1) Detecting patterns; associative analyses and identification of clustering
Associative analysis involve finding links or connections between two or more phenomena. These connections may be in the form of : (1) linkages between one or more sets of phenomena, or (2) attachments to subgroups.
LINKAGES BETWEEN SET OF PHENOMENA (MATCHED SET LINKAGES)
It is common in qualitative analysis to find that linkages repeatedly occur between sets of phenomena.
We have termed these matched set linkages.
Matched linkages cannot be verified until the full data set is reviewed although they may begin to emerge at a much earlier stage of analysis. When the data set is complete they can be found by looking across a range of different phenomena across all the cases. Such searches are rarely random - they may well be led by something that one of the participants has said, evidence or theories from other research or by a hypothesis which is being tested.
ATTACHMENT TO SUBGROUPS
It is often important in a qualitative study to investigate whether there are any patterns occurring in the data within particular subgroups of the study population. The groups concerned may be those determined by primary sampling criteria or by other socio-demographic characteristics or may have been established as important subgroups or typologies during the analysis.
Typologies and other group classifications are extremely useful in displaying associations in qualitative data by showing how particular views or experiences may attach to particular groups or sectors of the population (Hammersley and Atkinson, 1995).
The search for patterns or differences takes place exactly as described for matched set linkages. However, in this case, the focus of the search is known in advance (that is, to detect differences between identified groups) so the data can be ordered in a way that makes inspection easy to undertake. This can be done manually through summaries of the data or by reordering the data set on the computer.
Having found what appear to be linkages and associations in the data, it is necessary to then explore why they exist.This is because the relationship itself - that is, that there is a connection between X and Y - is not verifiable within the small, purposively selected samples used in qualitative research unless the explanation for their occurrence can also be found. The methods used to verify associations are the same for each of the types of associative analysis described above.
A first step is to check exactly how the level of matching between the phenomena is distributed across the whole data set. This is one of the few occasions when numerical distributions are used in qualitative research - but as a means, not an end, to gaining understanding. The counts will show how many times phenomenon A links with phenomenon B - and within which subgroups in the sample. It will also show where there is no matching of the kind under study.
A second step is to interrogate the patterns of association. Unlike largescale quantitative surveys where a correlation may be presented as an output in its own right, in qualitative research a pattern of association is used as a pointer towards further stages of analysis. The evidential base of a qualitative data set is a rich resource in offering explanations of why phenomena are occurring. Now is the time to use it.
A pattern has been found and appears significant - why is it occurring?
The way in which explanations are developed is discussed below, but it is important to stress that in the search for explanations the analyst looks not only at cases that fit the pattern, but also at cases that do not. In qualitative analysis, 'outliers' as they are sometimes termed, should never be ignored.This is partly because a qualitative analysis is not complete until all the scenarios discovered have been examined, even if they cannot be fully explained in the testing of explanations. For example, they may show that the original pattern was perhaps a false lead; or that other factors also have an influence on the phenomena under study such that a more refined or complex analysis can be developed. Search continues until all those that are out of pattern have been examined. This either brings further refinement to the tiers of explanation - or it leaves some individual cases as unexplained puzzles. Either way the continued search has a payoff in terms of deepening understanding of what is occurring in the data set.
The search for explanations is a hard one to describe because it involves a mix of reading through synthesised data, following leads as they are discovered, studying patterns, sometimes re-reading full transcripts, and generally thinking around the data. It involves going backwards and forwards between the data and emergent explanations until pieces of the puzzle clearly fit. It also involves searching for and trying out rival explanations to establish the closeness of fit. In essence, it is a stage at which the data is interrogated in a number of different ways to further understanding of what is causing or influencing phenomena to occur.
Explanations rarely just emerge from the data. As Richards and Richards comment, they are more often:
... actively constructed, not found, as Miles and Huberman nicely put it, like
Tittle lizards' under rocks. They will continue to be constructed by human
researchers. They are 'mental maps', abstracted webs of meaning, that the analyst
lays over bits of data to give them shape without doing violence to them (1984: 83).
The researcher must weave these webs ... see the links and draw the threads together, often by creative leaps of imaginative analogies. (1994: 170)
There are a number of ways in which the researcher can build an explanation, depending in part on the nature of the study, the emergent patterns within the data, and the researcher's own theoretical or epistemological perspective. To unpack the way in which explanations are developed, it is helpful to distinguish between different types of explanation. At an analytic level, explanations may be based on the explicit reasons that are given by participants themselves, or alternatively implicit reasons that are inferred by the analyst. Within these two approaches, explanations may be dispositional - that is, they derive from the behaviour and intentions of individuals; or they may be situational - that is, attributed to factors from a context or structure which are thought to contribute to the outcome (Layder, 1993; Lofland and Lofland, 1995).
