A proposal/idea to consider

[Einsteinemc2]

Advertisement - Members don't see this ad

The following is a proposal that would hopefully help end the barrage of “What are my chances” threads. The problem, of course, is that is ultimately impossible to objectively gauge one’s chance to get into medical school… perhaps this can serve as a guide. It is only a framework and not a completed project. I view this as a sort of open-source project for anybody who would want to contribute. I’ll present the idea, and then point out the flaws.

MCAT (40%)
0-22=0
23-24=10
25-26=20
27-29=30
30-34=40
35+=50
(If 40+, bonus of 60!)
GPA (30%)
0-2.5=0
2.5-3.0=10
3.0-3.3=20
3.3-3.6=30
3.6-3.8=40
3.8+=50

LOR (15%)
10=Writers had to ask how they knew you when asked to write the letter
20=Writers knew you from class only, no office hour visits etc (or any other weak relationship)
30=Writers knew you as a person, and you’re pretty sure they wrote positively
40=Writers knew you as a person, you know that they wrote extremely positive letters with hardly any negative comments
50=Same as above, except writers are famous/well known

EC (15%)
0=You did nothing
10= You did minimal activities with no depth; no leadership or clinical exposure
20=Participated in a few activities, had clinical exposure, but depth still lacking
30=You participated in a ncie range of activities, some leadership, had meaningful experiences. Strong longitudinal clinical exposure
40=Led/founded group at school, had unique/strong exposure to medicine (ie service trips to other countries), can talk at great length and detail about experience5
50=Did something a newspaper might write about, such as establishing AIDSS centers in Africa

.4(MCAT score)+.3(GPA score)+.15(EC score)+.15(LOR score)=your probability of getting into med school
Guide
~10=Not competitive anywhere
~20=Competitive at low tier or DO schools
~30=Competitive at mid range schools, very competitive at low tier
~40=Competitive at elite schools, very competitive everywhere else
~50=Very competitive everywhere

Problems:
1) Did not take PS or interview into account
2) It would be nice to use functions for the numerical portions (MCA/GPA) as opposed to a “step” system. In other words, devise an equation that would generally fit the points created. So a 32 mcat would not be 40, but maybe 37.2…etc I suck at math though.
3) I’m sure people have differing opinions as to how much weight certain aspects should carry, etc.

Discuss amongst yourselves

 

[Doc.Holliday]

for ECs you leave out research, there seem to be two paths, the research path or the clinical, some do a bit of both, some a lot of one or the other. research definitely deserves a place.

 

[DrHuang]

heh, i just plugged my numbers in and the outcome is similar to whats actually happening.

 

[Blue Planet]

This is an interesting thread ... I know individuals have done this for undergraduate admission.

As for the OP's comment on MCAT scores, I have taken the liberty of graphing the MCAT score function and fitting a curve to it. I thought a linear curve would be good, but the data seemed to indicate a parabolic curve would be better...

Anyway, here's the fomula:

Point value = 0.0468(MCAT score)^2 - 0.4848(MCAT score)
(R^2 = 0.9408)

(I have way too much time on my hands)

 

[Moto]

Punched in my stats, pretty close to my status, but yes your personal statement, life experiences, circumstances could compensate to a small degree. Also, number of schools you apply to could be a factor.

 

[doctorFred]

This is an interesting thread ... I know individuals have done this for undergraduate admission.

As for the OP's comment on MCAT scores, I have taken the liberty of graphing the MCAT score function and fitting a curve to it. I thought a linear curve would be good, but the data seemed to indicate a parabolic curve would be better...

