big_willy_style_333 wrote: R. Jeeves wrote: MikeSpivey wrote:
R. Jeeves wrote:Hello Spivey,
In your most recent data dump, how was the % change YTD calculated? Were you given the percentages, or did you calculate them after being given the raw #s.
Because im running some numbers and it looks like there is either a weird discrepancy with the % change YTDs, or I am very bad at math; given that I am posting on this board, that latter is probably true, but I'd appreciate it if you humored me.
I was given all of it, including %'s.
Every year there are a few weird percentage results, and every years LSAC has an explanation for them fyi (different dates compared, different metric, etc). So *likely* they aren't as weird as they seem but let it on me let's see what it is!
something leads me to suspect that the % change YTDs have been reported as
(Current # - Previous #)/(Current #)
(Current # - Previous #)/(Previous #)
Yeah, I was noticing this yesterday as well. Specifically looking at LSAT scores of 179. It says they are up 38.5% to 13, but there is no way for this to be true using the correct (second) formula. But it does work with the first/incorrect formula if there were previously 8.
Yeah here's what confused me:
The table which Spivey posted has the current number of applicants, and the Percentage Change YTD for each and every LSAT score. So if Percentage Change = (Current # - Previous #)/(Previous #)
, you should be able to use the table to calculate (Previous #) for every LSAT score using the formula: (Previous #) = (Current #)/(1+Percentage Change)
But then spivey also posted the percentage change for each band - 175-180, 170-174, etc.
You can calculate the percentage change for each band by taking the sum of the Current # of applicants for each score in that band, and then comparing that to the sum of the Previous # of applicants for each score in the band.
If you use the formula Percentage Change = (Current # - Previous #)/(Previous #)
- which i believe is the correct formula - to calculate the percentage change for each score and each score band, then the numbers Spivey posted in the table are NOT consistent with the percentage changes that he posted for each band.
But if you use the formula Percentage Change = (Current # - Previous #)/(Current #)
, then the numbers Spivey posted in the table ARE consistent with the percentage changes that he posted for each band. So I am almost certain that this is the formula that was used. But as far as I know percentage change YTD is expressed as a percentage of the previous year's value, not the current year's value.
Maybe I'm wrong about something and will end up with egg on my face but w/e, if I had any shame, I wouldn't have posted this either way.