How Harvard discriminates against Asian Americans in college admissions

Anonymous
Note that the Republicans (and maybe the Russians) are now using this issue (as they've used the charter school issue) to try to get people who actually like NATO and believe in global warming fighting with each other.

Somehow we have to help kids who live in education deserts, whether in inner cities or in meth suburbs and meth small towns, while also ensuring that bright, hard-working Asian-American kids go to schools that suit their talents and ambitions.

And having a peaceful country that's not in a hot civil war or physically invaded by Russia is good for all of those kids.

If we let Russia herd us into a hot civil war, then the only skills our children will really need are tent making, fire building, hunting, fishing, water purification and warlord dominion management.
Anonymous

It’s not as if Harvard is excluding Asian students. They are an over-represented minority. No one in this thread has offered *principled* explanations of why <22% represents discrimination and/or what % would be considered indicative of a non-racist admissions policy. As the Princeton data indicates, 9 out of 10 kids with perfect test scores and 9 out of 10 kids with a 3.9+ get rejected. This is not an admissions process in which those stats determine who gets in.

And no, the trade-off isn’t between well-rounded kids vs. world class talent. It’s among kids who stand out in different ways. And Harvard is trying to put together a class that is diverse in various ways (including racially). Diversity is a different value than well-roundedness or representativeness. There are various kinds of non-racial balancing going on — e.g. gender and geographic— in a not-always-successful attempt to create a class in which no one demographic or POV feels hegemonic (or completely marginalized).

I agree that the best potential evidence of discrimination here appears to be personality scores, but two things stand out. First, having interviewed for an HYPS, I know our input doesn’t really matter (unless, perhaps, the candidate does something egregiously awful and there’s something else in the file that lends credibility to the interviewer’s account). Personality is assessed primarily through things like letters of rec and essays. Also, the most detailed account I’ve seen (Yang’s NYT op-ed) leads me to think that the disproportionately low personality scores get assigned to Asian applicants who aren’t in the running for admission anyway. Asian kids in the top decile academically are given high personality scores 20+% of the time. The difference is that whites, blacks, and Hispanics in other academic deciles get similar personality scores. It’s a weird stat (and one formulated by the plaintiff’s expert witness), so maybe there’s something more or better that I haven’t seen and I’d be happy to have it pointed out to me.

What I’d love to see is a negotiated settlement involving an experiment in which applications are sufficiently anonymized that applicants are in control of whether colleges know their race and/or gender. (I believe URMs should be recognized as adding diversity to the class, so race can be treated as a positive attribute). How/would that change Harvard College’s demographics?



Are you playing dumb? or are you totally clueless. The reason is that currently Harvard is around 22% Asians. The question is if we removed all traces of race from the application and forced the Admissions office to score these applications, using the same system that they use today, what would be the percentage of Asians in that pool? I am not asking Harvard to use only Academic rating. Let them use their current rating. If you anonymize the applications, would the percentage of Asians increase materially. Only a fool would think it wouldn't. It would be closer to 30% or maybe even 40%. So the only real factor keeping the number of Asians at 22% is racial animus towards them. Nothing else. Everything else is smoke and mirrors to justify the blatant bias.

Harvard can swat that charge away in a second. Agree to score an entire year of applications twice. First have a firm look at all applications and anonymize the applications completely. No trace of race, name or zipcode or anything that could suggest anything about an applicant race should be visible. Now they can pick the class. They then release the data. Lets see what the data shows.
Harvard WILL NEVER DO THIS. Because they know what will happen. Race at Harvard is not a factor. It is THE FACTOR
Anonymous
Anonymous wrote:

It’s not as if Harvard is excluding Asian students. They are an over-represented minority. No one in this thread has offered *principled* explanations of why <22% represents discrimination and/or what % would be considered indicative of a non-racist admissions policy. As the Princeton data indicates, 9 out of 10 kids with perfect test scores and 9 out of 10 kids with a 3.9+ get rejected. This is not an admissions process in which those stats determine who gets in.

And no, the trade-off isn’t between well-rounded kids vs. world class talent. It’s among kids who stand out in different ways. And Harvard is trying to put together a class that is diverse in various ways (including racially). Diversity is a different value than well-roundedness or representativeness. There are various kinds of non-racial balancing going on — e.g. gender and geographic— in a not-always-successful attempt to create a class in which no one demographic or POV feels hegemonic (or completely marginalized).

I agree that the best potential evidence of discrimination here appears to be personality scores, but two things stand out. First, having interviewed for an HYPS, I know our input doesn’t really matter (unless, perhaps, the candidate does something egregiously awful and there’s something else in the file that lends credibility to the interviewer’s account). Personality is assessed primarily through things like letters of rec and essays. Also, the most detailed account I’ve seen (Yang’s NYT op-ed) leads me to think that the disproportionately low personality scores get assigned to Asian applicants who aren’t in the running for admission anyway. Asian kids in the top decile academically are given high personality scores 20+% of the time. The difference is that whites, blacks, and Hispanics in other academic deciles get similar personality scores. It’s a weird stat (and one formulated by the plaintiff’s expert witness), so maybe there’s something more or better that I haven’t seen and I’d be happy to have it pointed out to me.

