Study on McMaster admissions

[MedKeener101]

Found this on twitter today 

 

A lot of stats in table 1 for McMaster averages

 

https://bmcmededuc.biomedcentral.com/articles/10.1186/s12909-020-02126-0

 

 

[MedicineLCS]

Very interesting, thanks for sharing! 

Take a look at table 1, they report the avg. GPA (We knew that), CARS (ditto), and CASPer score (out of 9 presumably)+: https://bmcmededuc.biomedcentral.com/articles/10.1186/s12909-020-02126-0/tables/1

 

Measure	All	Accepted	Rejected	<P-value
CARS	126.2, ±2.3	129.2, ±1.5	126.0, ±2.3	< 0.001
Grade Point Average (Mean, SD)	3.6, ±0.4	3.9, ±0.1	3.6, ±0.4	< 0.001
CASPer (Mean, SD)	4.6, ±1.2	5.7, ±0.9	4.5, ±1.2	< 0.001
MMI (Mean, SD)a	6.4, ±1.1	7.0, ±0.7	5.4, ±0.8	0.10

The size of the SD on the "all/rejected" GPAs is interesting vs. the much narrower accepted band, but I supposee that's the way stats work. It does give some hope to anyone applying that the overwhelming majority of Mac applicants are not 3.9/130 level applicants. 

Also, interestingly, they found "No significant collinearity was found between these variables." where those variables were GPA, CARS, interview, CASPer, and income. https://bmcmededuc.biomedcentral.com/articles/10.1186/s12909-020-02126-0/tables/5The CASPer confidence intervals actually overlap, while still being statistically different. However, CASPer's odds ratio is ~9, vs. ~3 for CARS/GPA on their own. 

I wonder what this kind of study would look like at a school that actually looks at ECs. There is no admissions measure that high SES doesn't give an advantage in (Short of a Manitoba points based system, which would be very politically contentious) so I wonder if Mac's emphasis on stats may disproportionately favour higher SES applicants.