Qualitative analysis looks easy enough. You read your transcripts, summarize what they said and suggest what they mean. Unfortunately, it rarely works out like that. You get lost in a sea of quotes. One eloquent respondent sticks in the mind, and another tongue-tied interviewee gets forgotten. The data seems to provide exactly the conclusions you wanted – and to provide something entirely different to your boss.
Qualitative data is important. It can provide whole reams of meaning that bare statistics would miss. And done right, there is no need for late night tears amid a towering pile of transcripts. Here I lay out some tips I’ve picked up from a succession of short-deadline, high-pressure research projects.
Have a structure: Know your hypotheses and your overall table of contents inside out. Without knowing what you hope to do with the data, you cannot begin to classify it. (At the same time, you have to remain receptive to unexpected findings; it’s a tricky balance).
Use a data stripping sheet: This is an Excel document into which you paste the quotes. For each quote you then note in separate columns what section of the report it is likely to be most useful for, what the point being made in the quote is, and which interview it came from. This allows you to begin to sort and assess the data. Ensure you have a notes column, so that you can begin to record thoughts on the meaning of quotes.
Take out all quotes, no matter how dull: It takes a while, but if you don’t, then you never know how significant a point is, because you won’t know how widespread the sentiment was. You will also end up favouring the most eloquent respondents, who will almost certainly not be representative. Virtually every word of an interview ought to go into the data stripping sheet.
Label consistently: Be as consistent as you can in how you label the points being made (i.e. a point about mobility might be labeled ‘Moving house’ by one person, but ‘Leaving home’ by another – when you sort the quotes alphabetically, these two aspects of the same point will not appear together). Much of this can be addressed by preparing a list of labels in advance – but keep checking and updating it.
Source rigorously: Rigor in labeling where you got a quote from always pays off. Never use a quote without its source. There is – take it from me – no worse job than going through interviews late at night, trying to identify an orphaned quote.
Separate data and analysis: Simply putting two quotes together implies analysis, because you must have a reason for doing so. At every stage of the process you will be beginning to knit data into stories. However, it is a useful process to attempt to separate data and analysis, so that you do not fill holes in your data with secondary research or assumption.
Discussion, discussion, discussion: The significance of a set of quotes is almost never what you first think. You will reveal errors and deepen understanding by allowing plenty of time to discuss your findings with others conducting analysis. If you can, sit in the same room throughout.
Follow these tips, and your analysis will hopefully be a lot smoother!
Mission & Vision
The goal of the BDI has been to assemble and assimilate the full extent of creative and scientific knowledge on group behaviour and the dynamics of change. Read more
About Our Research
BDI has worked on projects across the world. For a sample of projects which BDI has either led or consulted on, click here.
The BDI's global network of expert members share their research and their wealth of practical and theoretical knowledge and experience. Read more
The Behavioural Dynamics Institute (BDI) was founded in 1989 and was formed out of the Behavioural Dynamics Working Group. Read more