3 ms·

If you read between the lines, it's clear the methodology used to extrapolate this data is shaky at best, and borders on "you've never taken a statistics class

If you read between the lines, it's clear the methodology used to extrapolate this data is shaky at best, and borders on "you've never taken a statistics class in your life" at worse. Complete garbage... Favorite snippets: " The paper further cautions that reliable depression surveys don’t even exist for some low-income countries -- a common issue with global studies -- forcing the researchers to come up with their own estimates based on statistical regression models." My interpretation: "There was no data so we guessed". Terrible.

greenyoda 13y ago
What's even more depressing (no pun intended) is that this kind of junk science is published by the Washington Post, one of our nation's most highly esteemed newspapers.
tgb 13y ago
It really annoys me to see people judging studies based off of press coverage of them. If you're going to accuse it of being shoddy, at least take five minutes to read the paper. I'm not saying you're wrong, since I just spent five minutes reading the paper and they do use an modelling tool in the way that you suggest (and also use it to combine multiple reports from over the years or reports with only partial information). It's just that it's very easy to see something and wisely claim "Oh I bet they're making this laughable mistake based off of some vague claim in an article I'm criticizing" which is kind of harmful. I've said things in the past like "I read a study that X" and immediately the other person was like "Oh but that's just correlation", which is a good thing to think but seriously, it's not like the people writing these studies don't know that correlation-isn't-causation and might have thought to address that. I hadn't mentioned any methods used in the study because I didn't recall any of them, yet the gross assumption that it was wrong was still lept to. So if communities like Hacker News want to ever believe any science report ever (and they probably do), then I really think people should slow down and at least give the study the benefit of the doubt. You shouldn't have to read between the lines - the actual study is free on PLOS. If you don't have the time or expertise to read the paper, just pose this complaint as a question instead and probably someone else here will do the work for you.

greenyoda 13y ago

What's even more depressing (no pun intended) is that this kind of junk science is published by the Washington Post, one of our nation's most highly esteemed newspapers.

tgb 13y ago

It really annoys me to see people judging studies based off of press coverage of them. If you're going to accuse it of being shoddy, at least take five minutes to read the paper. I'm not saying you're wrong, since I just spent five minutes reading the paper and they do use an modelling tool in the way that you suggest (and also use it to combine multiple reports from over the years or reports with only partial information). It's just that it's very easy to see something and wisely claim "Oh I bet they're making this laughable mistake based off of some vague claim in an article I'm criticizing" which is kind of harmful. I've said things in the past like "I read a study that X" and immediately the other person was like "Oh but that's just correlation", which is a good thing to think but seriously, it's not like the people writing these studies don't know that correlation-isn't-causation and might have thought to address that. I hadn't mentioned any methods used in the study because I didn't recall any of them, yet the gross assumption that it was wrong was still lept to. So if communities like Hacker News want to ever believe any science report ever (and they probably do), then I really think people should slow down and at least give the study the benefit of the doubt. You shouldn't have to read between the lines - the actual study is free on PLOS. If you don't have the time or expertise to read the paper, just pose this complaint as a question instead and probably someone else here will do the work for you.