Also, to
bring together, or fail to
distinguish,
ideas or
causes, particularly in
statistics and
experimental work. Often happens because, while one group may tend towards having an
attribute which causes
experimental result A, and another group has a different attribute, causing experimental result B, it's the attribute, and not the group membership, that's causing the result. Group membership and attribute are being conflated.
For example (a daft example): more women than men have long hair. If I were to ask a group of men and a group of women whether they use elastic hairbands, I'd almost certainly find a statistically significant difference: more women than men would answer 'yes'. I could conclude that men don't like hair accessories, and that Company X should market them in masculine colour schemes and designs, to appeal to men. I'd be pretty stupid to do so: men are simply less likely to need to tie their hair back. I'd've been conflating gender and hair length.