Comments on Analyzing Card Sort Results with a Spreadsheet Template

Joe Lamantia — Fri, 08 Jun 2007 19:05:39 GMT

Nice work Donna – and thanks for the acknowledgement!

Thanks also to Kate for a great set of suggestions on how to improve the maigc bit of the exercise – refining and normalizing the categories people create. I’ll roll those updates into the master template that’s linked from my site.

Donna Spencer — Fri, 07 Jan 2011 19:39:12 GMT

I have released a free spreadsheet I use to analyse card sort data. My spreadsheet is great for analysing open sorts – it manages participant results, helps you explore the data and does some basic statistical analysis. It is based on Joe’s spreadsheet described here. Thanks Joe!

Kate Walser — Fri, 07 Jan 2011 19:41:07 GMT

Joe, very helpful – thanks for the template! Couple suggestions for demystifying the “magic” very beginning step with the “super raw” data and also putting the cards in an easier-to-handle-and-tote electronic format.
1. Create a new tab (super_raw) and enter each user’s cards in a column. Use yellow to highlight suggested level 1 topic categories (and blue for level 2 if you go that deep, purple for orphans, etc.). (You can set up macros to automatically make the cell yellow, for faster entry)
2. Optional: Create another tab (super_raw_lookup) that’s a copy of the super_raw tab, and in place of the actual card, use a lookup (=LOOKUP) to reference the actual card’s label, for easy reference.
3. Create a copy of the super_raw tab and call it User_Categories. Get rid of all the non-yellow cells (I created a macro for this step – wasn’t coding worthy of an A+, but got the job done), and all you have left are the users’ suggested Level 1 categories.
4. Sort the cells in each column alphabetically and patterns start to emerge a bit.
5. Start moving around related cells and you can see even more trends.
... Once you get your categories set, you can copy/cut & paste the super raw data for each category and stick in the ‘Raw Data’ tab under the appropriate headings.

It may sound like a lot of work, but that initial up-front time to enter the data (especially if you can have someone else do some of that data entry while you jot down your qualitative observations) is well worth it as you sift through the electronic stacks later.

If you’re interested in presenting the different categories where cards appeared, you can do more on that front too and concatenate the category names and have them appear in your summary tab.

Hope the additions help!

Cheers, Kate

Grace Stoeckle — Tue, 01 May 2007 19:25:56 GMT

Joe—Some of my coworkers (who ARE Excel wizzes) figured out my error and fixed it. It’s working perfectly for me now. Thank you again for this excellent resource!

Grace Stoeckle — Tue, 01 May 2007 18:25:28 GMT

Hi Joe—This is such a terrific resource! I’m also glad to see that you’re still responding to questions/comments.

I have spent several hours entering data into this template spreadsheet, but I’m not getting results on the “Card Placement %” or the “summary” worksheets. All the cells have the same thing in it: ”#REF!”

The rest of the worksheets are looking good, but those two are not. Any ideas? I’m by no means an Excel wiz, but I was extremely careful not to delete important cells or mess with any formulas. I even attempted to start over in a fresh template and re-enter all the data but with the same results on the last two worksheets.

Any help you can offer would be VERY appreciated!

Thank you. Grace

Joe Lamantia — Tue, 24 Apr 2007 16:16:09 GMT

Linda – “the number of different cards per category” shows how many unique cards (not counting repeated cards) were placed in each category. This is different than the total number of cards (which counts repeated cards) in the category, because the same card may appear in any one category more than one time.

The ratio of total cards in a category (counting repeated cards) to unique cards (not counting repeated cards) helps make clear whether participants ‘agreed on’ the cards that belong in the category. Categories in which many participants placed the same collections of unique cards, and where there were many instances of repeated cards, are categories that you’ll want to be able to spot and (perhaps) preserve when you’re designing your structure.

Does this help?

Cheers,
Joe Lamantia

Linda Renuka — Fri, 07 Jan 2011 19:40:51 GMT

Can you pls let me know what you mean by ‘the number of different cards per category’ in the low & high card count section. i tried to find out what it meant but couldn’t find what it meant. Help pls

Thanks
Linda

Joe Lamantia — Fri, 26 Jan 2007 14:53:42 GMT

Javed,

One way is to conduct several rounds of sorting exercises. In the first round, have participants define the first two sets of categories. In the next round, have participants define additional categories, using the categories from the first round of sorting exercises as a reference.

You could also use one of the other tools available for card sorting (some are mentioned here: http://www.rosenfeldmedia.com/books/cardsorting/, and in other articles in this magazine). Several of these tools support statistical analysis for category structures with more than two levels depth, using clustering and other algorithms.

My spreadsheet is meant to be lightweight and simple, so does not offer such capability yet.

Cheers,

Joe Lamantia

Javed Anjum Sheikh — Fri, 26 Jan 2007 14:53:42 GMT

I have category, sub-category and sub-sub category, how can I manage them? It would be appreciated if you advice me in this regards.
1 Meat
1.1 White Meat
1.1.1 (10) Chicken
1.1.2 (36) Turkey
1.2 Red Meat
1.2.1 Halal
1.2.1.1 (05) Beef
1.2.1.2 (20) Lamb
1.2.2 Non-Halal
1.2.2.1 (30) Pork
1.2.2.2 (02) Bacon
Waiting for expert advice

Douglas Potts — Fri, 26 Jan 2007 14:51:40 GMT

Angelina,
The idea of the “Standardized Categories” is that you’ve created a category name for all of the raw categories that are similar. For example, if you have three raw categories, such as “Email Us”, “Contact Info”, and “How to get in touch with us”, you will create a Standardized Category called, for example “Contact Us”. You then place those three Raw Categories into the one Standardized Category. At this point in a cluster analysis, you are not interested in exact labels, but the the topic of the labels.

Timothy Pedersen — Tue, 11 Jan 2011 03:39:48 GMT

Check out this free beta software from IBM. EZSort can help you perform and analyse card sort data. http://www-3.ibm.com/ibm/easy/eou_ext.nsf/Publish/410

Mary Rackow — Fri, 26 Jan 2007 14:51:40 GMT

We are very new to the concept of IA and user experience/testing in general. I would like to take a group who is incharge of migrating our content into a new cms thru the card sort exercise.

Does anyone have an EASY way to do this? By easy, I’m easy to explain, and where to start. It’s a new concept, so I’ll be doing some selling in conjunction with explaining the exercise.
Thanks!
Mary

Joe Lamantia — Fri, 13 Apr 2007 09:12:14 GMT

The .xls file is available from my site at this link:

www.joelamantia.com/html/projects/card_sort_template_ba.xls

Enjoy!

Joe Lamantia — Fri, 26 Jan 2007 14:51:40 GMT

The .xls file is available from my site at this link:

www.joelamantia.com/html/card_sort_template_ba.xls

Enjoy!

Todd Warfel — Fri, 26 Jan 2007 14:51:40 GMT

Unfortunately, the XLS file doesn’t exist on the server. Oops!