How would you even do a randomized study? It’s not like you can pick 1000 people at random and tell them “Thou shalt cycle to work this year” and pick 1000 people at random and tell them “Thou shalt drive to work this year” and count how many of each group survive, and even if you do, the sample would be way too small to draw any definite conclusion, and the duration might be too short to show any long-term health effects of cycling. In some cases you have to resort to correlation studies, even if you can’t tell whether cycling makes people healthy or health makes people cycle or something else does both or some combination of the above.
That’s about the only thing that would ever work, and yes you can do it.
There was a real study where they gave people puppies randomly. There’s no reason why you cannot have a real study of giving people bicycles randomly. You could measure other health outcomes and extrapolate from that at least.
This study is worse than worthless—it has negative worth because it seems to be making some people on lesswrong stupid.
There are almost no correlation studies worth doing. Correlation studies correlate with being worthless really well.
“There are almost no correlation studies worth doing. Correlation studies correlate with being worthless really well.”
You can identify causes from observational studies with some assumptions. Some of these assumptions may be plausible sometimes. Have you ever looked at a paper doing a causal analysis of observational data?
There are almost no correlation studies worth doing. Correlation studies correlate with being worthless really well.
I don’t agree. Identifying correlations gives valuable information both about where to look for causes and how to related to the stuff that is correlated—all else being equal.
Anecdotal evidence has about as much value as correlation studies. It’s inspiration for a real study at best.
Oh, of course. I live my life primarily based on what I learn from anecdotal evidence. For example, if anecdotal evidence tells me that my girlfriend is likely to do X, where X is a particularly undesirable behavior I transition her to ‘ex’.
Funny that you mention it, I did some correlational studies of early stages of dating, since sample size was big enough there to get something of value (and effect sizes were massive). Single blinded even, since none of them had any idea. I’d say it was more valuable than most correlational studies getting published.
Single blinded even, since none of them had any idea.
For some reason I initially read that as “Single blinded men, since none of them had any idea” and I thought you were making a witty jest at the expense of guys that you had dated!
I’d say it was more valuable than most correlational studies getting published.
This is worth emphasizing: I don’t think you know what you are talking about. Observational studies can be very valuable for drawing causal conclusions—if you are careful.
It was a pretty large population—one tenth the population of Barcelona, if wikipedia is to be believed. You’d be hard pressed to get the same coverage out of a randomized study.
It’s an imperfect world, taw. We need all the data we can get.
This is so ridiculously far from a randomized study, I won’t even bother pretending it needs a serious answer.
How would you even do a randomized study? It’s not like you can pick 1000 people at random and tell them “Thou shalt cycle to work this year” and pick 1000 people at random and tell them “Thou shalt drive to work this year” and count how many of each group survive, and even if you do, the sample would be way too small to draw any definite conclusion, and the duration might be too short to show any long-term health effects of cycling. In some cases you have to resort to correlation studies, even if you can’t tell whether cycling makes people healthy or health makes people cycle or something else does both or some combination of the above.
That’s about the only thing that would ever work, and yes you can do it.
There was a real study where they gave people puppies randomly. There’s no reason why you cannot have a real study of giving people bicycles randomly. You could measure other health outcomes and extrapolate from that at least.
This study is worse than worthless—it has negative worth because it seems to be making some people on lesswrong stupid.
There are almost no correlation studies worth doing. Correlation studies correlate with being worthless really well.
“There are almost no correlation studies worth doing. Correlation studies correlate with being worthless really well.”
You can identify causes from observational studies with some assumptions. Some of these assumptions may be plausible sometimes. Have you ever looked at a paper doing a causal analysis of observational data?
I don’t agree. Identifying correlations gives valuable information both about where to look for causes and how to related to the stuff that is correlated—all else being equal.
Anecdotal evidence has about as much value as correlation studies. It’s inspiration for a real study at best.
Oh, of course. I live my life primarily based on what I learn from anecdotal evidence. For example, if anecdotal evidence tells me that my girlfriend is likely to do X, where X is a particularly undesirable behavior I transition her to ‘ex’.
Funny that you mention it, I did some correlational studies of early stages of dating, since sample size was big enough there to get something of value (and effect sizes were massive). Single blinded even, since none of them had any idea. I’d say it was more valuable than most correlational studies getting published.
For some reason I initially read that as “Single blinded men, since none of them had any idea” and I thought you were making a witty jest at the expense of guys that you had dated!
More fun to do too! Well, we can hope.
This is worth emphasizing: I don’t think you know what you are talking about. Observational studies can be very valuable for drawing causal conclusions—if you are careful.
Here’s an example: http://www.biostat.harvard.edu/~robins/publications/Long-term.pdf
There are many other examples.
Evidence is still evidence, no?
This doesn’t count as evidence of anything any more than a study correlating number of pirate attacks with population of elephants in India would.
It was a pretty large population—one tenth the population of Barcelona, if wikipedia is to be believed. You’d be hard pressed to get the same coverage out of a randomized study.
It’s an imperfect world, taw. We need all the data we can get.