage is that it enhances the internal validity of a study by ensuring it addresses the issues important to respondents, and also reduces the likelihood that a researcher will ‘lead’ respondents towards some bias of their own.

4.1.4 Group interviews (focus groups)

A group interview enables a number of respondents to be brought together to discuss an issue. Group interviews can use unstructured or semi-structured approaches; a structured group interview is unlikely.

In the former, the group is encouraged to respond flexibly, as in an unstructured/focused interview, with the researcher using a list of topics rather than specific questions to stimulate the discussion. The objective is to generate a broad-ranging output, to maximise the data that can be collected, and to explore group dynamics and a level of reflection that might not be achieved in a one-to-one interview. This kind of interview is also known as the focus group. In this methodology, the researcher may take on the role of group moderator. The objective is to stimulate group dynamics, so that data reflects the interactions between people as well as individual responses. For example, a focus group conducted with teaching staff and managers to discuss how a school is run may provide data on the power relations between staff.

However, it is also possible to apply a semi-structured or approach, in order to elicit more specific information. For instance, a researcher may wish to gather data on a variety of relatively factual topics, all of which a single respondent might not know. Thus, at the start of a piece of exploratory research or ethnography, a group interview of different ‘experts’ could be used to efficiently generate the background information that a researcher needs, rather than undertaking a series of one-on-one interviews.

Group interviews do seems to generate richer data, with respondents sometimes more willing to speak out in a group then individually (May, 2011: 137-9).

There is one specific type of group interview that warrants a separate section: this is the so-called Delphi approach.

5. The Delphi methodology

The Delphi methodology, named after the Oracle of Delphi in ancient Greece (said to be able to predict the future and supply knowledge to these who consulted it), is a means to gain consensus views among a range of experts in a field. It uses multiple ‘rounds’ or waves of questions, with the results from a round fed back to the same respondents during a subsequent round, until consensus is reached.

While often using quantitative approaches such as questionnaires or surveys, sometimes a Delphi study may incorporate group interviews.

Somewhat like a focus group, Delphi methodology tends to use very focused questions in order to stimulate discussions. However, the intention here is not to stimulate group dynamics, but to ensure consensus. What is aimed for is an output that reflects what all involved can agree upon. Sometimes this is fairly limited, but can be used as the basis for a policy or standard. For example, when developing guidelines on school meal provision, a Delphi approach could bring together different professionals to seek an agreed strategy. Clearly it is important that those involved are leaders in their fields, and can represent the shared views of their professional colleagues.

6. Issues in interviewing

There are a number of issues in using interviews to gather data.


It is sometimes hard to persuade people to be interviewed, especially powerful or high status individuals. Setting up group interviews can be difficult, as many people need to find a suitable time


There is no guarantee that respondents will tell the truth, and may refuse to speak about some subjects. Unlike participant observation, a researcher is dependent on the subject to provide valid data.


Researchers can affect how people respond in interviews.  Males may not be willing to speak freely to females about some subjects and vice versa.  Researchers may also bias findings intentionally or unconsciously if they have strong views about a subject.  These issues can be addressed by recording interviews, checking analysis of interview data and feeding back findings to respondents for ‘member-checking’ to see if they agree with conclusions.


It is now common to record interviews using either audio or video. The latter has the advantage of capturing non-verbal behaviour. When transcribing interviews, some researchers — especially those using a form of detailed transcription called discourse analysis – code all utterances and gaps as well as formal speech. 

Advantages and disadvantages of interviews

Advantages Disadvantages
  • Simple to organise and well-understood by respondents
  • Relatively easy to learn techniques
  • Can generate data that can be recorded for ease of analysis
  • Cheap and quick
  • Validity may be suspect as respondents may be unwilling to talk openly or truthfully
  • Open to ‘researcher-effect’ in which the presence or characteristics of the researcher affects how interviewees respond
  • Transcription of data is time-consuming

SAQ 8.3 Interviewing

[table id=46 /]

7. Documentary and material culture research

Documentary analysis is an approach that derives data from public or private documents, including diaries and other personal records, documents produced by government, policy-making bodies, pressure groups and other organisations.  The range of documents that may be consulted include:

  • personal documents
  • official and historical documents, including government policy documents and legislation
  • official statistics
  • mass media output
  • online documents, including web pages, blog posts, online videos, etc.

