Why do low LizzyM scores have such low acceptance rates?

This forum made possible through the generous support of SDN members, donors, and sponsors. Thank you.

hj3818

Full Member
10+ Year Member
Joined
Dec 19, 2011
Messages
24
Reaction score
0
I was browsing through the MSAR database and noticed schools with lower average cGPAs and MCATs have terrifyingly low acceptance rates, while schools like Wash U in St. Louis or Duke have pretty high acceptance rates for such a high LizzyM scores.

Is this basically because lower LizzyM score schools attract more applicants, and as every other medical school in the country, their space is very limited, resulting in these low acceptance rates? I keep thinking even for schools who don't attract as many applicants, it should still be relatively proportional, where they would accept even less applicants for interviews...?

Just wondering.

Members don't see this ad.
 
I was browsing through the MSAR database and noticed schools with lower average cGPAs and MCATs have terrifyingly low acceptance rates, while schools like Wash U in St. Louis or Duke have pretty high acceptance rates for such a high LizzyM scores.

Is this basically because lower LizzyM score schools attract more applicants, and as every other medical school in the country, their space is very limited, resulting in these low acceptance rates? I keep thinking even for schools who don't attract as many applicants, it should still be relatively proportional, where they would accept even less applicants for interviews...?

Just wondering.
Acceptance rate is based on applicant number vs. seat number, with number of applicants that drop their acceptances thrown in as a confounding variable. It doesn't matter if the average student is a 3.8+gpa with a 36+ MCAT or a 3.4gpa with a 28 MCAT, number of applicants vs number of seats, with influences based on number of dropped acceptances, dictates the acceptance rate.

Beyond that, not sure what you are asking if the above isn't the answer to your question.
 
Members don't see this ad :)
Acceptance rate is based on applicant number vs. seat number, with number of applicants that drop their acceptances thrown in as a confounding variable. It doesn't matter if the average student is a 3.8+gpa with a 36+ MCAT or a 3.4gpa with a 28 MCAT, number of applicants vs number of seats, with influences based on number of dropped acceptances, dictates the acceptance rate.

Beyond that, not sure what you are asking if the above isn't the answer to your question.

Me either. It's simply based on merit.
 
Acceptance rate is based on applicant number vs. seat number, with number of applicants that drop their acceptances thrown in as a confounding variable. It doesn't matter if the average student is a 3.8+gpa with a 36+ MCAT or a 3.4gpa with a 28 MCAT, number of applicants vs number of seats, with influences based on number of dropped acceptances, dictates the acceptance rate.

Beyond that, not sure what you are asking if the above isn't the answer to your question.

Ah - I understand now. I wasn't looking at it from the right angle. Thanks!
 
In general a 3.4/28 is less desirable than a 3.8/35. It's that simple. Who would you want?
 
Based on absolutely no other information you would choose a 3.4/28 applicant over a 3.8/35 applicant? Why?

It was a joke...the answer was obvious enough that I didn't want to give the question credit. The OP wasn't asking who is the better applicant.
 
It's because literally everyone (like actually 1/3 of the total applicant pool, which is huge) applies to lower tier private schools like GW or Drexel. They are private, so no disadvantage no matter where your from. The awesome applicants apply to them as safeties, and a lot are rejected for yield protection. The mediocre and cruddy applicants apply as matches, and the usual low number get in, because that's just how med school works.
 
In general a 3.4/28 is less desirable than a 3.8/35. It's that simple. Who would you want?


The OP was asking why schools with a lower LizzyM average have a lower acceptance rate than schools with a higher LizzyM average. Not why applicants with a lower LizzyM score have lower acceptance rates than applicants with higher LizzyM scores.
 
Members don't see this ad :)
The OP was asking why schools with a lower LizzyM average have a lower acceptance rate than schools with a higher LizzyM average. Not why applicants with a lower LizzyM score have lower acceptance rates than applicants with higher LizzyM scores.

Yes - sorry, that is what I was asking. My bad on the misunderstanding. But... I do now understand why, thanks for everyone's responses, :D
 
I always thought it was just arbitrary because at one point it used to be +1. But as the applicants became more competitive, I saw it becoming a -1. But I could be wrong.

