Size and layout of e-commerce product grids: a user research case study
Hugo Jenkins

Very nice write up! I’m curious if you were using the standard average, or a geometric average.

I ask because I’ve just been doing a lot of task-time research myself, and it seems that the geometric mean gives consistently more accurate averages of sample sizes <= 25. (This is because it uses a log transform on the data, and time-task data is almost always positively skewed.)

Show your support

Clapping shows how much you appreciated Carl Pearson’s story.