Anyway, here's the fomula:

Point value = 0.0468(MCAT score)^2 - 0.4848(MCAT score)
(R^2 = 0.9408)

(I have way too much time on my hands)

alright then, let me just punch that into my computer real quick.


sorry, i guess i can't seem to crunch the numbers.
has anyone ever told you guys that you're geeks? 😀

 

[DrMassacre]

works for me too

 

[DarkFark]

I think it's alright, but flawed. It works for most people, but it doesn't work on extreme cases. In real life, if you scored a 0 in any of those catagories but well in everything else, the numbers would say you'd be highly competitive, but you'd struggle to get in anywhere. The most realistic example is the all-stats-no-ECs formula. (3.8 GPA, 36 MCAT and no real clinical experience or research). They would haver a tough time surviving an interview at palooka state, but the formula says theyd be competitive for the top tier. In short, the process favors the well rounded, and the formula doesnt.

Someone also mentioned research, which has a difficult-to-chart effect, because its an optional add-on for applying to some schools, but almost a requirement for others, particularly at the top.

You could modify it to take all this (and other) stuff into account, but wouldn't it to be easier to just look at someone's numbers and ECs and estimate it yourself?

 

[medhacker]

LOL

Nice formula, pretty much describes what is happening in my case. Now, how do we define low tier, middle range and elite? Thanks

 

[Einsteinemc2]

I think it's alright, but flawed. It works for most people, but it doesn't work on extreme cases. In real life, if you scored a 0 in any of those catagories but well in everything else, the numbers would say you'd be highly competitive, but you'd struggle to get in anywhere. The most realistic example is the all-stats-no-ECs formula. (3.8 GPA, 36 MCAT and no real clinical experience or research). They would haver a tough time surviving an interview at palooka state, but the formula says theyd be competitive for the top tier. In short, the process favors the well rounded, and the formula doesnt.

Someone also mentioned research, which has a difficult-to-chart effect, because its an optional add-on for applying to some schools, but almost a requirement for others, particularly at the top.

You could modify it to take all this (and other) stuff into account, but wouldn't it to be easier to just look at someone's numbers and ECs and estimate it yourself?

Your comments are, for the most part, well taken. The specific example you give I would disagree with--I think that person would be very competitive at palooka state as long as they had done something. Research and even clinical exposure are not requirements...

However, your point about extremes is right on. So, how do we get around that? What I envision is eventually preparing a spread sheet for people to enter their data. Within the formula we could include "trips" for extreme cases (for example, a 15 mcat would actually subtract from your score, etc).

Finally, the reason I think this is a worthwhile endeavor is that a lot of people are objectively minded; in other words, they like numbers and concrete ideas. So, if we can produce some working model that would actually provide a number (with the disclaimer that you can actually never predict admission), it might benefit certain applicants state of mind.

THe consensus model, however, will take some time. The suggestions so far have been excellent. Of course research must be taken into account. Furthermore, I would eventually like a fxn for the GPA.

With all the brillant minds on SDN, I'm sure we can come up with a viable model.

The ultimate test will be to run the data with tons of applicants who have completed the process to measure its accuracy and precision.

 

[DarkFark]

Your comments are, for the most part, well taken. The specific example you give I would disagree with--I think that person would be very competitive at palooka state as long as they had done something. Research and even clinical exposure are not requirements...

However, your point about extremes is right on. So, how do we get around that? What I envision is eventually preparing a spread sheet for people to enter their data. Within the formula we could include "trips" for extreme cases (for example, a 15 mcat would actually subtract from your score, etc).

Finally, the reason I think this is a worthwhile endeavor is that a lot of people are objectively minded; in other words, they like numbers and concrete ideas. So, if we can produce some working model that would actually provide a number (with the disclaimer that you can actually never predict admission), it might benefit certain applicants state of mind.

THe consensus model, however, will take some time. The suggestions so far have been excellent. Of course research must be taken into account. Furthermore, I would eventually like a fxn for the GPA.

With all the brillant minds on SDN, I'm sure we can come up with a viable model.

The ultimate test will be to run the data with tons of applicants who have completed the process to measure its accuracy and precision.

If you're going to go that route, you probably need some data to fit the curves to. MSAR and mdapplicants might help...

 

[neovenom]

nice try to put a stop to those threads. it's unfortunate that they'll keep coming though. at least the formula was fairly accurate for me.