You are right. It was originally that you took your gpa(10)+MCAT and compared it to a school's average gpa(10) + avg MCAT - 1. The idea is that you have a shot even at schools where you are just a bit below the avg but in range. The new thinking is that you need to be above a school's average to have a good shot at getting an interview thus the formula for the schools should be changed to +1.

Multiplying gpa x 20 in essence doubling the worth of gpa in comparison to strong gpa would be a major change. Currently a 3.7/35 is a 72 and a 3.5/37 is also a 72. The new formula would change those to 109 and 105 where a 3.0/30 would be a 90 and a 4.0/40 would be 120. I'm going to need to look at some applicatons and see if that formula holds up for me.
 
You are right. It was originally that you took your gpa(10)+MCAT and compared it to a school's average gpa(10) + avg MCAT - 1. The idea is that you have a shot even at schools where you are just a bit below the avg but in range. The new thinking is that you need to be above a school's average to have a good shot at getting an interview thus the formula for the schools should be changed to +1.

Multiplying gpa x 20 in essence doubling the worth of gpa in comparison to strong gpa would be a major change. Currently a 3.7/35 is a 72 and a 3.5/37 is also a 72. The new formula would change those to 109 and 105 where a 3.0/30 would be a 90 and a 4.0/40 would be 120. I'm going to need to look at some applicatons and see if that formula holds up for me.

While certainly more complicated, I would imagine the most accurate way to generate a metric would be to scale both GPA and MCAT to a sigmoidal curve (if I had to guess, the center of the sigmoids would be 3.6 for GPA and 30 for MCAT) as this would better account for balance being preferred over a strong weakness in one being made up for in the other.
 
While certainly more complicated, I would imagine the most accurate way to generate a metric would be to scale both GPA and MCAT to a sigmoidal curve (if I had to guess, the center of the sigmoids would be 3.6 for GPA and 30 for MCAT) as this would better account for balance being preferred over a strong weakness in one being made up for in the other.

I'll tell you what... fit this curve (it's actually going to result in something more like a surface plot) to a formula with GPA and MCAT as arguments and I'll buy you a stethoscope. :D
 
While certainly more complicated, I would imagine the most accurate way to generate a metric would be to scale both GPA and MCAT to a sigmoidal curve (if I had to guess, the center of the sigmoids would be 3.6 for GPA and 30 for MCAT) as this would better account for balance being preferred over a strong weakness in one being made up for in the other.

The curve would be different for each school... the purpose is not to determine "what are my odds of getting into any school" but "to which schools should I apply?"

A 4.0/29 application is not going to get much traction at a school where the average MCAT is 35 although he would have a LizzyM 2.0 score that is the same (109) as someone with a 3.6/37.

If I were to reform the formula, I might consider gpa+ gpa (10)+MCAT = gpa(11)+MCAT with the understanding that you'd be giving extra weight to average school gpa when looking as school's LizzyM scores, too.
 
I'll tell you what... fit this curve (it's actually going to result in something more like a surface plot) to a formula with GPA and MCAT as arguments and I'll buy you a stethoscope. :D
How long do I have, and will you actually do so? Cuz I aced calc 3, so I might be able to do so given some time. :smuggrin:
 
How long do I have, and will you actually do so? Cuz I aced calc 3, so I might be able to do so given some time. :smuggrin:

Sadly, I am too poor to keep that promise, but it would still be cool if someone could. There would, of course, be limitations on a school-specific level due to fewer data points as compared to the AAMC summary table. You would only be provided a 10th, 50th, and 90th percentile GPA and MCAT for each school (and it's not safe to assume a normal distribution).
 
May be a silly question but, how does sGPA factor into this "score"? For non-traditionals such as myself, cGPA and sGPA may be vastly different, possibly even encompassing two entirely different time periods. Should a low(er) cGPA still carry so much weight?
 
May be a silly question but, how does sGPA factor into this "score"? For non-traditionals such as myself, cGPA and sGPA may be vastly different, possibly even encompassing two entirely different time periods. Should a low(er) cGPA still carry so much weight?

sgpa doesn't factor in except that it is part of your total gpa.

