[harrybr]

A great example of the wrong way to report user research.

- How many participants were tested? - Who was tested? - What kind of test was it? Survey? Lab study? etc? - Between subjects or within (i.e. did the same people see all the icons)?

I could go on.

[gus_massa]

I guest that the number of participants is 99. (I don't have the real data.)

                |  Yes U.    Not U. |   Yes U.    Not U.
  --------------+-------------------+--------------------
   Bold         |  99/99      0/99  |  100.00%     0.00%
   Italic       |  99/99      0/99  |  100.00%     0.00%
   Underlined   |  99/99      0/99  |  100.00%     0.00%
   Strike-t.    |  54/99     45/99  |   54.55%    45.45%
   L. Chain     |  35/99     64/99  |   35.29%    64.71%
   L. Globe     |  25/99     74/99  |   25.00%    75.00%
   Ordered L.   |  99/99      0/99  |  100.00%     0.00%
   Unordered L. |  93/99      6/99  |   93.94%     6.06%
   Image        |  78/99     21/99  |   78.79%    21.21%
   Undo         |  57/99     42/99  |   57.58%    42.42%
   Redo         |  51/99     48/99  |   51.52%    48.48%

PS: Your decimals leak information. :)

[epo]

None of which really matters, at all. It's a blog post not a journal paper or a research report. It would have been nice to see a link to a full account but perhaps they couldn't decide how to represent that.

[harrybr]

Omitting these details turns it into "a fun discussion". Including these details (even as a footnote) would turn it into a potentially useful piece of research.

[calpaterson]

Apparently it was his bachelors thesis though, so he's (hopefully) written to academic levels of detail before before...