So you’ve decided that you need some user feedback on how to organize your information, and you think that a card sort sounds like the way to go. Now what?
There are different ways to set up a card sort that will affect the kinds of results that you get. How to set it up depends on your situation and your goals. The following is a list of the most common types of card sorts to run and the best times to use them:
Open
The most fundamental of card sorts, an open sort simply provides the participant with a list of items to be sorted into groups/categories. The number of categories, labels for the categories, etc. are all up to the participant. This type of card sort is by far the best place to start when trying to obtain user feedback on the organization of your items (whether for a website or for general information architecture purposes). Even if there is an existing set of categories for the items, beginning research with an open sort ensures the best quality user feedback (see Card Sorting: Current Practices and Beyond).
There are some pitfalls to be aware of. First of all, when asking participants such an open-ended question, you have to be prepared to get a lot of different answers. To make sure that the information you get is going to be what you need to make good decisions, you need to be very clear in your instructions what the study is for (website menu structure? retail catalog organization?) and you need to cull good participants. Think about who is going to use what you are trying to build and ask them. Don’t just post a link to your study on Facebook and think this is going to do the trick (unless your target demographic is your set of Facebook friends, of course).
One last thing – don’t get too many participants. This may sound odd, but it is backed up by some research. Also, each participant is going to submit unique results. Names of categories don’t often match exactly, items may differ widely in where they are put – in short, you will have a lot of information to sift through that isn’t going to be quickly assimilated. The more participants you have, the more data you will have, and it goes up exponentially. If you’ve worked hard to have quality instructions and quality participants, you won’t need all that many of them to get some good data. In this case, less is more.
To conduct this type of sort using the WebSort interface, simply do not enter anything into the Categories tab of the Study Manager.
Closed
Unlike an open card sort, a closed card sort does not allow participants to make any decisions about the categories into which the items are sorted. Instead, the study administrator provides a pre-existing set of labeled categories into which the participants are asked to place items. While we strongly caution against using a closed sort without first obtaining user feedback with an open sort, a closed sort is excellent for validating results obtained from previous user-based research. It is also useful when adding new items to an older, previously user-validated structure.
To conduct this type of sort using the WebSort interface, enter in the desired labels/titles of the pre-defined categories in the Categories tab of the Study Manager. Be sure to uncheck the ‘Allow creation of new categories’ box at the bottom of the Categories tab.
Mixed
This type of card sort combines the flexibility of user-defined groups and names with the ability to present the participants with a pre-existing set of labeled categories. Participants in a study are given a set of administrator-defined categories, just as in a closed sort. However, unlike a closed sort, participants can rename the groups, delete them, and create new ones.
Some researchers consider this type of sort a ‘guided’ open sort, as the existing categories provide information to the participants as to how the items might be perceived or how an existing organizational structure may be set up. A mixed sort is very useful when you have new items to be added to an existing structure that is old enough to need fresh feedback. And when combined with clear instructions, it can be used when you want to limit the participants to a certain number of categories but allow them to provide the names for them.
To conduct this type of sort using the WebSort interface, enter in the desired labels/titles of the pre-defined categories in the Categories tab of the Study Manager. This time, be sure to check the ‘Allow creation of new categories’ box at the bottom of the Categories tab.
While there are other types of card sorts out there (which we will post about in the future), these are the tried-and-true methods that provide the foundation of good card sorting. And don’t forget – you can try them all out for free at WebSort.net. 
