How do I solve this simple math problem?
January 16, 2007 11:26 AM   Subscribe

This is probably a simple math question for someone who knows math. Schools create an index number by using this equation: (X * testscorenumber) + (Y * GPA) + constant = indexnumber Some schools index numbers max out at 5, some at 200. Nonetheless, the difference between schools is in how they weight GPA vs. testscore. What formula would allow me to compare the importance of once factor over another among schools. I.e., School A weighs testscores 2x more than school B? Is it a simple ration between X and Y?
posted by JakeWalker to Education (12 answers total)
 
The question is a little vague, but yeah, the most meaningful value for comparing the different school's weightings would be the ratio of X to Y (or vice versa).
posted by chrismear at 11:30 AM on January 16, 2007


A ratio of X to Y would be useful for RELATIVE comparisons of schools' weighting, but if you want to get a real world number like the 2x mentioned in your example, it would be helpful to know the range of possible testscore numbers. Is this a percentage of 0-100, or a "letter grade" 0-4 like the GPA?
posted by tkolstee at 11:54 AM on January 16, 2007


Response by poster: GPA would be 0-4, TESTSCORE would be 120-180.
posted by JakeWalker at 11:57 AM on January 16, 2007


I would just use your actual numbers and then the average numbers for students admitted to that school. This will show where you place.
posted by caddis at 11:58 AM on January 16, 2007


Response by poster: That's not what I'm trying to do. I'm trying to compare for all applicants what schools emphasize one factor over another, and to what extent.
posted by JakeWalker at 11:59 AM on January 16, 2007


Yes, if you just want a comparison between school A and school B, you can compare XA/YA vs. XB/YB.

To get an answer to how much school A values testscore compared to how much school A values GPA, you'd need to incorporate more data about testscore and GPA--possibly what their ranges are, and even better if you can get them, average values and standard deviations for them. But that doesn't appear to be what you're asking.

Personally, I'm curious about how you got information about such formulae at all. When I was applying to colleges nearly 20 years ago, they gave out little to no information about their internal processes for making admission decisions. Has that changed? (On re-reading, you didn't say this was about college admissions at all, that was just my assumption.)
posted by DevilsAdvocate at 12:53 PM on January 16, 2007


The range on GPA is 4, the range on TESTSCORE (should I just call it LSAT?) is 60. So the same number for X and Y would give one point of LSAT as much weight as one point of GPA -- or, 15 times the weight. So one way to do it would be to calculate (Y*15/X), and if that number is greater than 1, the school weighs GPA more heavily.

A more accurate way to look at it would be to look at the real-world ranges for GPA and LSAT. If the real pool of applicants that matter are those with GPAs between 3 and 4, and those with LSATs between 160 and 180, then the formula would be (Y*20/X) since the range for LSAT is 20 times as big as for GPA.
posted by raf at 12:54 PM on January 16, 2007


Alternatively, if you're just trying to compare schools, then you really want to look at the baseline for all schools. So if the "average" school cares about GPA twice as much as LSAT, then the relative numbers could be compared to that average. Then what you would want to do is calculate Y/X for all schools and compare that number for any particular school to the average Y/X. So our relative rank r_school = (Y_school/X_school)/(Y_average/X_average). That r_school is >1 for schools that care about GPA more than the average school does, and <1 for those that care about LSAT more than the average school does.
posted by raf at 12:57 PM on January 16, 2007


Those are good points raf. Although when comparing two schools, the denominator in r_school (the numbers for an average school) fall out and you're left with just a comparison of the two Y/X's.
posted by vacapinta at 1:09 PM on January 16, 2007


To see this comparison in action with 2000 law school data (and to answer DevilsAdvocate's question about availability of said data), check out this analysis, which does exactly what you want to do. It shows quite distinctly how much more weight is generally put on the LSAT than elsewhere (note certain outliers, e.g. UC-Berkeley)

Of course, remember that the index is not the only way decisions are made, etc.
posted by Partial Law at 1:16 PM on January 16, 2007


Response by poster: Awesome, Partial Law -- Thanks.
posted by JakeWalker at 4:15 PM on January 16, 2007


Response by poster: FWIW, it's for more work that I am doing with this:

http://www.LSATreport.com/table.php

And to the person who asked, the Law School Admissions Council publishes the index formula for almost every law school in the country. About 25 refrain from publishing their index score (or claim to not use one), including many of the top ranked schools.
posted by JakeWalker at 4:16 PM on January 16, 2007


« Older Are there sports blogs for those uninterested in...   |   local business information Newer »
This thread is closed to new comments.