 

[Law2Doc]

The following is a proposal that would hopefully help end the barrage of “What are my chances” threads. The problem, of course, is that is ultimately impossible to objectively gauge one’s chance to get into medical school… perhaps this can serve as a guide. It is only a framework and not a completed project. I view this as a sort of open-source project for anybody who would want to contribute. I’ll present the idea, and then point out the flaws.

MCAT (40%)
0-22=0
23-24=10
25-26=20
27-29=30
30-34=40
35+=50
(If 40+, bonus of 60!)
GPA (30%)
0-2.5=0
2.5-3.0=10
3.0-3.3=20
3.3-3.6=30
3.6-3.8=40
3.8+=50

LOR (15%)
10=Writers had to ask how they knew you when asked to write the letter
20=Writers knew you from class only, no office hour visits etc (or any other weak relationship)
30=Writers knew you as a person, and you’re pretty sure they wrote positively
40=Writers knew you as a person, you know that they wrote extremely positive letters with hardly any negative comments
50=Same as above, except writers are famous/well known

EC (15%)
0=You did nothing
10= You did minimal activities with no depth; no leadership or clinical exposure
20=Participated in a few activities, had clinical exposure, but depth still lacking
30=You participated in a ncie range of activities, some leadership, had meaningful experiences. Strong longitudinal clinical exposure
40=Led/founded group at school, had unique/strong exposure to medicine (ie service trips to other countries), can talk at great length and detail about experience5
50=Did something a newspaper might write about, such as establishing AIDSS centers in Africa

.4(MCAT score)+.3(GPA score)+.15(EC score)+.15(LOR score)=your probability of getting into med school
Guide
~10=Not competitive anywhere
~20=Competitive at low tier or DO schools
~30=Competitive at mid range schools, very competitive at low tier
~40=Competitive at elite schools, very competitive everywhere else
~50=Very competitive everywhere

Problems:
1) Did not take PS or interview into account
2) It would be nice to use functions for the numerical portions (MCA/GPA) as opposed to a “step” system. In other words, devise an equation that would generally fit the points created. So a 32 mcat would not be 40, but maybe 37.2…etc I suck at math though.
3) I’m sure people have differing opinions as to how much weight certain aspects should carry, etc.

Discuss amongst yourselves

This is not a formula I would rely on. To weight MCAT scores equally in the range of eg 30-34 is just silly -- one end of this range is "average" and the other well above average. The formula is also hugely flawed in areas like the LOR section -- a bad LOR should be worth more like "negative 40" (i.e it could keep you out of med school, as might a bad PS). Also, as you noted the formula assumes the interview is a formality (not part of the formula), and in fact an awful lot of people who pass muster in your formula won't get in thanks to a lukewarm or bad interview while a lot of people who aren't at the top score in your formula who manage to get interviews may get a much better shot. (An argument could be made that, along the numerical scores, the interview is the most important part of the application, not something that can be ignored). Finally, your equation assumes no other factors are considered, which isn't the case - it ignores various blemishes people can have in their records, and things that make them stand out in the applicant pool (other than extracurricularly). This kind of attempt to boil down an applicant's chances into a numerical formula is what creates the misperception seen on here that there is randomness in the system -- when someone who scores well in your formula doesn't get in while someone who scores poorly does, you will say it is random, and that is not really the case. Adcoms are people making partially objective and partially subjective decisions based on their respective school's applicant needs (trying to put together a diverse class), and they are seeing factors not limited to the ones you listed, and so it does not work to try and create an objective formula to attempt to guage their thought processes.

 

[HessExpress]

While on an interview recently, I found it strange how both the interviewer and the admissions rep described how they could "simply use an algorithm" to pick the top 300 students each year and send them acceptances.

They were joking but it was a strange coincidence.

 

[osli]

MSAR would be a good start for data mining. To do better than what they present would require meaningful incorporation of non-numerical data. SDN members would be a good place to gather some of that data.

