"Race" in College Applications FAQ & Discussion 12

Applying to College
4001

Medical care often does involve close personal interaction between the provider and patient, so it is likely to prejudicial assumptions (by either party) can be triggered quite easily.

https://www.thecut.com/2017/11/too-many-doctors-still-believe-dangerous-racial-stereotypes.html
https://www.statnews.com/2017/10/18/patient-prejudice-wounds-doctors/

In the local context of medical care, there may be no good way to solve these types of problems, although there may be some ways that are less bad than others.

4002

The issue with getting a proportional amount of URMs into medical school deals with a pipeline that needs to be fixed way before medical school admissions. Trying to fix medical admissions instead of the actual underlying problems that lead to pipeline issues has not really worked for elite school admissions (Those underrepresented are still underrepresented after 65 years, while those at the bottom tend to stay there with current AA policy) and will probably fail for medical school admissions as well. One possible alternative would be to put more emphasis in medical school on physician/patient interactions.

4003

The Notre Dame study did not separate by race. The average stats by race for the Duke study are below. The HS curriculum was on a 1-5 scale. Non-URMs averaged >4.7 out of 5, and Black students averaging 4.3 out of 5. This measure essentially flags the small minority of Duke students who do not get the maximum 5 rating for HS curriculum, which are far more likely to occur among Black students than non-URMs. Note that while gender and HS curriculum rating were the most significant analyzed admission predictors of engineering attrition with full controls, they were not the only contributing factors. All the other listed application ratings except personal qualities appear to have some degree of influence, even though most did not reach statistical significance at a 90% level.

**Average Duke Sample Group Stats/b
Percent Female – Black: 72%; White: 48%; Asian: 45%
Average SAT – Black: 1275; White: 1416; Asian: 1457
HS Curriculum (1-5) – Black: 4.3; White 4.7; Asian 4.9
HS Achievement (1-5) – Black: 3.7; White 4.2; Asian 4.6
Essays (1-5) – Black: 3.1; White 3.4; Asian 3.7
LORs (1-5) – Black: 3.5; White 3.8; Asian 3.9
Personal Qualities (1-5) – Black: 3.2; White 3.5; Asian 3.4

4004

While NEJM is a highly respected medical journal, I was originally unconvinced by this opinion piece. However, I then started to read the underlying referenced study (www.nber.org/papers/w24787.pdf) and this appears to be a solid study. I am generally a strong opponent of AA, but I could be convinced that it makes sense here.

Medicine is a bit different. A person deciding to go into math, physics, finance, computer science, consulting, investment banking, etc. is for the most part just making an individual career decision. A person going into medicine is both making an individual career decision and directly impacting the health of others. If the race of the doctor truly makes a difference for certain patients then that is something to be considered, as long as a minimum quality level can be maintained.

In any case, I suspect that the results of this study are interesting enough that others will try to replicate the results. Let’s see if they are able to do so.

4005

Some other professions where one interacts with the general public may have similar characteristics. For example, if a city has a lot of racism and segregated neighborhoods, would police officers who are of the same race as more of the people in the neighborhood be more effective and more trusted? But would hiring police officers and assigning their patrol neighborhoods by race be illegally discriminatory with respect to the officers themselves? And would catering to the pre-existing racism to address a near term problem reinforce the racism over the longer term?

4006

Representation of African-Americans in police forces is pretty much at parity with the population as a whole – about 13% – see https://datausa.io/profile/soc/333050/

So I don’t think the analogy really holds.

I’d add that there are ways that medical schools can increase outreach. I know a doctor couple who came through the Charles Drew program in Los Angeles - so their medical training is via UCLA, but their path toward getting there is somewhat different. (Current stats for Drew College of Medicine are 47% African American, 43% Hispanic … 1% White)

4007

I am looking at this issue holistically and I think this is what is happening.

I know from looking at decades of research that there are two main components to academic and job success. General learning aptitude (R=0.5 or more ) and conscientiousness (R=.2 to .4). Since the two are not well correlated, the sum of the two + a bit of luck is all one needs to explain the problem.

What I think the colleges are trying to do is to hack the system. They are data-mining for ways to get more URM into the school without having to improve their academic qualification. If all they are interested in is to increase URM enrollment and up their graduation rate, all they have to do is to offer more majors where grading is soft, easy, and subjective. There is no need to torture the data until they confess.

