Gurwinder
@gurwinder.bsky.social
Examining how technology is affecting human psychology, and how to stay rational in the digital age.
Join 75,000 readers: gurwinder.blog
Join 75,000 readers: gurwinder.blog
18. Simpson’s Paradox:
A trend can appear in groups of data but disappear when these groups are combined. This effect can easily be exploited by limiting a dataset so that it shows exactly what one wants it to show. Thus: beware of even the strongest correlations.
A trend can appear in groups of data but disappear when these groups are combined. This effect can easily be exploited by limiting a dataset so that it shows exactly what one wants it to show. Thus: beware of even the strongest correlations.
October 30, 2025 at 5:19 PM
18. Simpson’s Paradox:
A trend can appear in groups of data but disappear when these groups are combined. This effect can easily be exploited by limiting a dataset so that it shows exactly what one wants it to show. Thus: beware of even the strongest correlations.
A trend can appear in groups of data but disappear when these groups are combined. This effect can easily be exploited by limiting a dataset so that it shows exactly what one wants it to show. Thus: beware of even the strongest correlations.