In addition, we can include the study of material culture products, by which we mean all artefacts produced by human endeavour: typically, art and craft objects, ritual objects and technological products, including media outputs and the Internet. Material culture products have traditionally been collected and analysed as part of anthropological studies, and often form part of museum collections of ethnology.

There are three main approaches to studying documentary sources and material culture:

  • Realist approach: the document is a medium that provides a window on to the We can use the document to ‘read off’ the underlying reality that it describes, for instance, a pattern reflected in official statistics or an account of an historical event. A typical example would be the use of private papers by a retired politician to explore the discussions that led to a government policy change.
  • Constructionist: a document or object is considered as a representation of the circumstances that led to its creation, for instance the social context within which it was created, or the specific objectives of its author. Thus, a policy document often reflects a desire to change existing practices, while an autobiography may reflect not only what ‘actually happened’ but also the ‘spin’ that the author would like to promote to the world about her/himself. The task of the researcher here is hermeneutic, to discover the ‘hidden meaning’ of the text.
  • A document or object is seen as part of a ‘discourse’ that mediates power and knowledge: what can be spoken about a subject at a particular period in This perspective derives from the work of the social philosopher Michel Foucault and has been applied in the social sciences to offer insights into changing patterns of power and authority in society, and how this is mediated through knowledge. For example, by studying documentary sources associated with the growth of public health in the 20th century, Armstrong (1983) argued that this reflected a new construction of the community as the object of medical power.

7.1 Issues in documentary research

The following issues need to be considered in documentary studies.


Access to documents may be restricted and analysis based on available documents always runs the risk that inaccessible documents may shed an entirely different light on the issue under investigation.


What is the guarantee of a document’s authenticity? Documents may be fabricated to suggest false versions of events, and it may only be possible to determine authenticity by reference to other sources. Sometimes errors are unintentionally introduced into documents when they are copied or translated.


The approach to analysing a document or object will depend on the theoretical stance taken by the researcher (see above).

8. Action research

Action research is not specifically a qualitative methodology, as it can use both qualitative and quantitative approaches. However, it fits neatly in this unit and is a growing methodological approach in social research.

It seeks not only to understand and analyse a setting, but to test out how it might be changed in some way for the better. For example, a teacher might evaluate whether incorporating five-minute breaks into a lesson can improve student concentration.

Carr and Kemmis (1986) describe three categories of action research:

  • technical (in which an outside expert undertakes the research);
  • practical (in which the researched are encouraged to participate in the research process); and
  • emancipatory (in which the researcher takes on the role of a ‘process moderator’ assisting participants to undertake the research themselves).

Emancipatory action research is the most radical, and is predicated on political commitments to participation and improvement in quality of life among participants. Schensul (1987) suggests that it can

  • bring together people with diverse skills and knowledge;
  • de-mystify the research process, allowing practitioners to shape the data collection process;
  • build a research capacity into a community which can operate independently;
  • increase the likelihood of the use of research findings by practitioners; and
  • improve the quality of research by enabling access to key bodies of knowledge in a community.

Action research is sometimes also known as ‘practitioner research’, and in this guise describes research undertaken by a professional (for instance, a teacher, health care worker or a community worker) on the setting in which they themselves work.  For example, a teacher might research how to get parents more interested in the work she is doing with their children.  Practitioner research poses some challenges for objectivity, which we will consider later.

The main advantages and disadvantages of action research are summarised below.

Advantages and disadvantages of action research

Advantages Disadvantages
  • Enables practitioners to research their own workplaces.
  • Focuses on change rather than understanding: the outcome is improvements in service, not academic reports.
  • Can spread research capacity, enabling people to research themselves rather than depending on outside academic researchers.
  • Is setting–specific. Hard (and inappropriate) to generalise.
  • May be of limited use where practice cannot easily be changed (e.g. where there are national guidelines on practice standards).
  • Can be hard to separate dual role of practitioner and researcher leading to questions of reliability of data.

9. Sampling in qualitative methods

You will recall various issues concerning sampling in quantitative research from Unit 5. The issues for qualitative studies are somewhat different, and the approach to sampling is distinctive, as can be seen below. For example, a study may wish to consider attitudes towards eating seasonal produce. A small sample of local allotment owners may be interviewed. The sample is thus chosen on the basis of ‘theory’: this kind of theoretical sampling aims to maximise the range of responses, but does not strictly seek to ‘represent’ all the people in the community and other stakeholders.