Here is an interesting article on college grading. I suspect similiar trend exists in high school as well:

http://conversableeconomist.blogspot.com/2011/11/grade-inflation-and-choice-of-major.html

All in all, they are not trying to solve the problem; they are trying to hide it.

4008

@calmom I am not sure if Charles Drew is a good example because it is a HBCU. HBCUs have a mission dedicated to helping historical underrepresented groups become educated and serve our communities, but most schools have different institutional goals. The 4 HBCU Medical schools (Meharry Medical College, Howard University College of Medicine, Morehouse School of Medicine, and Charles R. Drew University of Medicine and Science) graduate a much larger footprint of African American doctors than would ever be expected from 4 medical schools. The admissions competition for Medical school is fierce and hitting what medical schools are look for (GPA/MCAT score/volunteering/research) along with the debt collected from years of schooling affects the number of URM students that even attempt to become doctors.

@Canuckguy My favorite movie line of all time sums up what is going on with grade inflation. “The greatest trick the devil ever pulled was convincing the world he didn’t exist,” Roger “Verbal” Kint, The Usual Suspect. If everyone has an A, it means that no one is “unqualified” for elite college admissions, especially in the absence of standardized testing. It also can hide the inferiority of the education received at terrible schools and can make it very hard to differentiate between students academically. The difference is today’s schools are not as good at hiding in plain sight as Keyser Söze was.

4009

Agree with you, changethegame, that the problem is definitely much earlier in the pipeline.

But what we’ve got now is a marked discrepancy by race in MCAT scores and grades for medical school applicants who are accepted:

Average MCAT scores (percentiles) for applicants who were accepted (2014-16):
Asians 32.8 (90%)
White 31.7 (85%)
Hispanic 28.1 (68%)
African American 27.3 (62%)

Also to have a 50% change of getting accepted an african american needs to have a
GPA of 2.4-2.6 if one has an MCAT of 33-35 or a
GPA 2.89-2.99 if one has an MCAT of 27-29

whereas an asian american needs to have a
GPA of 3.2-3.39 if one has an MCAT of 36-38
GPA of 3.8-4.0 if one has an MCAT of 27-29

Found these figures from a poster on a different website, but I checked these numbers from the www.aamc.org tables and these appear correct: https://www.aamc.org/download/321512/data/factstablea24-1.pdf, download tables 24-1 to 24-4).

While the Association of American Medical Colleges (AAMC) states in their recent reports that “Research shows that physician diversity adds value to the health-care system by expanding access to health care… Racial and ethnic minority physicians are more likely to practice primary care than their white peers. Black or African-American, Hispanic or Latino, and American Indian or Alaska Native physicians are also more likely to practice in medically underserved areas.”

What they don’t say is that the top medical students (by grades or by final summary recommendation from the medical school) generally go into subspecialty practice (think neurosurgery or plastic surgery) and very, very few practice general family care in medically underserved areas.

A really cynical (but perhaps not entirely untrue) interpretation of the AAMC report would be … we need to have diversity in admitting medical students because those that do relatively poorly in medical school (more often URMs) are those who are more likely to practice primary care in medically underserved areas.

The flip side argument of course, is who cares about super high MCAT scores and grades if it’s just one more white/asian kid (and the occasional affluent URM) who will just end up as a high priced dermatologist or plastic surgeon in manhattan? The goal of course is to get smart docs, of all races, willing to practice in medically underserved areas. Maybe a program that offers free medical school tuition for those willing to go to underserved areas, or a federal program that attaches research dollars for medical schools to a requirement for medical care in underserved areas might help.

I think the point is, we DO want to have higher rates of URMs in tippy top colleges and medical schools, and we DO want to have smart docs of all races taking care of patients where they are most needed. The question is whether the social engineering programs of “holistic” admissions accomplishes these aims or whether they backfire terribly and result in cynicism, more tribalism, and in certain ways actually harm the group they are attempting to help.

I can see that affirmation action might in some form be worthwhile. In fact, I’m not sure that even quotas by race is such a horrible idea (as long as its called what it is). What gets me is the idea that “holistic” review has somehow evolved to be the best we can do, and therefore doesn’t need to examined fully and transparently and often to see if we can improve the process for everyone.

4010

If I needed brain surgery, I would probably want the smartest, most brilliant neurosurgeon I could find. But that’s not the criteria I use for my GP – there, I want someone who will listen to me. Someone who will take the time to answer my question and explain what is going on, and someone who will talk to me in a kind and respectful way. Same deal with a pediatrician or family doctor when my kids were growing up. Personal qualities & bedside manner means a lot for primary level health care. I don’t need a rocket scientist to treat my kid’s ear infection.

