Sunday, May 1, 2016

Back to Changing Society: the Dubious Effects of Progressive Agitation

Here's a round-up of some disturbing findings from social psychology that one has to consider when thinking about any plan to change society.

On complaints that tests are biased against minorities:
We know exactly what happens when minorities are told tests are biased against them: they do worse on those tests. This is the essence of the idea of “stereotype threat” – for example, one can improve women’s performance on a math test simply by telling them that the test is not biased against women. So maybe we should stop doing exactly the thing that we just proved hurts women and minorities’ educational performance.
On educating children away from drugs:
nearly every study on DARE programs has found that they increase drug use, sometimes as much as 30%.
On "sensitivity training":
A comprehensive review of 31 years of data from 830 mid-size to large U.S. workplaces found that the kind of diversity training exercises offered at most firms were followed by a 7.5 percent drop in the number of women in management. The number of black, female managers fell by 10 percent, and the number of black men in top positions fell by 12 percent. Similar effects were seen for Latinos and Asians.
Besides, just talking to people about diversity makes them more likely to think in terms of racial and ethnic categories.

Plus, calling people racists makes them more racist. In fact all attempts to change people's attitudes run afoul of things called the "boomerang effect" and the "backfire effect." Basically, people hate to be criticized, and the main thing that happens when you criticize them is that they get defensive and find ways to justify their beliefs and actions.

It should of course be said that this is all social psychology, so none of these findings need be taken very seriously. There are studies that purport to show the opposite. On balance, though, the findings that show these reverse effects are stronger and better attested than anything showing that these measures help. There is simply no evidence that complaining about rape culture reduces rape, or that fulminating against racism decreases racism.

On the other hand there has been a measurable decline in both sexism and racism over the past 70 years, so these things can change. Maybe measures that fail in the short term have some effect on a time scale of generations. Or maybe the changes are being driven by forces that have little to do with moralist hectoring.

7 comments:

Unknown said...

Perhaps a more effective way to change behavior is to hold up examples of the behavior you want to discourage, with the implied message, "you don't want to be like that, do you?"--and perhaps with a bit of the message also, "Aren't you glad you and I aren't like that?" This may be especially effective at the grade-school level.

Unknown said...

Another effective method would be to tell stories that promote identification with the victim and blame on the perpetrator. Again, most effective probably on the grade-school level. These sorts of things at this level might promote a base line of shame and restraint that could be most helpful.

Unknown said...

A third thought: both moralistic hectoring and crude legal changes can have effects over the long term, as you suggest. This seems to be a large part of what happened with both smoking and seat belt use.

Then again, hectoring, package warnings, and the constant reportage of statistics has not greatly changed the American diet. The percentages are probably up for healthy eating (since, say, 1970), but still not extremely high (and I say this as a person whose diet is quite unhealthy).

I suppose the question is, why has moralistic hectoring worked with some things and not others? What's the differential?

John said...

That's a great question. When I look at something like seat belt use, or the decline in drunk driving, I see several trends coming together. You had tougher laws, and stepped-up police enforcement, and hectoring by groups like MADD, and plenty of publicity around the actual dangers of driving drunk, and the generally declining tolerance for risk in the country as a whole.

Compare this to attempts to stamp out marijuana use, where the hectoring was not really supported by scientific research, and legalization found support among both libertarian-leaning westerners and liberals nostalgic for the 1960s, and the country was turning against imprisonment for non-violent offenders, and so on.

What I see is that hectoring in itself is not likely to change anything and may very well hurt, but maybe it can support social changes for which there is also other pressure.

John said...

As for healthy diet, it's just hard to generate the kind of social forces that have curtailed smoking and drunk driving. Your eating just doesn't hurt anyone else, so there won't be any support from that side, and the culture seems to be squeezing the time for sit-down meals and pushing us toward gobbling donuts on the way from one appointment to another, plus the genetic draw of sweet and salty foods is strong.

pootrsox said...

Re diet: long article in NYTimes about how losing large amounts of weight simply sets you up to gain it back, b/c your body, which has learned to burn calories more slowly, simply never goes back to a "normal" rate of caloric burning. So to maintain weight loss over years (the study was 6 years long), people would have to eat many fewer calories than their weight would normally be able to burn.

Not sure what hectoring and moralizing can do to change that :(

szopen said...

Huh, while I think all stereotype threat studies are basically junk science, there is one thing about "backfire effect" - I know that I was far less conservative and extremist before I went into facebook, where I got in touch with many former friends from US. They were mostly liberal, they were shocked how bigoted, racist and sexist I was and after two years of moral preaching they completely succeeded in convincing me that indeed I am f* bigot, racist and sexist, who would rather die than allow their kind to ever invade Poland's universities or politics.