9.1 Comparisons of sampling approaches

Quantitative Qualitative
Sample big enough for statistical inference (see Unit 5). Often very small, occasionally a single case.
Selected to be representative. Rarely attempts to be representative: sample chosen to maximise range of responses.
Assumes respondents are independent and do not affect each other’s responses. Respondents’ interactions may be part of the topic for study.

The issue of sampling is closely associated with ensuring validity and reliability. Before looking at specific methods of data collection and analysis, we will not turn to considerations of how to make qualitative research valid and reliable.

10. Validity and reliability in qualitative research

Having read Unit 4, you should be familiar with the principles behind the reliability (or consistency) and the validity (or accuracy) of research, and how to maximise both in the research which you are undertaking. If you need to remind yourself of the main kinds of reliability and validity, you should review this material now.

Validity and reliability are issues for qualitative researchers as they are for others. However, it has been argued that the way in which the former need to address validity and reliability differ from the traditional approach of positivist researchers. The naturalistic paradigm, Lincoln and Guba (1985) argued, places different issues before researchers. However, the question of the trustworthiness of a piece of research still stands. Lincoln and Guba suggest four key areas that we need to consider of a piece of naturalistic (qualitative) research, relating to four questions:

  • How truthful is the finding? (Credibility)
  • Can the findings be generalized? (Transferability)
  • Could the findings be replicated? (Dependability)
  • Can we rule out researcher bias? (Confirmability)

These four areas relate to traditional notions of study validity and reliability, as can be seen below.

Positivist and naturalistic analyses of validity and reliability

Positivist Naturalistic Asks:
Internal validity Credibility Are the findings believable?
External validity Transferability Are the findings applicable elsewhere?
Instrument reliability Dependability If the study were repeated, would the same findings emerge?
Intra-observer reliability Confirmability Has the researcher biased the findings?

Note that just as in the positivist paradigm, validity is dependent on reliability, while the reliability of a study does not guarantee its validity.

Qualitative researchers thus have to address the four issues of credibility, transferability, dependability and confirmability when they undertake their research. However, the way Lincoln and Guba (1985) argue that we need to think about some of these issues differ quite markedly from the perspective of positivist research. Let us look at each area in turn.

10.1 Credibility

Related to the internal validity of a study, credibility requires that we concern ourselves with the accuracy of description in a piece of qualitative research. We need to state the precise parameters of the study – who was studied, where and when, and by what methods. If we identify these aspects, and if we have a reliable means of measurement (dependability and confirmability), our study will be valid for the specific setting investigated.

10.2 Transferability

We noted earlier that in most qualitative research, the method of sampling is not representative, but will be aimed at maximising the diversity within the study setting, to ensure as ‘rich’ a picture of the setting as can be gleaned. Clearly, this method of sampling will not supply external validity in the way that is usually sought in quantitative research, which wishes to generalise from the sample to the population from which it has been drawn.

Thus in a major divergence from positivist approaches, Lincoln and Guba argue that no claims can be made about the applicability of the findings to other settings. If other researchers wish to generalise from a study to other situations, the onus must be on them rather than the original researcher to demonstrate a study’s applicability elsewhere.

10.3 Dependability

Whereas positivist research has to assume an unchanging world, so that if an identical study were to be performed the assumption would be that the same findings would emerge, the naturalistic paradigm acknowledges that the world, especially the social world, is constantly changing. In particular, a study might itself affect the world it is trying to document. (Thus, interviewing someone about their attitudes to vegetarianism might make them think for the first time about what they really do think. Repeat the interview, and the answers given may reflect a new perspective gleaned from the respondent having considered the subject more fully.)

If change is inevitable, then all a research can do is to try to predict as much as possible of what these changes may be, and account for them by casting widely for data within the setting. You should be able to see how the ‘dependability’ of a study is like the different kinds of reliability we looked at in Unit 4: internal reliability of an instrument, test/re-test reliability and inter-observer reliability. You should also be able to recognise the need for dependability in order to obtain credibility.

10.4 Confirmability

In a naturalistic paradigm, we have to accept that observer bias is a fact of life: we all have values and we cannot wholly avoid allowing these to colour the way we interpret data in a qualitative analysis.

To minimise this bias, qualitative researchers need to recognise their biases, and seek to fault their own assumptions or ‘pet theories’ about what they are researching. (This is a particular challenge for practitioner researchers.)  Bringing in colleagues to offer alternative readings, and feeding back results of an analysis to the original respondents can help to reduce these biases. Again, you should be able to see how the confirmability of a study is a pre-requisite for its credibility.

