Qualitative data synthesis/analysis
Methods for what you do to move from raw data (like interviews or observations) to conclusions.
Factors to balance when planning synthesis and analysis
When we’re conducting synthesis and analysis, we try to balance:
- Inclusivity. How much can people other than researchers participate in analysis?
- Rigor. How thorough is the analysis process? How likely are we to discover surprising patterns instead of repeating our own assumptions?
- Speed. How long does it take to do the synthesis analysis? Are the results ready by the time my team needs them?
Bad ways to do analysis
- Gut feelings. It’s good to ask people for their hot takes after research sessions. But that shouldn’t be the end of the analysis. Gut feelings are particularly sensitive to bias. Hot takes often reflect our own assumptions more than the data.
- Finding a set of categories or findings that encompasses all the data. You may not be able to find a set of categories that explains everything.
Basic steps of analysis and synthesis
In any tool (Trello, stickies, or spreadsheets), in any place (online, offline), the steps are the same.
- Remind yourself of your research question. Go back to your research plan and remind yourself of the question your study is trying to answer. Conduct your analysis and synthesis to discover answers to that question.
- Read over your notes and split them into bits (like quotes, behaviours or paraphrases). Put them each on a sticky note, row or card of their own.
- Categorize the bits (first round). Look through the bits and start putting them into categories. Maybe you’ve come up with the categories in advance, or maybe you’re coming up with them as you go. Try labeling each category.
- Split or combine your categories. If you have lots of categories with a few bits in each, figure out ways to combine them into bigger categories. If you have a few categories with many bits in each, figure out ways to split them out. Repeat the process until the categories are useful answers to your research question. Label categories as you go.
- Scrap everything and try categorizing again. Forget all your previous categories. Recategorize your bits again. You may find a new way of thinking about your data.
Increasing speed, rigor or inconsistency.
- To go faster, you can do fewer steps of these steps. You can stop after step two, three or four.
- To be more rigorous, you can do more of these steps. The more you do, the more likely you are to find suprising answers.
- To be more inclusive, you can include your team in more of these steps.
- You can’t have all three. You have to pick two, or do a moderate job of both.
For ideas about how to balance these three, try the analysis/synthesis sequence ideas spreadsheet. You can select how much you care about each factor, and it will suggest some possible analysis/synthesis steps.