4 comments

[ 2.9 ms ] story [ 23.5 ms ] thread
One thing they did right with this study and paper was to make the data public, which allowed a better analysis to be published the following year. See https://idl.cs.washington.edu/files/2015-BeyondWebersLaw-Inf...
Thanks, your link is a superior study. The OP is really not very good in the problems it considers. Bottom line of either study seems to be:

"We can give a clear recommendation for designers in the vast majority of circumstances: use scatterplots to visualize bivariate correlation"

Didn't expect to see my paper on the front pages of HN today :)

Lane (first author) here -- happy to take questions on behalf of the co-authors.

xangg already pointed out the nice followup from Matthew Kay and Jeff Heer (https://idl.cs.washington.edu/files/2015-BeyondWebersLaw-Inf...). I'll also direct folks to Ron Rensink's original paper and his several followup papers targeting the underlying mechanisms of correlation perception (https://scholar.google.com/citations?hl=en&user=s_5xXxQXpU0C...)