dudester wrote: Space_Cowboy wrote:
dudester wrote:Fifth post from the top.
Ah missed that. As I conceded, the 95+% reporting requirement was silly. Lets just go with those who don't report salary as earning $0. I'd rather bias the data toward lower salaries and give schools a very strong incentive to get complete information. The school can be eligible for GradPLUS loans but prospectives get to see how incomplete the data is. That good?
It would be irresponsible for schools to bias the information either way rather than just reporting the information they actually have, especially since people should be able to easily modify the data in the way you suggest if they wanted to.
Class size: 300
Response rate: 60%
Average salary: $80,000
Didn't respond: 120 students
New average salary: $48,000
Thanks. I didn't know how math works.
Dude, they already bias the information by doing everything thing they can to follow up with only those graduates who were reportedly employed at graduation, while not bothering to check up on those that didn't have jobs at graduation. The response rates so salary data suck because schools don't want to collect data on the lower end.
Back to your introduction to mathematics example. Averages are terrible measures to gauge outcomes. Outliers (very high or very low salaries) distort the impression you get from a data set. Publish 25th percentile, median and 75th percentile salary data. That way, making non-reporters = $0 doesn't affect data too terribly (shifting the median instead of dragging down an average), unless we're talking about abysmal response rates, which, I'm sorry is indicative of a problem.
Is this an ideological block for you? You know, the second the evil government requires disclosure freedom is destroyed for all mankind, or something?