Those personal qualities are things that might be considered with a holistic approach – qualities that may be most evident through LOR’s or personal interviews – so if you want good primary care doctors meeting the needs of the population as a whole, then holistic admissions will help reach that goal.

4011

For a surgeon, wouldn’t you also want steady hands?

For a primary care physician, wouldn’t you also want one who is broadly knowledgeable (in addition to being good at talking with patients), since you may come in with anything that s/he has to figure out what is? Even if s/he refers you to a specialist, you want him/her to refer you to the correct kind of specialist.

4012

@ucbalumnus - I was thinking the same thing, our primary care physician diagnosed our 6 year old with a major illness and it was very good that she caught it early.

4013

I watched the video and it was reminded me of some of those nights. So stressful for all.

I know that all three are star students. But my take is that it is not accurate to say that Ryan was clearly the best candidate.

Che had a 1580 SAT , Cai had a 36 superscored ACT and Henry had a 1460 SAT and 34 ACT with 800 in math subject test.

I’m made those up. Just like our judgements. Maybe it was the opposite.

They were all great. One great kid got in. Not sure if he’s a better writer or had better scores or both. Or maybe race was the key with all the rest.

No dog in the hunt. But to say who was more qualified is a completely unqualified assessment

4014

just looking at these stats it looks like Asians need higher scores than whites to gain admission to medical school. Asian discrimination vs whites is not just a Harvard problem but no doubt an issue in all the professional schools as well.

4015

Employee job performance is not the same as college academic success. Studies of college success tend to find different results. The corresponding metric to job performance during college would be college GPA, and the corresponding metric to GLA would be SAT score. The first study of college GPA that came up in a Google search is https://journals.sagepub.com/doi/full/10.1177/2332858416670601 . It looked at how well HS GPA, SAT, and standardized HS math and English subject testing predicted first year GPA among 10k kids in the CUNY system and 10k students at Kentucky public colleges. Results are summarized below. In both groups, SAT was the weakest analyzed predictor and added relatively little to the prediction beyond the other analyzed variables. While this example is not representative of all colleges, particularly selective privates, the point is it’s not a given that GLA type standardized test result is the primary component to explaining college academic success, particularly for URMs.

CUNY System FY GPA
SAT explains 14% of variance in first year GPA
HS NYS Regents test explains 16% of variance in first year GPA
HS GPA explains 25% of variance in first year GPA
HS GPA + SAT explains 28% of variance in first year GPA

Kentucky Public Colleges FY GPA
SAT explains 16% of variance in first year GPA
HS KCCT test explains 17% of variance in first year GPA
HS GPA explains 32% of variance in first year GPA
HS GPA + SAT explains 34% of variance in first year GPA

The first study that came up in search for conscientiousness at http://citeseerx.ist.psu.edu/viewdoc/download?doi=10.1.1.419.3808&rep=rep1&type=pdf was more surprising to me, It found a combination of self and informant conscientiousness was a notable predictor of college GPA, on par with more traditional metrics. However, the informant conscientiousness was far more useful than the self report, which is difficult to obtain accurately in selective college admission. For example, I wouldn’t assume a LOR or alumni interview will give an accurate and unbiased review of applicant conscientiousness. Note that sample size only 131.

UCR Cumulative GPA
SAT explains 10% of variance in cumulative GPA
SAT + HS GPA explains 28% of variance in cumulative GPA
SAT + HS GPA + Conscientiousness 38% of variance in cumulative GPA

At selective colleges like Duke, I expect few of the students who are switching out of quantitative majors are doing so because they are failing out and/or unable to graduate in their initially planned major. For example, the Duke study mentions Black students had an average first year GPA of ~2.9, while white students had a GPA of ~3.3. A first year 2.9 GPA is well above failing and not too low to allow a career in tech, but students who do not pull up GPA to 3.0 by graduation may have issues with resume filters when applying for a job. This low a GPA would also be problematic for more selective graduation/professional schools.

I expect some of the switching out more relates to getting a lower grade in engineering (or other) than they received in their other classes, even if that grade was well above failing. For example, when I was a freshman at Stanford, they had psych counselors sent to our dorm as preventative measures after the first midterms because it was relatively common for students to freak out after getting their first ever B/B+. I’d expect some students whose first ever B occurs in an intro engineering type class decide engineering is not the right major for them. The Duke study found that harness of grading was one of the more significant factors in switch out behavior. Many switch out for reasons unrelated to GPA as well.