What I’d love to see is a negotiated settlement involving an experiment in which applications are sufficiently anonymized that applicants are in control of whether colleges know their race and/or gender. (I believe URMs should be recognized as adding diversity to the class, so race can be treated as a positive attribute). How/would that change Harvard College’s demographics?



Are you playing dumb? or are you totally clueless. The reason is that currently Harvard is around 22% Asians. The question is if we removed all traces of race from the application and forced the Admissions office to score these applications, using the same system that they use today, what would be the percentage of Asians in that pool? I am not asking Harvard to use only Academic rating. Let them use their current rating. If you anonymize the applications, would the percentage of Asians increase materially. Only a fool would think it wouldn't. It would be closer to 30% or maybe even 40%. So the only real factor keeping the number of Asians at 22% is racial animus towards them. Nothing else. Everything else is smoke and mirrors to justify the blatant bias.

Harvard can swat that charge away in a second. Agree to score an entire year of applications twice. First have a firm look at all applications and anonymize the applications completely. No trace of race, name or zipcode or anything that could suggest anything about an applicant race should be visible. Now they can pick the class. They then release the data. Lets see what the data shows.
Harvard WILL NEVER DO THIS. Because they know what will happen. Race at Harvard is not a factor. It is THE FACTOR


+1000. Simply knowing the race of the applicant affects how Harvard scores them. There is either severe unconscious racial bias or blatant race balancing going on. Both of which is illegal. It's like the Heisenberg college uncertainty principle is at play here for the Asian applicant. You can be Asian or have high personality scores, but not both. Once the Harvard admission reader realizes that the applicant is Asian, they somehow find ways to eff them over, consciously or unconsciously.
Anonymous
Anonymous wrote:

It’s not as if Harvard is excluding Asian students. They are an over-represented minority. No one in this thread has offered *principled* explanations of why <22% represents discrimination and/or what % would be considered indicative of a non-racist admissions policy. As the Princeton data indicates, 9 out of 10 kids with perfect test scores and 9 out of 10 kids with a 3.9+ get rejected. This is not an admissions process in which those stats determine who gets in.

And no, the trade-off isn’t between well-rounded kids vs. world class talent. It’s among kids who stand out in different ways. And Harvard is trying to put together a class that is diverse in various ways (including racially). Diversity is a different value than well-roundedness or representativeness. There are various kinds of non-racial balancing going on — e.g. gender and geographic— in a not-always-successful attempt to create a class in which no one demographic or POV feels hegemonic (or completely marginalized).

I agree that the best potential evidence of discrimination here appears to be personality scores, but two things stand out. First, having interviewed for an HYPS, I know our input doesn’t really matter (unless, perhaps, the candidate does something egregiously awful and there’s something else in the file that lends credibility to the interviewer’s account). Personality is assessed primarily through things like letters of rec and essays. Also, the most detailed account I’ve seen (Yang’s NYT op-ed) leads me to think that the disproportionately low personality scores get assigned to Asian applicants who aren’t in the running for admission anyway. Asian kids in the top decile academically are given high personality scores 20+% of the time. The difference is that whites, blacks, and Hispanics in other academic deciles get similar personality scores. It’s a weird stat (and one formulated by the plaintiff’s expert witness), so maybe there’s something more or better that I haven’t seen and I’d be happy to have it pointed out to me.

What I’d love to see is a negotiated settlement involving an experiment in which applications are sufficiently anonymized that applicants are in control of whether colleges know their race and/or gender. (I believe URMs should be recognized as adding diversity to the class, so race can be treated as a positive attribute). How/would that change Harvard College’s demographics?



Are you playing dumb? or are you totally clueless. The reason is that currently Harvard is around 22% Asians. The question is if we removed all traces of race from the application and forced the Admissions office to score these applications, using the same system that they use today, what would be the percentage of Asians in that pool? I am not asking Harvard to use only Academic rating. Let them use their current rating. If you anonymize the applications, would the percentage of Asians increase materially. Only a fool would think it wouldn't. It would be closer to 30% or maybe even 40%. So the only real factor keeping the number of Asians at 22% is racial animus towards them. Nothing else. Everything else is smoke and mirrors to justify the blatant bias.

Harvard can swat that charge away in a second. Agree to score an entire year of applications twice. First have a firm look at all applications and anonymize the applications completely. No trace of race, name or zipcode or anything that could suggest anything about an applicant race should be visible. Now they can pick the class. They then release the data. Lets see what the data shows.
Harvard WILL NEVER DO THIS. Because they know what will happen. Race at Harvard is not a factor. It is THE FACTOR


+1 Although I do agree with the PP that suggested race still be factored in regarding URMs, otherwise the representation of black and latino students would diminish too drastically. IIRC those populations were halved when California implemented race blind admissions.
Anonymous
Anonymous wrote:
Anonymous wrote:

It’s not as if Harvard is excluding Asian students. They are an over-represented minority. No one in this thread has offered *principled* explanations of why <22% represents discrimination and/or what % would be considered indicative of a non-racist admissions policy. As the Princeton data indicates, 9 out of 10 kids with perfect test scores and 9 out of 10 kids with a 3.9+ get rejected. This is not an admissions process in which those stats determine who gets in.

