Criteria for interview invite vs. acceptances

[235009]

---

Members don't see this ad.

https://www.aamc.org/download/261106/data/aibvol11_no6.pdf

sorry if it has already been posted. the figure on page 2 says it all.

 

[flatearth22]

Thanks for this....however I'm a little skeptical that there are no green qualifiers (demographic data) on either list. I find it hard to believe that that state residency, racial background, rural experience, etc. isn't important.

 

[NickNaylor]

Wow, great article. Thanks for posting.

 

[lkthlttr]

Thanks for this....however I'm a little skeptical that there are no green qualifiers (demographic data) on either list. I find it hard to believe that that state residency, racial background, rural experience, etc. isn't important.

I don't know why this would be the case, because it obviously does make a difference.

 

[kautionwirez]

Thanks for this....however I'm a little skeptical that there are no green qualifiers (demographic data) on either list. I find it hard to believe that that state residency, racial background, rural experience, etc. isn't important.

I agree flatearth22. Those factors are definitely add points to your application.

In my opinion, once you get an interview, the major factor it is how well you sell yourself to the interviewer so he can advocate for you when presenting your file. Getting an invite is pretty much you are well qualified for the school (GPA/MCAT/ECs/LORS) but they are wondering if you are a good fit/superstar via interview.

 

[flatearth22]

I agree flatearth22. Those factors are definitely add points to your application.

In my opinion, once you get an interview, the major factor it is how well you sell yourself to the interviewer so he can advocate for you when presenting your file. Getting an invite is pretty much you are well qualified for the school (GPA/MCAT/ECs/LORS) but they are wondering if you are a good fit/superstar via interview.

I mostly agree but you also have to consider the accepted:interviewed ratio. At some schools (like Tulane which accepts ~80% of those interivewed) that ratio is very high which kind of implies that they already really like you and are looking for reasons NOT to accept you based on the interview (basically if you don't come across as a complete jerk you should get in)....whereas the stats for a place like GWU (where about 1/3 of those interviewed are accepted) implies that they are interested in you but that you really have to shine at the interview to earn your spot there.

 

[centrigeugle]

What a pointless paper. Acually, I'll go a step further. What a ******ed paper.

For example:

Interview Offer Importance criteria

1. GPA: Cumulative science and math (3.7)
2. GPA: Cumulative (3.6)
3. MCAT Total scores (3.5)
4. Letters of recommendation (3.4)
5. Community service: medical (3.3)
6. Personal statements (3.2)
7. Medical/clinical work experience (3.2)
8. Community service: non-medical (3.1)
9. Leadership experience (3.0)
10. Completion of premedical requirements (3.0)
11. Experience with underserved populations (2.7)

Acceptance Offer Importance criteria

1. Interview recommendation (4.5)
2. Letters of recommendation (3.8)
3. GPA: Cumulative science and math (3.7)
4. Community service: medical (3.6)
5. GPA: Cumulative (3.6)
6. MCAT Total scores (3.4)
7. Personal statements (3.4)
8. Medical/clinical work experience (3.4)
9. Community service: non-medical (3.3)
10. Leadership experience (3.2)
11. Completion of premedical requirements (3.1)
12. Experience with underserved populations (3.0)

So... ...so what? Seriously? What does this table tell us? It tells us that your GPA and MCAT are in the top criteria for both selection processes. In fact, if you remove "Interview Recommendation" (which you should, since there's no real basis for comparison and it's blatantly obvious the interview is important) your "numbers" account for three of the top five criteria. In both cases.

Oh, also:

The admissions data presented standard deviations ranging from 0.9 to 1.7, indicating variation in
importance across medical schools.

So actually what this table tells us is that there's absolutely no difference between the importance of your grades at various points. None of the differences are even close to statistically significant. This article asserts that uGPA, sGPA and MCAT become less important after you get an interview invite. In fact, they're exactly as important as before.

The biggest change of any criteria is for letters of recommendation, which increases by an entire 0.4 (out of 5.0).

Once you remove the interview recommendation, those two charts are, functionally, exactly the same. If there are any surprising facts to take out of this study, I would actually say it's how unimportant leadership and volunteering for the underserved turn out to be.

Without the raw data or more school-specific studies, this is worthless for everybody.

 

[NickNaylor]

The admissions process is incredibly subjective - the utility of any objective data is going to be dubious at best. That said, it doesn't mean that the data is as a whole useless.

 

[centrigeugle]

The admissions process is incredibly subjective - the utility of any objective data is going to be dubious at best. That said, it doesn't mean that the data is as a whole useless.

I agree - and should have probably clarified. As you know I'm sometime prone to over-snarking my posts.

I think that the dataset, taken as a whole, may be useful. As I said, I think it's interesting to note that the "stats" tend to stay on top, non-medical EC's of any kind stay on the bottom, and recommendations become important at the interview stage.

Considering how important leadership and unserved community service is, according to SDN gospel, it's an interesting (and somewhat cynical) confirmation that this process is, indeed selective. But it's also highly numbers-driven.

But as a method of comparison, for "what's-important-after-you-get-the-interview" which is what this paper seems to be implying, it's useless. It's not as if you can change any aspects of your application once you've applied and been invited.

I guess I'm just irritated since the two tables are, as I said, exactly the same from a statistical standpoint. When your standard deviation is 50% of your entire measurement scale, there's something wrong with the experimental design.

 

[kpcrew]

thanks, that was an interesting article. i've always wondered what criteria they favored when choosing between applicants.

 

[alexandertg6]

I agree - and should have probably clarified. As you know I'm sometime prone to over-snarking my posts.

I think that the dataset, taken as a whole, may be useful. As I said, I think it's interesting to note that the "stats" tend to stay on top, non-medical EC's of any kind stay on the bottom, and recommendations become important at the interview stage.

Considering how important leadership and unserved community service is, according to SDN gospel, it's an interesting (and somewhat cynical) confirmation that this process is, indeed selective. But it's also highly numbers-driven.

But as a method of comparison, for "what's-important-after-you-get-the-interview" which is what this paper seems to be implying, it's useless. It's not as if you can change any aspects of your application once you've applied and been invited.

I guess I'm just irritated since the two tables are, as I said, exactly the same from a statistical standpoint. When your standard deviation is 50% of your entire measurement scale, there's something wrong with the experimental design.

I agree, the article should be taken at face value for what it is, a wild estimation for something incredibly subjective. That being said, a wild estimation is way better than having no idea. Interesting article, and cool find OP.

 

[KLycos]

This is how a lot of adcoms operate, by creating some sort of score after your interview, and getting a certain score based on these factors will either get you acceptance, waitlist, rejection, etc.

 

[phltz]

I guess I'm just irritated since the two tables are, as I said, exactly the same from a statistical standpoint. When your standard deviation is 50% of your entire measurement scale, there's something wrong with the experimental design.

Standard deviations do not work that way. The standard deviations vary from 0.9 to 1.7, but because of the sample size, the uncertainty in the values of the means varies from 0.08 to 0.16.