Social psychologists in particular are fond of what Lee Ross once called "demonstration experiments," that is, experiments that show in more or less dramatic fashion the weird things you can get people to do if you try hard enough. Even social scientists have derided demonstration experiments as goofy, sometimes useless, but they have their value.
The most famous of these is the Milgram experiment, in which psychologist Stanley Milgram got people to electrically shock somebody in another room simply by telling them to do so. (Don't worry. The person in the other room was just a tape recording.)
Scary, isn't it? Milgram had actually set out to show that we Americans simply weren't capable of being monsters — we were supposed to be the control group, and bad Germans were the treatment — and he found out that all it took to turn us into monsters was a simple, calm command.
Here's something scarier: the bizarre plethora of things that affect your vote without having the slightest thing to do with politics. Facial similarity between voters and candidates, the subtle effects of grammar on candidate perceptions, the fact that how high on a ballot a candidate appears affects your likelihood of voting for that candidate, and this one, which shows that if you happen to be squeezing a foam stress ball, you're more likely to think a random stranger is a Democrat.
Scary? Or just weird? Yes, it is weird that a stress ball could affect your political perceptions, but it's scary, because it really, really shouldn't. For all those who actually care about The People making good decisions, this is important: you've got an uphil battle, and one that you don't totally understand.
Scarier still: there's a long-running effort to stifle this kind of research because it's weird, especially when it's weird and has something to do with politics. The Washington Post has a nice editorial on weird but important natural science research, but let's remember the importance of the weird in social science, too.