Card Sort Study
Prepare cards for a navigation or information architecture study, then create one session per participant. Each participant names their own groups and puts every card in exactly one group. Only completed sessions count in the pairwise co-group matrix. Explore the fictional library example to see both completed and unfinished sessions. Save the versioned JSON yourself before leaving this tab.
Key features
- Stable card IDs and separate editable participant sessions
- Participant-defined group names and an explicit assignment for every card
- Completion gate for missing groups or unassigned cards
- Deterministic co-group count, completed-session denominator and percentage for each card pair
- Strict JSON import and formula-safe session and matrix CSV exports
How to use
- Load the fictional example or start a new study and add cards before the first session.
- Create a participant session, name that participant's groups and assign every card once.
- Complete the session only when no card is unassigned; repeat with other participants.
- Review each pair's co-group count over the number of completed sessions. Drafts do not enter the denominator.
- Download JSON for re-editing and CSV files for session records or the matrix.
Use cases
- Explore where readers expect website navigation items
- Compare multiple users' categories for a product menu
- Find card pairs that participants often put together
- Share raw session assignments and the aggregate matrix for further analysis
Frequently asked questions
What does 2/3 in the matrix mean?
Two of the three completed participant sessions placed that card pair in the same participant-defined group. The denominator is every completed session, because every card must be assigned exactly once to finish.
Do unfinished sessions change the score?
No. Draft sessions are excluded entirely. Editing a completed session reopens it as a draft, removes it from the denominator, and increments its revision. Complete it again to include the revised result.
Are group names matched between participants?
No. Groups belong only to one participant's session. Similarity compares whether each pair shares a group within each session, even when participants use different group names.
Can I change cards after participants start?
No. Card IDs and labels are frozen while any session exists, so old answers cannot silently change meaning. Remove all sessions first if you need to change the card set.
Does this replace research analysis or store private results?
No. The matrix describes co-group frequency, not statistical confidence, clustering or a recommended taxonomy. Data stays in tab memory and is not autosaved or sent to a server. Download JSON before closing.
Privacy
Browser-local, tab-memory study. No autosave or server submission. The example is fictional. Avoid putting real participant identities or secrets into exported files. This is descriptive co-group frequency, not statistical significance or an automated taxonomy recommendation.
Comments & questions