Another factor related to URM graduation rate is gender balance. 72% of Black students in the Duke sample were female, while 45-49% of all other races were female. That’s quite a difference. The Cornell study offers a reason why that difference may exist. It found that SAT score and class rank were not statistically significant predictors of GPA and graduation rate among Black students. The author writes, “We also wonder why SAT scores and high school rank are significant predictors of cumulative GPA for all groups except for blacks… These findings lead to a complex question: why is it that admissions characteristics, generally, predict college level performance for all groups except blacks?” The only noteworthy predictor of graduation rate the Cornell study found for Black students was being female. This implies one way to increase URM graduate rate at a selective college would be favoring admitting Black women over Black men; which would lead to a male/female ratio similar to the 72% female in the Duke sample… Or it may just be a coincidence. The Harvard lawsuit analysis did not find that Black women were favored over men, and most selective colleges show a reasonable gender balance between Black males and females.

4016

@ucbalumnus It’s pretty hard to complete any medical school without being broadly knowledgeable.

It seems that people on this thread have confused or confounded the concepts of “capable” vs. “highest scoring.”

When college or graduate admissions are highly competitive, they end up excluding many very capable students.

I don’t know of any study that correlates MCAT scores with treatment outcomes or patient satisfaction.

Given the amount of medical training that depends on practical, hospital-based work through rotations and residency, I’d also think interpersonal skills and work ethic would be particularly important to long-term success — and a lot of the work along the way is scut work – lower level, repetitive tasks that somebody has got to do, but is not something most would consider intellectually challenging. So really – if I had access that sort of information, I’d be a lot more interested in the subjective evaluations during rotations and residency than test scores. Because a lot of very brilliant doctors make horrible mistakes because they aren’t listening to what their patients or staff are trying to tell them.

4017

“I don’t know of any study that correlates MCAT scores with treatment outcomes or patient satisfaction.”

If your point is that people don’t generally ask a doctor their MCAT scores when deciding on a doctor, agree there. But we assume I think that the undergrad and med schools do a good job of vetting the pre-meds and medical students.

“Because a lot of very brilliant doctors make horrible mistakes because they aren’t listening to what their patients or staff are trying to tell them”

Are you implying that brilliant doctors make more mistakes than non-brilliant ones because they may be less humble and not as open to feedback, criticism etc? That seems a bit of a stretch since med schools place a lot of emphasis on the interview to look for that kind of attitude. I’m not saying there aren’t any of course.

“interpersonal skills and work ethic would be particularly important to long-term success”

Sure, but again you imply that the high scorers don’t have as good interpersonal skills and work ethic as the low scorers/low gpa. That would seem to contradict the large amount of work needed to get through a pre-med program. Again I would think the med school would be able to differentiate a 3.8 lazy applicant in a fluff program vs a 2.8 hard working applicant from a tough program. I’m probably depending too much on med schools, I do realize that.

4018

The 2.8 student wouldn’t get an interview, anywhere.

4019

“The 2.8 student wouldn’t get an interview, anywhere.”

Not true. That’s highly dependent on the race of the applicant. Read Data10’s post #4008 above. African Americans with that GPA have a 50% chance of med school admission, so they’re getting interviews somewhere.

4020

A grid showing the acceptance rate by MCAT and GPA for Black medical school applicants is at https://www.aamc.org/download/321514/data/factstablea24-2.pdf and summarized below. This suggests a decent chance of acceptance with a 2.8 GPA, if MCAT is 27+. 27 was a slightly below average score (across all races) during the listed admission cycle.

Black Applicants
2.8 GPA + 33-35 MCAT – 73% acceptance rate
2.8 GPA + 30-32 MCAT – 52% acceptance rate
2.8 GPA + 27-29 MCAT – 49% acceptance rate
2.8 GPA + 24-26 MCAT – 31% acceptance rate
2.8 GPA + 21-23 MCAT – 10% acceptance rate

White Applicants
2.8 GPA + 33-35 MCAT – 23% acceptance rate
2.8 GPA + 30-32 MCAT – 22% acceptance rate
2.8 GPA + 27-29 MCAT – 12% acceptance rate
2.8 GPA + 24-26 MCAT – 7% acceptance rate
2.8 GPA + 21-23 MCAT – 2% acceptance rate