Once you've run card sorts, you need to be able to organize and understand the data collected. In this video. learn to break down and organize data so you can look for patterns and insights.
- Once all the session are complete, you'll have to tackle organizing all the data and synthesizing insights that the team can use moving forward. There are several ways to analyze card sort data. And your process will be slightly different, depending on what type of test you've ran. Regardless of the type of card sort, your first focus will be to organize everything you've collected. I'd recommend using a single spreadsheet to enter all the data. Donna Spencer has a brilliant template that she's made available, and a detailed set of instructions for organizing card sort data in a spreadsheet.
Essentially, she suggests creating a tab for each participant that has one column for the card names, and another for the category or group that they put the card in. Donna's spreadsheet has a formula built-in to summarize each of the participant's results by card. If you've ran an open or hybrid study, where participants suggested their own labels, one important thing to do is identify and merge very similar categories. Participants may have conceptualized data similarly but come up with slightly different names or misspelled things.
For instance, think about a card sort of menu items for a restaurant. One person may have a category labeled small plate. Whereas another person may have the same items together but labeled it small plates. For purposes of understanding how people group information, you can consider both those groupings as a single category and combine them for further analysis. However, if there are clear differences in language for similar groupings, like if another participant labeled that same group starters, you'll want to keep those categories separate for analysis.
Neither label is right or wrong, but you'll want to do some further digging into how closely all the items are related. I also like to add a single summary tab that lists all the cards in individual rows, and the major categories across the top as columns, with a count of how many times each card appeared in the category in the matrix. This view makes it easy to quickly spot trends where there is high agreement or where the participants diverged greatly. For instance, check out the example summary seen here with 30 total participants.
You can quickly see that the card popcorn was put in the snack category 27 times, meaning that there was high agreement between participants, and snacks is likely popcorn's logical home. On the other hand, the fruit card was almost equally placed in snacks and breakfast, so you'll need to do some more digging there. This kind of high-level view is a great starting point. But you'll likely want to dig deeper. As mentioned, Donna's spreadsheet has some built-in formulas to begin more detailed analysis. You can also utilize other statistical analysis methods, such as multi-dimensional scaling or cluster analysis.
But let me tell you a secret, all the digital card sorting tools that I know of, will tackle much of the statistical analysis for you. I recommend using one of these tools unless you really love and are really good at manual statistical analysis. They'll save you tons of time. If you use the digital tool to begin with, the tools will automatically do some calculations for you. You should still go through each participant's responses to clean up data and pair any reasonably similar groupings. If you did a physical card sort and entered the data into a spreadsheet, I'd recommend adding that data into a digital tool for the analysis.
Although the data entry can be a bit tedious, it saves time in the long run. These tools will do a variety of types of analysis for you very quickly.
- What's card sorting?
- Open, closed, and hybrid card sorts
- Card planning
- Category planning
- Finding, selecting, and screening participants
- Scheduling and incentivizing participants
- Running sessions