Statistical Analysis of the Most Important Factors to gain an Acceptance

[pathdr2b]

---

Members do not see this ad.

Be careful w/ the interpretation. Remember, to reference the regression for MCAT date, something had to be left out. That one thing was those who took the MCAT both months. Therefore, those that took the August MCAT had never taken the MCAT before. They got their results back late in the game. I can change the variables in the OLS so that August or April is left out instead of both. Just ask.

Please do so for April and August test takers. I'm sure I'm not the only one curious about this result. 😉

 

[Tippy007]

You are a god of econometrics. Thanks for the data.

 

[trauma_junky]

The model is an OLS, which requires PARTIAL differentiation. True, you probably won't get looked at. But this isn't the aim or scope of the model. The idea is to show how increasing (decreasing) one variable will increase (decrease) your chance of getting accepted. So your question isn't an issue. Race was already described. About the lies, well, we know how to lie with stats. This isn't the case (I didn't throw data away). If you want the data, I will send it to you so that you can reproduce my results (true scientist).

Sorry, it wasn't a slam, just the first thing my stats prof told me in undergrad. Thought it was funny and wanted to share it. Is it simple summation if I want to consider multiple factors here?

 

[Medikit]

I just ran a regression model with a full mode analysis and I found that there is a 101% chance that I don't understand stats very well.

 

[trauma_junky]

What about the personal statement? How much weight does that hold?

If you think of med school admissions as 3 parts then this part is only 25% of the total factor.

So the 3 factors: 1) MCAT & GPA 2) LOR's, PS, EC's 3) Interview. An adcom once told me that that parts one and 2 make up 50% of the admissions decision and the interview the other 50%. now as far as getting an interview I would say the numbers are the biggest factor. Anyone can get good LOR's, fork over $250 to essayedge for a PS and say they volunteered forever, but your GPA and MCAT are the an absoulute measure (even though not all undergrad institutions are created the same, that is what the MCAT was supposed to do, prior to TPR and KAplan classes. Now it is not so much the case.) Bottom line, all factors involved, I believe that the most imporatant factors are simply the GPA and MCAT to get an interview, for their it is all personality and the ability to sell yourself to the interviewer.

 

[gary5]

Cool statistics, although your interpretation is over-simplied.

I added the MCAT values and got a total of .2895, and a total GPA value of .2755. That's what I expected: that each contributes about roughly 1/3 to getting accepted. Of course, the model does not include LORs, personal essay, and interviewing. It makes sense to me that the WS value was low - I've heard that schools rely more on the numeric MCAT value (e.g. 30).

WHY VERBAL: Here's why a high verbal score correlates well with getting accepted. If your verbal skills are excellent, then you communicate well, which means you are more likely to get better LORs (professors know you by what you communicate to them). Also, excellent verbal skills are critical for writing a good personal essay, writing good secondary essays, and interviewing well. So, your high verbal value makes perfect sense to me. Note, however, that verbal skills are developed over a lifetime. They are the least influenced by last-minute cramming.

WHY SCIENCE GPA: The total contribution of MCAT and GPA are similar (.2895 and .2755). However, the MCAT value is split among 4 factors (bio, physical, verbal, writing), while GPA is split among 2 factors (science and non-science). So, on average, the GPA values are going to be twice the MCAT values in percent contribution. That explains part of it. However, it is interesting that the science part of GPA outweighs the non-science, while the science part of MCAT does not outweight the non-science MCAT portion. Interesting...

So, perhaps the overall lesson is that medical students are those that did very well in science (high science GPA) but who also have excellent verbal skills (verbal MCAT). Ask an admissions counselor and I'm sure they'll agree that this is fundamentally what they want. The interview is meant to ensure this.

The # of schools applied to is interesting. It's cool to see a stat on that. I've also noticed that people that apply to too many tend to not get in (50+). Something about lack of focus. 🙂

The April MCAT value is interesting. So, relatively, an August MCATer has a 6.72% disadvantage compared to an April MCATer. That makes sense to me. I'm surprised, however, that the August MCAT value is 0%. I took the August MCAT and got in, and I wouldn't have gotten in without it (obviously), so it must have some weight. I'd expect it to be significantly less than the April value, like 1/3, or roughly 2%. Any thoughts on this?

Anyway, very cool stats! Thanks for sharing them.

 

[elephunt]

hey viking,

as a person with an econ undergrad heading into med school now, i just wanted to say i enjoyed and appreciated your analysis. thanks for taking the time to share it with SDN, and thanks for being patient and helpful with the questions/comments posted. econometrics definitely wasn't my favorite class, so i am all the more impressed by your stats "prowress."
you go on with your bad self.

