This is a really good point. An interesting psychology experiment could measure the susceptibility of subjects to stereotype threat after they read a given newsfeed. I hypothesize that because of the "priming" effects of each newsfeed, subjects who read the same set of tweets would score similarly on the stereotype test. This would be in like with the results of the famous John Bargh study where subjects subconsciously decreased their walking speed, for instance, after reading words related to the elderly.

It's a scary thought that the opinions of society can be clustered by similarity of Twitter newsfeeds. That sounds ripe for manipulation and propaganda dissemination.