Verified by 
Not 100% science fiction after all
How good are politicians at keeping their campaign promises?
Most people think the answer is somewhere between zero and deliberate dishonesty just to annoy them personally. Politicians will say ANYTHING to get elected. A politician's promise is about as good as Lindsay Lohan's promise to show up in court on time. Right?
But the real truth is somewhere closer to Mr. Smith Goes to Washington than The Manchurian Candidate.
A recent article by my favorite political scientist, Jonathan Bernstein, points out that the consensus in his field is that campaign promises are a very good indicator of what policies a president will enact. Obama promised to make health care a priority, and he did. George W promised testing-based education reform, and then did so. And remember when George HW said? "Read my lips. Brussels Sprouts will not become the national vegetable."*
[* Don't worry. It's still the cauliflower].
In one major study Bernstein discusses, politicians kept 75 percent of their promises. And the ones they failed to keep were prevented by obdurate congress, rather than by the president maniacally laughing and saying, "those suckers believed every word!! Muahahahaha."
Which means, watching the Republican debates is not just a fun excuse to take a shot of Maker's every time you hear "job creators," "Obamacare," or "Reagan." It actually tells you what these people will try to do when they control the reins of power.
The real question is why so many Americans are so skeptical of political promises. I think the answer is that broken promises are so much more salient than fulfilled ones. Once troops leave Iraq, then we mentally mark that as done and forget that it was a campaign promise—we just think of it as one of Obama's things that he did (whether we agree with it or not). But when George HW promised no new taxes and then raised taxes anyway, that's highly salient to us.
Psychologists have long been fascinated by peoples' innate abilities (or inabilities) to estimate simple statistical patterns in events they experience in real time. It turns out we do alright, but we have a lot of biases that impact our judgments. One of the most glaring ones is that we tend to over-weight salient events in our estimates. That's moderately interesting when it applies to undergraduate study participants guessing the next number on a computer screen. But it's of global importance when it comes to the action of 300 million participants choosing the next president.
In the election of 2000, many people believed that the only difference between Gore and Bush was the color tie they preferred. We now know that sense wasn't accurate, but, as Bernstein points out, we should have known it, because the Red and Blue teams made very different sets of promises.
Over then next 10 months, we'll hear a lot of promises from everyone who is running for office, from president down to dog-catcher. Don't just screen them out. They are useful predictors.
