Qualitative data can have several forms, but usually it consists of text. Use our seven steps to get correct data and use our five steps to analyse these data.
What is qualitative data?
Objects and situations can be shown in real life or on a photo.
To compare two objects or situations text is used.
Text may contain numbers (the fast road is 10 km and the sightseeing road is 12 km), but it still isn’t quantitative data. As a matter of fact, a digit is basically a qualitative datum. Only when many cases are examined in a study and the characteristics of these objects are statistically compared, one has to do with a quantitative research. In all other cases one has to deal with qualitative research and qualitative techniques have to be used. A special case is a case study in which only one object is included in the research.
Objects, situations or photos of it, are the actual data. In this way they are hard to analyse. To describe, compare, evaluate changes of these objects or situations, it is done in a text.
When interviews are being held or if a discussion with focus groups has been recorded, the spoken words are transcribed into text. Again, it is text that has to be analysed.
How to analyse qualitative data
When text is your data, how can it be analysed to support an idea, vision or theory?
Three things can be done with it, and we use them all the time in daily life: refer to it, summarizeit or interpret it. This looks pretty easy, but be aware that mistakes are easily made. You might refer to a sentence spoken by someone, but maybe it was in a special context and in another context it doesn’t make sense.The summary is what you think what the headlines are, and an interpretation is even a step further away from the original spoken words. A lot of bias might occur. To reduce bias special techniques can be used.
For instance if words are frequently used in combination, this can be explored. For a list of words (for instance: man woman, good, bad and pudding) texts are investigated and the words in between are counted. Now the mean distance can be computed and plotted. Sorry if I’m wrong, but to me this seems to be a semantic way of analysing the data and is only sideways linked to content. Maybe the result is that ‘woman – pudding – bad’ is a strong combination opposite to ‘man – pudding – fat’, but the meaning of it is not very clear. Besides, a lot of data has to be used and the type of analysis is more quantitative than qualitative.
A way to analyse text and to reduce bias is to use a special technique.
It contains 7 steps:
Transcribe all the spoken text on the voice recorder to a written text.
Delete all irrelevant information.
Split the text into fragments of topics.
Label every single fragment.
Make an overview of the labels and re-order them systematically in a table of contents.
Replace the fragments under the labels of the table of content.
Rewrite the text fragments to an easy to read text.
The technique can be used top down – the table of contents was made before the interview was taken – or bottom up – based on the topics during the interview the table of contents is made.
More details can be found in our paper How to analyse data from interviews.
Many more types of research can be done, like comparing groups, changes in time due to evolution or due to an incident. As a matter of fact, all six types of research designs can be applied.
To apply scientific comparisons with qualitative data, five steps must be taken:
The first step is always: check your data. If the text is not transcribed correctly or if the respondent withdraws his statement, then try to fix it. You need to have a good basic text to compare the statements of the respondents.
The second step is to describe the response. You have to find out if the persons who cooperated in the study are more or less representing a community or organisation. Also check if all subgroups are represented.
The third step is testing your instruments. The ideas or vision you want to compare are usually summaries or interpretations of the original texts. It is better to check and double check whether these summaries are correct and interpreted well.
Step four is: do the analysis to answer your research questions. Now the texts of different respondents (or response groups) are compared. What do they agree with and what do they disagree with?
Step five is executing deep comparisons. A lot of statements can be compared, but now it is time to look for deeper backgrounds. The way things are said can come from a different view. If two (groups of) respondents have the same goal, but still do not trust each other, then try to find out why.
These five steps are rather similar as in a quantitative research.