I came up with this after seeing applicants not get in anywhere despite applying early to 20 schools. Come to find out they were applying to the top 20 with a MCAT of 30 and a gpa of 3.3. Good enough, perhaps to get into some schools (particularly 5 years ago when I started this) but not good enough for the top 20.

If you want to be ballsy, calculate your total gpa giving double weight to every course taken as a post-bac (e.g. a 4 credit course for which you earned an A would be 4.0 x 8 credits with 8 credits in the denominator, too). That may overestimate your chances of getting an interview but it might take into account the boost you may get for having an excellent post-bac after a lackluster undergrad record.
 
sgpa doesn't factor in except that it is part of your total gpa.

I came up with this after seeing applicants not get in anywhere despite applying early to 20 schools. Come to find out they were applying to the top 20 with a MCAT of 30 and a gpa of 3.3. Good enough, perhaps to get into some schools (particularly 5 years ago when I started this) but not good enough for the top 20.

If you want to be ballsy, calculate your total gpa giving double weight to every course taken as a post-bac (e.g. a 4 credit course for which you earned an A would be 4.0 x 8 credits with 8 credits in the denominator, too). That may overestimate your chances of getting an interview but it might take into account the boost you may get for having an excellent post-bac after a lackluster undergrad record.

Ah - interesting.. :idea: Thanks again for your input, LizzyM. There may be a couple of different calculations when it comes to formulating a list of schools when the time comes!
 
Sadly, I am too poor to keep that promise, but it would still be cool if someone could. There would, of course, be limitations on a school-specific level due to fewer data points as compared to the AAMC summary table. You would only be provided a 10th, 50th, and 90th percentile GPA and MCAT for each school (and it's not safe to assume a normal distribution).

It's alright, I understand. If only this had come up before xmas when I'd have lots of free time to work on it. Well, if I'm not doing anything this weekend I'll toy with it.

The curve would be different for each school... the purpose is not to determine "what are my odds of getting into any school" but "to which schools should I apply?"

A 4.0/29 application is not going to get much traction at a school where the average MCAT is 35 although he would have a LizzyM 2.0 score that is the same (109) as someone with a 3.6/37.
Why would the curve need to be different? Isn't the example you gave congruent with the point of a sigmoidal scaling? The LizzyM score appears to break down when the MCAT and GPA are far off. People seem to suggest that the answer is to weight the GPA more, but I believe this only appears to be the case because aplicants with high MCATs but held back by low GPAs seem to be much more common than the other way around. Instead, wouldn't it stand to reason that the most significant fluctiations are those centered around the "standard acceptible" scores?

I will admit after doing some test calculations, it might be better to scale GPA to a logrithmic curve instead (since these scores are arbitrary, they can easily be adjusted to balance properly)
 
It's alright, I understand. If only this had come up before xmas when I'd have lots of free time to work on it. Well, if I'm not doing anything this weekend I'll toy with it.


Why would the curve need to be different? Isn't the example you gave congruent with the point of a sigmoidal scaling? The LizzyM score appears to break down when the MCAT and GPA are far off. People seem to suggest that the answer is to weight the GPA more, but I believe this only appears to be the case because aplicants with high MCATs but held back by low GPAs seem to be much more common than the other way around. Instead, wouldn't it stand to reason that the most significant fluctiations are those centered around the "standard acceptible" scores?

I will admit after doing some test calculations, it might be better to scale GPA to a logrithmic curve instead (since these scores are arbitrary, they can easily be adjusted to balance properly)

Yes, it has been pointed out that the LizzyM score doesn't seem to work well when one's gpa is low (<3.2) and the MCAT is high >35) or when the gpa is high (>3.7) and the MCAT is low (<25). Well, those oddball applicants who make up a small fraction of the total pool are not going to fit into an easy formula because they either get filtered out very quickly or they require a great deal of weighing and judging on the part of the adcom ("the MCAT is so low" "but the gpa is excellent" and reasons offered back & forth for why the appliant should or shouldn't be interviewed given the major, rigor of the academic courses taken, school attended, other life experiences, qualilty of the LORs and essays, extenuating circumstances and on & on. I'm not sure that you an make a formula that will accurately predict which schools will take a chance at someone with an unusual combination of MCAT & gpa.
 
