Why doesn't this exist? If it does, where is it?
I want a tool not unlike gapminder that takes summary statistics from tons of studies and allows me to mash them together.
Traditionally, this has been done with meta-analyses which are studies of studies that compare these outcomes and using different weights and probably other things (you want me to read wikipedia on this?) to come up with some sort of broad conclusion based on all the evidence.
Granted, this may not be as scientifically sound, but hey, everyone's doing homemade ecological studies (these are the gapminder kind) and I don't see anyone cryin'.
Here's the experience (for anyone I haven't lost)
I have a kid who hates to wear his seatbelt (hypothetical here of course). I think what are the actual risks if you don't wear a seatbelt. So I look up the Meta-analysis explorer under seat-belts for kids and I narrow it down to around his age range and I get some statistic like, A kid in the US who doesn't wear his seatbelt is 45 times more likely to die in a car crash. Hm, then I think, what's the chance of having a car crash? And then I look that up and it says "on average in your state 80 serious crashes occur every day. If you are on the road today, you have a 1 in 10,000 chance of being in a crash. If you drive everyday this year, you have a 1 in 400 chance of being involved in a crash." Hm, I guess I better make him enjoy his seatbelt.
Okay you're right, some of that data would come from highway crash reports, not a specific study. But why can't we have all this evidence in one place? And not have to swim through a hundred hefty journal articles to get to it?
I want a tool not unlike gapminder that takes summary statistics from tons of studies and allows me to mash them together.
Traditionally, this has been done with meta-analyses which are studies of studies that compare these outcomes and using different weights and probably other things (you want me to read wikipedia on this?) to come up with some sort of broad conclusion based on all the evidence.
Granted, this may not be as scientifically sound, but hey, everyone's doing homemade ecological studies (these are the gapminder kind) and I don't see anyone cryin'.
Here's the experience (for anyone I haven't lost)
I have a kid who hates to wear his seatbelt (hypothetical here of course). I think what are the actual risks if you don't wear a seatbelt. So I look up the Meta-analysis explorer under seat-belts for kids and I narrow it down to around his age range and I get some statistic like, A kid in the US who doesn't wear his seatbelt is 45 times more likely to die in a car crash. Hm, then I think, what's the chance of having a car crash? And then I look that up and it says "on average in your state 80 serious crashes occur every day. If you are on the road today, you have a 1 in 10,000 chance of being in a crash. If you drive everyday this year, you have a 1 in 400 chance of being involved in a crash." Hm, I guess I better make him enjoy his seatbelt.
Okay you're right, some of that data would come from highway crash reports, not a specific study. But why can't we have all this evidence in one place? And not have to swim through a hundred hefty journal articles to get to it?
Comments