The researcher can choose how to analyse output of the focus group discussions. This might be qualitative, ethnographic, or based on systematic content coding. Even when using a systematic coding method it is a mistake to use focus groups as though they were questionnaire surveys, and to try to use the results to generate quantitative findings. The group members are not randomly selected, they are purposively selected and too small a number to generate statistically valid findings. In addition, the responses given may be influenced by ‘the experimenter effect’ whereby respondents tell the researcher what they think s/he wants to hear and by ‘the conformity effect’ whereby the group norms influences what is and is not said. The best approach to analysing focus group results appears to be thematic.
Most methods of analysis are based on transcriptions of the focus group discussion. The researcher will need to decide whether these notes need to be verbatim or summarised against pre-determined criteria. Researchers may simply wish to extract selected comments, and use selected material to generate short case studies to illustrate findings generated using other methods. This loses a lot of detail, but the detailed coding and analysis of the output from focus group discussions can be very time consuming due to the quantity of detailed information and the diversity of opinions (Sherraden, 2001), and researchers may instead choose to use only some of the information the discussions generate.
Researchers can choose between two systematic approaches:
- the coded key word approach, or
- searching the text for particular phrases or words.
This latter approach is often taken by ethnographic research, but the coded key word technique allows for more precision in identifying comments and does not risk "losing" a comment merely because your search for particular words fails to pick up phrases with a similar meaning. Risks of this increase where your ‘population’ includes individuals from a wide range of backgrounds, as their language patterns and choices of words may be highly varied.
The coded key word approach involves reading the focus group notes thoroughly and assigning a code/ key word to each comment in order to identify themes or categories within the text. Coding sets the stage for systematic analysis of focus group comments through the application of a "text management" or "ethnographic retrieval" program. CAQDAS – computer assisted qualitative data analysis software – can be used to support data analysis. GoFer, NUD*IST, Ethnograph, CODE-A-TEXT. The GoFer (as in "go for this and go for that") program has been found it to be very suitable for analysing and summarising focus group records. It permits work on multiple files of several thousand pages simultaneously. GoFer will operate on raw text or on coded text, applying logical operations of "and," "or," or "nearby." It will produce a count of instances in which the desired combinations occur, and it will, if desired, show each one on the screen and allow transfer to another document. GoFer is a "resident" program, i.e., it can operate simultaneously with a word processing program. This makes it ideal for counting types of comments and locating examples for illustration. CODE-A-TEXT can be used with digitised sound files, instead of transcribing and coding the written material. Coding, ‘memoing’ and analysis can be done in the software while listening to the sound file. CODE-A-TEXT can also be used for content analysis with transcribed data.