Yes, it has been pointed out that the LizzyM score doesn't seem to work well when one's gpa is low (<3.2) and the MCAT is high >35) or when the gpa is high (>3.7) and the MCAT is low (<25). Well, those oddball applicants who make up a small fraction of the total pool are not going to fit into an easy formula because they either get filtered out very quickly or they require a great deal of weighing and judging on the part of the adcom ("the MCAT is so low" "but the gpa is excellent" and reasons offered back & forth for why the appliant should or shouldn't be interviewed given the major, rigor of the academic courses taken, school attended, other life experiences, qualilty of the LORs and essays, extenuating circumstances and on & on. I'm not sure that you an make a formula that will accurately predict which schools will take a chance at someone with an unusual combination of MCAT & gpa.

Yeah, I was about to say, if multiplying the GPA by 20 instead of 10 gives a more accurate score to gauge competitiveness, that really bodes well for me. LizzyM score = (3.98 x 20) + 25 = 104.6, which puts me within 1 point of a school I like: (3.67 x 20) + 32 = 105.4. As you mentioned, this goes to show how important the other factors are. If you're an "oddball applicant" as far as numbers are concerned (as I am), having stellar work experiences, ECs, LoRs, etc, may give admissions committees that extra nudge in your favor.
 
Yeah, I was about to say, if multiplying the GPA by 20 instead of 10 gives a more accurate score to gauge competitiveness, that really bodes well for me. LizzyM score = (3.98 x 20) + 25 = 104.6, which puts me within 1 point of a school I like: (3.67 x 20) + 32 = 105.4. As you mentioned, this goes to show how important the other factors are. If you're an "oddball applicant" as far as numbers are concerned (as I am), having stellar work experiences, ECs, LoRs, etc, may give admissions committees that extra nudge in your favor.

why not retake the MCAT and show that the first score was not a good measure of your ability?
 
why not retake the MCAT and show that the first score was not a good measure of your ability?

Doing just that! In the meantime, I applied to a couple of my favorite schools for this cycle, too.
 
A 4.0/29 application is not going to get much traction at a school where the average MCAT is 35 although he would have a LizzyM 2.0 score that is the same (109) as someone with a 3.6/37.

If I were to reform the formula, I might consider gpa+ gpa (10)+MCAT = gpa(11)+MCAT with the understanding that you'd be giving extra weight to average school gpa when looking as school's LizzyM scores, too.

I acknowledge that my proposal was based on the systematic disconnect I noticed in the *overall* admission rates for applicants with identical LizzyM scores. It may be that *individual* schools may place greater weight on the MCAT than the collective admissions process would suggest, and I of course defer to LizzyM's judgment and experience in that regard. I would be interested to see, however, whether the increased GPA weighting might track better with applicants' success at individual schools, particularly if we consider more common GPA/MCAT combinations (4.0/29 seems rare to my mind).

Nevertheless, my original argument centered on the following two applicants: a 3.3/36 applicant and a 3.8/31 applicant (LizzyM scores of 69). Both should have equal chances of overall admission, yet the 3.8/31 applicant has a ~77% chance while the 3.3/36 applicant has only a ~59% chance of admission (based on the 2010 AAMC data, graphed here by sector9: http://i55.tinypic.com/v4o2om.jpg).

I then looked at a hypothetical applicant with a 31 MCAT and estimated how many MCAT points would be required to make up a 0.2 pt GPA deficit going down the scale:

3.9/31 = 82% = 3.7/35 (+4)
3.8/31 = 77% = 3.6/36 (+5)
3.7/31 = 72% = 3.5/36 (+5)
3.6/31 = 64% = 3.4/35.5 (+4.5)
3.5/31 = 56% = 3.3/35 (+4)
3.4/31 = 48% = 3.2/34.5 (+3.5)

From that trend, it looks like 4 MCAT points roughly equal 0.2 GPA points, at least in *this quadrant* of the graph. To track that data for overall admissions, the GPA weighting should be roughly doubled. That is:

LizzyM 2.0 = GPA*(20) + MCAT

This 0.1:2 relationship also seems to hold up well for a 0.1 pt GPA deficit (e.g., 3.8/31 = 3.7/33 and 3.6/31 = 3.5/33). For a 0.3 GPA deficit, though, you'd need more than 6 extra pts on your MCAT (e.g., 3.8/31 = 3.5/39) to have an equal overall chance at admission.