 

[1viking]

Please do so for April and August test takers. I'm sure I'm not the only one curious about this result. 😉

here it is. The variables, well you can figure it out (they haven't changed, except for MCAT Both). First numbers are coefficients. Second numbers are p-values.
Indexed against the spring:
V 0.36007E-01 0.000
P 0.20628E-01 0.010
B 0.13005E-01 0.137
W 0.65489E-02 0.254
BCPM 0.26175 0.013
Allother 0.10872 0.378
TOT 0.50032E-01 0.816
NUMAPP 0.94769E-02 0.000
MCATF -0.78816E-01 0.004
MCATB 0.78839E-01 0.031


Indexed against the Fall:
V 0.36007E-01 0.000
P 0.20628E-01 0.010
B 0.13005E-01 0.137
W 0.65489E-02 0.254
BCPM 0.26175 0.013
AO 0.10872 0.378
TOT 0.50032E-01 0.799
NUMAPP 0.94769E-02 0.000
MCATS 0.78816E-01 0.005
MCATB 0.23148E-04 0.999

To interpret P-values, think like this: if greater than 5%, it is safe to assume that the variable coefficient is actually zero, with a 95% confidence interval.

Be human. If you did poorly on the MCAT because you had Mono (like myself), then take it in August so that you can get a better score, because you will (common sense). If you did your best in April, then I would stay away from the August MCAT.

 

[ornis4]

Now, in a completely un-mathematical context, hear me out...for those of you familiar with blackjack...you can get what's known as a "hard" or "soft" version of a number when dealt your hand, referring to whether or not you have an Ace(which can count as either 1 or 11) and also to your chances of winning. Pertaining to this study and MCAT scores, let's use me as an example...I got a 30P, with 9's in both science sections and a 12 in Verbal. My GPA is a little below med school matriculant average(3.4), and my EC's are good and I'm about to complete my MS degree. Despite the lack of double digits in each section, should I consider my score a "hard" 30, in that I seem to have done well on the most important section? Do you think it's just as good as 10's across the board? If this study is correct, then did I achieve the best possible score of 30? Or would this model prefer something like 15 verbal, 10 PS and 5 BS? I find the analysis interesting, but I also agree that there are waaaaay too many other factors that come into play...especially when you consider the percentages of interviewees who don't get accepted...perhaps a pure numerical study would be more accurate for those who get interviewed only? At least that would take away one of the unaccounted-for factors?

 

[1viking]

Now, in a completely un-mathematical context, hear me out...for those of you familiar with blackjack...you can get what's known as a "hard" or "soft" version of a number when dealt your hand, referring to whether or not you have an Ace(which can count as either 1 or 11) and also to your chances of winning. Pertaining to this study and MCAT scores, let's use me as an example...I got a 30P, with 9's in both science sections and a 12 in Verbal. My GPA is a little below med school matriculant average(3.4), and my EC's are good and I'm about to complete my MS degree. Despite the lack of double digits in each section, should I consider my score a "hard" 30, in that I seem to have done well on the most important section? Do you think it's just as good as 10's across the board? If this study is correct, then did I achieve the best possible score of 30? Or would this model prefer something like 15 verbal, 10 PS and 5 BS? I find the analysis interesting, but I also agree that there are waaaaay too many other factors that come into play...especially when you consider the percentages of interviewees who don't get accepted...perhaps a pure numerical study would be more accurate for those who get interviewed only? At least that would take away one of the unaccounted-for factors?

the 15, 10, 5 won't happen--it is very rare. But I think you have the "hard" 30. I like your idea of using only those that were interviewd. if you can get data, I'll regress it.

 

[1viking]

hey viking,

as a person with an econ undergrad heading into med school now, i just wanted to say i enjoyed and appreciated your analysis. thanks for taking the time to share it with SDN, and thanks for being patient and helpful with the questions/comments posted. econometrics definitely wasn't my favorite class, so i am all the more impressed by your stats "prowress."
you go on with your bad self.

--------------------------------------------------------------------------------

Econ didn't help me too much with apps.--too hard with too little respect. People think that Econ is a fluff major. Our program is a good PhD prep. I was offended when AMCAS wouldn't count those classes as science. We had as much math (if not more) as my physics core.

Econometrics was my favorite class! 🙂

 

[1viking]

Cool statistics, although your interpretation is over-simplied.

The # of schools applied to is interesting. It's cool to see a stat on that. I've also noticed that people that apply to too many tend to not get in (50+). Something about lack of focus. 🙂

Anyway, very cool stats! Thanks for sharing them.

The max on number applied is 59 schools. There were people that applied to many schools and got in. They can't look too low on many apps.

oversimplified because some need oversimplification.

Glad you liked it. I hope it puts some rumors to rest.

 

[hattrics]

You know, "there are three kinds of lies. Lies, damn lies, and statistics." - Benjamin Disraeli.

Now before this post ruffles any feathers, I must say, I'm only kidding...