Instead of taking a guess at the weighting of each category and specific curves within them, the appropriate process would be to start a database with a wide range of options. For instance, a spreadsheet where on person could indicate their GPA, MCAT, undergrad degree, age, gradschool or not, state of residency, estimated competitiveness of schools applied to, where they were/were not accepted and/or invited for interviews, and a list of options describing various combinations of EC's and LOR's from which they could choose the one best describing them.

If you are lucky, you might data mine such a collection and find somewhat reliable patterns. If you are really lucky, a function could handle with some degree of confidence outliers too. That would probably require some piecewise or conditional statements for situations where it appears that an extremely strong showing or extremely glaring lack in one category negates the merit of other categories and produces a step in an otherwise smooth curvefit.

Remember that there a lot of combinations to choose from and there almost certainly isn't enough data just among SDN members to make an effort robust. Any fit, especially one attempting to encompass relatively rare cases, will not likely have a stellar correlation with so few data points.

In my case, I have tons of EC's, work experience and leadership roles in both my academic life and subsequent working life, varied interests and in general am a well rounded person. Zero clinical experience other than some long heart to heart discussions with a few physicians about whether medicine is the right place for me. Where does that fit in the current rating? A function to handle these is possible (but probably not going to be particularly prescient) but complex.

 

[HMSNeuro]

anything ever come of this?

 

[swtiepie711]

/

 

[Depakote]

I scored a 35.5 (somewhere between Competitive at mid range schools, very competitive at low tier and Competitive at elite schools, very competitive everywhere else on your scale) ... I'm on my 3rd round of applications.

Objective assessments do not apply to subjective judgements.

 

[mshheaddoc]

I have to agree with the above poster

 

[GujuMD]

I am a re-applicant (2nd cycle). I scored 37 on the scale...we'll see if the proposal works for me.

 

[adam64897]

I think it's alright, but flawed. It works for most people, but it doesn't work on extreme cases. In real life, if you scored a 0 in any of those catagories but well in everything else, the numbers would say you'd be highly competitive, but you'd struggle to get in anywhere. The most realistic example is the all-stats-no-ECs formula. (3.8 GPA, 36 MCAT and no real clinical experience or research). They would haver a tough time surviving an interview at palooka state, but the formula says theyd be competitive for the top tier. In short, the process favors the well rounded, and the formula doesnt.

Someone also mentioned research, which has a difficult-to-chart effect, because its an optional add-on for applying to some schools, but almost a requirement for others, particularly at the top.

You could modify it to take all this (and other) stuff into account, but wouldn't it to be easier to just look at someone's numbers and ECs and estimate it yourself?

I very much agree. Let's face it, only half of applicants in any given year actually get in to medical school. Therefore, there is a lot of subjectivity in determining those that will make up any given school's class for that year. This formula is a good "rule of thumb" for possibly deciding those who will be interviewed for a spot in a class but not necessarily those who will be offered admission. I have never been a part of an admissions committee but I am willing to bet not too many people applying have GPAs sub 3.0 and MCAT scores lower than 24. Therefore, a lot of people that apply each year have very similar academic credentials and this formula just doesn't allow us to separate those who will be admitted from those who won't based solely on the criteria listed. I would be willing to bet more than half of applicants meet the criteria for being competitive at med. schools described by the key for this formula. I think it mostly boils down to a matter of luck and how well one interviews that determines if he/she will get in to medical school since not everyone can be admitted to medical school that applies especially if objective measuring of one's competitiveness like this formula is used to determine admissions.

I think this formula is very clever though and could possibly be used to determine potential interviewees. Then again a 4.0 GPA with a 0 MCAT awesome LORs and tons of ECs would seem to bid well for an applicant using this formula. Obviously someone with a 0 MCAT would not be given an interview anywhere in the U.S. Maybe include a clause that says if an applicant has a 0 in any category then the entire score is to be 0.

 

[defrunner]

Damn!

You guys are putting so much thought into this, it's scary. I'm not sure yet if this is a good thing for you guys to do or a bad thing. Could be worse, I suppose.