In any case, I'd be interested to hear your thoughts on this.
 
(4.0/29 seems rare to my mind).

Yeah, I was about to say, if multiplying the GPA by 20 instead of 10 gives a more accurate score to gauge competitiveness, that really bodes well for me. LizzyM score = (3.98 x 20) + 25 = 104.6, which puts me within 1 point of a school I like: (3.67 x 20) + 32 = 105.4. As you mentioned, this goes to show how important the other factors are. If you're an "oddball applicant" as far as numbers are concerned (as I am), having stellar work experiences, ECs, LoRs, etc, may give admissions committees that extra nudge in your favor.

And then there's this poster with a 4.0/25 :). At any rate, I did the calculations in the 3.3+ and 30+ "quadrant" of the graph, so they'd probably be most applicable there.
 
Hm, could one of you tell me where you found those probabilities from AAMC? That curve looks interesting to me.
 
Doing just that! In the meantime, I applied to a couple of my favorite schools for this cycle, too.

If you don't get in this cycle you will be branded a re-applicant which LizzyM has pointed out makes her assume you are damaged goods.
 
If you don't get in this cycle you will be branded a re-applicant which LizzyM has pointed out makes her assume you are damaged goods.

True, which I understand is a bad thing, unless you've significantly improved your application. By June 2012, I will have shadowing hours, a letter from a physician, and mentoring experience, all of which I was lacking this time around. In addition, I will be volunteering as a basketball coach and retaking the MCAT. So, reapplying shouldn't hurt, right?
 
True, which I understand is a bad thing, unless you've significantly improved your application. By June 2012, I will have shadowing hours, a letter from a physician, and mentoring experience, all of which I was lacking this time around. In addition, I will be volunteering as a basketball coach and retaking the MCAT. So, reapplying shouldn't hurt, right?

That will be a stronger application.... without those things you had no business applying in this cycle.
 
Wait, I figured it out!!!

LizzyM, that point about making individual curves for each school wasn't a shortcoming, but a great idea for implementing the metrics I mentioned. We would only need to make a template equation, likely with MCAT scaled to a sigmoidal curve and GPA scaled to a logrithmic curve, and the GPA worth somewhere between 1 and 2 times what the MCAT is, and centered around the school's average, and your score is based on where it stands comparedto each school's average. Thus we would be using not an absolute scoring scale like the current LizzyM score does, but a relative scoring system for each school. Since the intention is to give you a good idea on where to apply, this may be more effective as it could better account for some schools placing more emphasis on higher GPAs or higher MCATs.

Edit: And that's actually something that could easily be made into an excel spreadsheet (given the time to enter it all in)
 
Last edited:
That will be a stronger application.... without those things you had no business applying in this cycle.

Point taken, although DO schools have taken somewhat of an interest in me. I have no experience specifically with mentoring, but I have worked as a CNA, served as a medical director, tutored math/science, taught piano, etc.

But I didn't mean to turn the discussion toward me. :laugh: Among Navier-Stokes "existence and smoothness," I think this problem belongs with the Millennium Prize Problems in mathematics:

http://en.wikipedia.org/wiki/Unsolved_problems_in_mathematics
 
If you want to be ballsy, calculate your total gpa giving double weight to every course taken as a post-bac (e.g. a 4 credit course for which you earned an A would be 4.0 x 8 credits with 8 credits in the denominator, too). That may overestimate your chances of getting an interview but it might take into account the boost you may get for having an excellent post-bac after a lackluster undergrad record.

Thank you. I was wondering how much a postbac program boosts a LizzyM score, besides the actual increase in cGPA. That raises my score considerably.
 
Top