Sunday Links #9

Peter Singer discusses in the New York Times about why giving to charity is important and how much one should give.

You may know about The Lean Startup. But can this apply to non-profits? Probably yes?

“The Four Habits that Form Habits”. If you want to learn how to form good habits, you need to start from base habits that will support you in your quest for future good habits. Otherwise, you’ll constantly derail and give up.

It’s difficult to give good advice about whether to go to graduate school or not, but here’s good advice on avoiding bad advice.

Five stories on the power of reinforcement. In interactions with people, one should pay attention to reinforcement, or the idea that rewarding people for certain behaviors actually does increase that behavior.

The overlap effect to skills is kind of interesting. The idea is that certain skills share common traits, meaning you probably could speed up learning two different things if they are sufficiently related (like French and Spanish, not French and mountain climbing).

Most of the time, I think, government social services helping the poor go right. But sometimes, they go wrong, like by incentivizing parents to pull their children out of literacy programs because the parents are afraid of losing government support for raising “disabled” children.

The United States spends a lot of money on preventing terrorist attacks. But how cost-effective is this spending? It’s really hard to tell, because attacks that are prevented never happen, and therefore don’t get counted. For every attack that is foiled, dozens more might have been deterred entirely, and never started. However, there’s one thing we can calculate – the expected damage from terrorist attacks compared to our spending on preventing them. So how many terorrist attacks would we have to prevent or foil in order to make our big spending worth it? One thousand, six hundred, sixty seven.

People blame gerrymandering a lot for why the US Democratic party tends to underperform it’s demographics – winning a majority of the popular vote but a distinct minority of the seats in the House of Representatives, for example. However, the evidence for gerrymandering is weak and it’s a lot more likely that Democrats are “unintentionally gerrymandered” [PDF] because their core voting population are concentrated in large cities. Though, further thoughts reveal a bit stronger of a case for gerrymandering than first thought.

I’m probably overusing quantification and should quantify with more care. Repeat after me: (1) one counter-intuitive study at 95% confidence doesn’t mean I should take that study at face value and assign 95% probability to it’s conclusion, (2) studies that seem to good to be true often are, (3) focusing too much on a correlation can derail it’s effect, (4) beware of needless complication, (5) don’t use big samples when small samples work fine, and (6) just because it sounds cool and impressive doesn’t mean it’s better than a simple and dumb solution.