Trying to quantify a crapshoot like this is like trying to make sense of the nonsensical; you can put all your time and effort into it and leave having accomplished exactly nothing. But hey, you get the A for effort!

Think of it this way; here's an example.

The bear jogged along a light pole in the middle of the ocean while timing her pork roast's cooking so that he can fall asleep in the middle of the desert.

Now, try to make sense of that statement. What does it mean?

Absolutely nothing -- I just completely made that up off the top of my head.

Soooo... you can fiddle around all you want with quantifying the qualitative aspects of this process (and I'm sure that even the numerical aspects are extremely accurate themselves), but in the end, it really doesn't mean a thing.

Hey, I do have a sub-3.0 undergrad GPA so it makes it slightly easier for me to not put any stock into things like this -- my chances are very, very slim anyway no matter how I slice it and dice it. Good luck to all of you; if you put half as much effort into your applications as you did with this, I've no doubt that you're getting in somewhere.

 

[WantsThisBad]

The following is a proposal that would hopefully help end the barrage of “What are my chances” threads. The problem, of course, is that is ultimately impossible to objectively gauge one’s chance to get into medical school… perhaps this can serve as a guide. It is only a framework and not a completed project. I view this as a sort of open-source project for anybody who would want to contribute. I’ll present the idea, and then point out the flaws.

MCAT (40%)
0-22=0
23-24=10
25-26=20
27-29=30
30-34=40
35+=50
(If 40+, bonus of 60!)
GPA (30%)
0-2.5=0
2.5-3.0=10
3.0-3.3=20
3.3-3.6=30
3.6-3.8=40
3.8+=50

LOR (15%)
10=Writers had to ask how they knew you when asked to write the letter
20=Writers knew you from class only, no office hour visits etc (or any other weak relationship)
30=Writers knew you as a person, and you’re pretty sure they wrote positively
40=Writers knew you as a person, you know that they wrote extremely positive letters with hardly any negative comments
50=Same as above, except writers are famous/well known

EC (15%)
0=You did nothing
10= You did minimal activities with no depth; no leadership or clinical exposure
20=Participated in a few activities, had clinical exposure, but depth still lacking
30=You participated in a ncie range of activities, some leadership, had meaningful experiences. Strong longitudinal clinical exposure
40=Led/founded group at school, had unique/strong exposure to medicine (ie service trips to other countries), can talk at great length and detail about experience5
50=Did something a newspaper might write about, such as establishing AIDSS centers in Africa

.4(MCAT score)+.3(GPA score)+.15(EC score)+.15(LOR score)=your probability of getting into med school
Guide
~10=Not competitive anywhere
~20=Competitive at low tier or DO schools
~30=Competitive at mid range schools, very competitive at low tier
~40=Competitive at elite schools, very competitive everywhere else
~50=Very competitive everywhere

Problems:
1) Did not take PS or interview into account
2) It would be nice to use functions for the numerical portions (MCA/GPA) as opposed to a “step” system. In other words, devise an equation that would generally fit the points created. So a 32 mcat would not be 40, but maybe 37.2…etc I suck at math though.
3) I’m sure people have differing opinions as to how much weight certain aspects should carry, etc.

Discuss amongst yourselves

Disclaimer: I only read the first 2 lines of your thread.

There will never ever be an end to the "what are my chances" threads.

 

[DoctorPardi]

The way to solve the question of "what are my chances?" is to look in the MSAR. It is easier, and probably more accurate. You're talking about chances at individual schools and not all schools or tiers of schools. The thing is people who post "what are my chances" threads want advice, approval or support.

Posters make these threads to either feel confident about their application, or find out what they can do to make their application better. It would be convenient for schools to use some standard equation for acceptance, but that is not the case. So there is no point in trying to formulate a fake one. We don't want to know what arbitrary number our stats compute to be, but rather what we can do to become more competitive. If not that, posters just want reassurance, that yes their 32 mcat and 3.7 gpa will give them a good chance.

Regardless of what you come up with, or why posters make these posts. There will never be an end to "what are my chances" threads.