Card Sort cluster analysis tool 1.1 results

Processing completed Thu, 09 Sep 2010 05:44:47 +0300. Processing took 0.15 seconds.

Please send feedback to aapo.puskala@userpoint.fi. Back to front page.

Users

User feedback

• No feedback was sent.

Cards

The following cards were used:

• 1. car
• 2. motorcycle
• 3. bicycle
• 4. bus
• 5. boat
• 6. airplane
• 7. private jet
• 8. zeppelin
• 9. rollerskates
• 10. subway
• 11. tram
• 12. taxi
• 13. helicopter
• 14. segway
• 15. ship
• 16. train
• 17. feet

Similarity matrix

Each card is compared to each other card. The higher the number, the more similar the cards. The maximum score is 9.

Dendrogram

The dendrogram shows a hierarchy based on the groupingsmade by the respondents.

SVG format dendrogram. Try hovering your cursor over the green labels.

Note: If you use Internet Explorer, you need to install Adobe SVG Viewer (no longer supported) to see SVG files. Most likely you also need to install a hotfix and to modify the registry. To download the hotfix, you need to give them your email address as the hotfix is password protected. After 2 hours of struggling I still could not get Internet Explorer 7 to work with embedded SVG images. My warm recommendation is to trash Internet Explorer and switch to a better browser, such as Firefox. It will take much less time and nerves. SVG was recommended by W3C in 2001 but as of January 2009 Internet Explorer still does not support it, being the only major browser not to.

JPG format dendrogram

Clusters based on similarity

In the table below the cards are divided into different number of groups based on card similarity.

List of clusters

Cluster ID = the ID number of the cluster

Contains card IDs = a list of all cards in the cluster. The card ID is the same as the cluster ID (first column).

Score = the average number of pairings between the cards in the group. If three users have participated, and two of the users have grouped two cards together, the score is 2 (out of maximum of 3). The higher the score, the more similar the cards within the cluster. The maximum score is the number of participants (users), 9 in this study. If there is only one card in the cluster, the score can not be calculated as there can be no pairs of single cards.

Match = score / number of participants (9). A percentage indicating how similar the cards within the group are. If all participans have grouped to cards into the same group, match equals 100%.

Member of = the ID of the 'mother' cluster, i.e. the cluster to which this cluster belongs to

Name(s) = the top three best matching user-given name for each cluster