SAQ 8.4 Validity and reliability in qualitative data analysis

Answer the following questions in relation to the validity and reliability of qualitative data collection. 

[table id=47 /]


The approaches to research that have been described in this unit underpin all qualitative research studies. Sometime a mix of methodologies is used, and sometimes they are melded with quantitative methodologies.

These methodologies have in common an emphasis on interpretation, and seek to maximise the internal validity of findings (ensuring that a research question generates the right data to provide the answer).

In contrast, there is lower generalisability (external validity) than quantitative research, and there are significant problems of objectivity and dangers of bias, which require meticulous attention to methods of data collection and analysis.

There are also fundamental questions about the status of knowledge generated by research (is it ‘truth’ or one among many ‘truths’?) Although this question is common to all research, it comes to the fore in qualitative research because these methodologies have acknowledged the role of interpretation in the growth of knowledge.

Naturalistic methods are invaluable in researching the meanings and perceptions of people, either alone or in conjunction with quantitative methods. In health and social care research, qualitative research is now firmly established, and these methodologies are important strategies for the researcher to master.

Now please complete the following exercise for your log book.

Reflective exercise 8.1 Collecting qualitative data

Please write down a research question on a topic that interests you, using one or more of the qualitative methodologies described in this unit. Describe which methodology you would use, why you consider the methodology appropriate, and suggest what problems in data collection and validation you might encounter, were you to undertake this research.

[table id=48 /]


Armstrong, D. (1983) The Political Anatomy of the Body.  Cambridge: Cambridge University Press.

Carr, W. and S. Kemmis (1986) Becoming Critical: Knowing through Action Research. Victoria: Deakin University Press.

Fox, N.J. and Ward, K.J. (2008) Health, ethics and environment: a qualitative study of vegetarian motivations. Appetite, 50 (2-3), 422-429.

Lincoln, Y.S. and E.G. Guba (1985) Naturalistic Enquiry. California: Sage.

Schensul, J.J. (1987) ‘Perspectives on Collaborative Research’, in D.D. Stull and J.J. Schensul (eds.) Collaborative Research and Social Change.  Boulder, Colorado: Westview Press.

Further reading

Bryman A (2012) Social Research Methods. 4th edition. Oxford University Press.

Denzin N and Lincoln Y (eds.) (2011) The SAGE Handbook of Qualitative Researc4th edition.

May T (4th edition, 2011) Social Research: Issues, Methods and Research. 4th edition. Buckingham: Open University Press.

Answers to SAQ 8.1

  1. Interviews with older adults.
  2. Documentary analysis of newspapers, online, television and radio coverage.
  3. Action research, perhaps incorporating ethnography and interviews.
  4. Ethnography, incorporating key informants.
  5. Focus group aiming to achieve consensus (Delphi study).

Answers to SAQ 8.2

  1. The researchers attended public services in the selected churches, and did not tell anyone that they were conducting research.
  2. Participant observation: the researcher participated fully in the services.
  3. Note taking during observational activities that lasted throughout the services.
  4. This is not explicitly discussed, but understanding emerged through the application of a thoeretical framework to the data.
  5. There were no key informants
  6. Covert research may be considered unethical as it breaches the rights of these observed to privacy.  It is unclear whether the researchers were believers.  If not it might be offensive to those observed (for example, if the researchers had participated in communion/mass) if they subsequently learnt of the research.

Answers to SAQ 8.3

  1. Having a schedule of questions will ensure that all respondents are asked the same questions and that nothing is missed. However in an unstructured interview, an interview plan can be used to avoid missing key issues.
  2. The value of a group is that is enables respondents to discuss topics more naturally with each other during the session. Using a structured list of questions prevents such interactions from occurring.
  3. No. However, efforts should be taken to assess in advance whether the gender of the interviewer might be an issue when interviewing about a topic such as medical history or sexual behaviour (a threat to inter-observer reliability).  If so, then consideration should be given to how to avoid gender affecting how respondents answer questions.

Answers to SAQ 8.4

  1. This is a way to check the accuracy of transcripts, and will assist in achieving internal validity (credibility).
  2. A field diary will ensure details are not lost, insights are recorded and can enhance the quality of interpretation of a setting.
  3. You could conduct interviews following field work. This can help to check interpretations or corroborate factual information.

[box type=”info”]This unit is part of our course on Social Research Methods. You must be registered and logged in to access course content. Back to courses Welcome page.[/box]