It is important to note a difference in the nature of the evidence used to generate and support explicit and implicit accounts.
For explicit accounts, the evidence appears overtly in the reasoning within the participants' responses (see below). For implicit accounts, on the other hand, the researcher may draw on patterns within the data
Alternatively, the researcher may deliberately put together different pieces of evidence in order to develop or construct an explanation.
Where reasons are implicit and inferred by the researcher, the process may entail: (1) searching for a possible underlying logic within what people have said; (2) using common sense to search for explanations;(3) applying powerful analytic concepts; (4) comparing findings with those in other studies; or relating findings to a more theoretical framework.
Again, each of these is discussed below.
Using explicit reasons and accounts: During an effective in-depth interview, participants will always be asked why they feel, act and believe as they do and these explicit accounts are of immeasurable value in understanding motivations and intentions. The researcher may decide to simply present the recurrence, range and diversity of explanations given by participants themselves, or to look for patterns among and offer explanations for these explicit accounts. These may be dispositional - for example, the aspirations and requirements that lead to the choice of a particular vocation or career; or situational, for example, the features of dental surgery delivery that have put people off going to the dentist.
Inferring an underlying logic : It may be the case that deeper explanations of a phenomenon are not immediately conveyed, or even clearly understood, by the individuals themselves and the researcher will want to identify factors which are not initially evident in the data.
Three approaches are possible on the basis of such evidence: In the first, the researcher may tease out an explanation based on the juxtaposition or interweaving of two apparently unconnected themes.It is also possible to use the absence of phenomena to inform the underlying logic of an explanation. This can arise, for example, when a feature that is formative in some people's accounts is entirely missing from others. This raises an important question as to why this is the case which then needs to be investigated. This can be done through examining the accounts of people who have not mentioned a factor or reason to see if an explanation for its lack of relevance or influence can be found; or comparing the two sets of accounts to see what differences might explain its presence or absence.
Using common sense to search for explainations: The researcher may follow common sense assumptions when attempting to explain patterns within the data. These premises or assumptions may either fit a pattern commonly known to exist or simply make straightforward 'sense' through something seen in the data. However, once they have been made explicit, they will need to be fully interrogated across the whole dataset to ensure that their explanatory base is supported.
Developing explanatory concepts: Sometimes a powerful analytical concept which is developed in the course of the study can itself explain a phenomenon. Often these are underpinning or 'meta' concepts that make it possible to place important emergent themes within a broader explanatory framework.
Drawing from other empirical study: Ideas and hunches about possible explanations can also come from comparing the researcher's own study with others which have been carried out in the same or a similar field. Here the researcher may 'borrow' concepts or explanations to see how well they fit his or her findings.
Using theoretical frameworks: Where researchers are interested in a particular field or body of literature, or where they are committed to a particular theoretical perspective, they may wish to relate their local findings to a broader context and develop 'local' explanations in accordance with their chosen theoretical or analytical framework.
Explanations developed in this way must be carefully checked to ensure that they reflect the uniqueness and diversity of the data and do not 'bully' the findings to fit preconceived ideas. We have argued that, at the beginning of the analytical hierarchy, the researcher should stay close to the participants' own language and accounts and then, later in the analytical process, introduce theoretical concepts or theories in as far as they actually match, the data. If a theoretical framework is applied to the data too early in the analytic process, much of the detailed richness of the data will be lost.
Seeking wider applications
The final tier of analysis involves a consideration of whether evidence from the study has some wider application. This might be a contribution to theory or to a theoretical debate, suggested strategies for the formulation or realignment of a social policy, or recommendations about practices within a public service. The next chapter is devoted to a discussion of how qualitative data can be generalised and the different kinds of wider inference that can be drawn. But, at the end of this analysis chapter, it is important to recognise that any consideration of the wider applications of research findings forms part of the analytic output from a study. As such it needs to be strongly supported by evidence with a clear exposition of how the inferential or explanatory arguments have been developed.
Source: Miles, M. B., Huberman, A. M., & Saldana, J. (2013). Qualitative data analysis. Sage.