And no, the trade-off isn’t between well-rounded kids vs. world class talent. It’s among kids who stand out in different ways. And Harvard is trying to put together a class that is diverse in various ways (including racially). Diversity is a different value than well-roundedness or representativeness. There are various kinds of non-racial balancing going on — e.g. gender and geographic— in a not-always-successful attempt to create a class in which no one demographic or POV feels hegemonic (or completely marginalized).

I agree that the best potential evidence of discrimination here appears to be personality scores, but two things stand out. First, having interviewed for an HYPS, I know our input doesn’t really matter (unless, perhaps, the candidate does something egregiously awful and there’s something else in the file that lends credibility to the interviewer’s account). Personality is assessed primarily through things like letters of rec and essays. Also, the most detailed account I’ve seen (Yang’s NYT op-ed) leads me to think that the disproportionately low personality scores get assigned to Asian applicants who aren’t in the running for admission anyway. Asian kids in the top decile academically are given high personality scores 20+% of the time. The difference is that whites, blacks, and Hispanics in other academic deciles get similar personality scores. It’s a weird stat (and one formulated by the plaintiff’s expert witness), so maybe there’s something more or better that I haven’t seen and I’d be happy to have it pointed out to me.

What I’d love to see is a negotiated settlement involving an experiment in which applications are sufficiently anonymized that applicants are in control of whether colleges know their race and/or gender. (I believe URMs should be recognized as adding diversity to the class, so race can be treated as a positive attribute). How/would that change Harvard College’s demographics?



Are you playing dumb? or are you totally clueless. The reason is that currently Harvard is around 22% Asians. The question is if we removed all traces of race from the application and forced the Admissions office to score these applications, using the same system that they use today, what would be the percentage of Asians in that pool? I am not asking Harvard to use only Academic rating. Let them use their current rating. If you anonymize the applications, would the percentage of Asians increase materially. Only a fool would think it wouldn't. It would be closer to 30% or maybe even 40%. So the only real factor keeping the number of Asians at 22% is racial animus towards them. Nothing else. Everything else is smoke and mirrors to justify the blatant bias.

Harvard can swat that charge away in a second. Agree to score an entire year of applications twice. First have a firm look at all applications and anonymize the applications completely. No trace of race, name or zipcode or anything that could suggest anything about an applicant race should be visible. Now they can pick the class. They then release the data. Lets see what the data shows.
Harvard WILL NEVER DO THIS. Because they know what will happen. Race at Harvard is not a factor. It is THE FACTOR


+1000. Simply knowing the race of the applicant affects how Harvard scores them. There is either severe unconscious racial bias or blatant race balancing going on. Both of which is illegal. It's like the Heisenberg college uncertainty principle is at play here for the Asian applicant. You can be Asian or have high personality scores, but not both. Once the Harvard admission reader realizes that the applicant is Asian, they somehow find ways to eff them over, consciously or unconsciously.


No - both of which is legal. Where in the world do you get that use of race in college admissions is illegal?
Anonymous
Harvard wants students who will pay off for them in some way in the end. Whichever race they may be. They do not seem to want kids who have been prepped within an inch of their lives. Whichever race they may be. It just so happens many of their applicants are. I’m sure the parents of these students would love for admission to Harvard to be based on one test. It’s how it’s done in China. It’s not a mess over there at all... google Chinese cram schools.
Anonymous


No - both of which is legal. Where in the world do you get that use of race in college admissions is illegal?


No. You are wrong. Read what I said carefully. The Supreme Court has said that using race as A FACTOR is legal. Racial Balancing is using race based quotas. That is illegal. and deliberately scoring Asians low on certain scores to bring their admission rates down so that you can balance your class racially is also illegal.
Anonymous
Anonymous wrote:

It’s not as if Harvard is excluding Asian students. They are an over-represented minority. No one in this thread has offered *principled* explanations of why <22% represents discrimination and/or what % would be considered indicative of a non-racist admissions policy. As the Princeton data indicates, 9 out of 10 kids with perfect test scores and 9 out of 10 kids with a 3.9+ get rejected. This is not an admissions process in which those stats determine who gets in.

And no, the trade-off isn’t between well-rounded kids vs. world class talent. It’s among kids who stand out in different ways. And Harvard is trying to put together a class that is diverse in various ways (including racially). Diversity is a different value than well-roundedness or representativeness. There are various kinds of non-racial balancing going on — e.g. gender and geographic— in a not-always-successful attempt to create a class in which no one demographic or POV feels hegemonic (or completely marginalized).

I agree that the best potential evidence of discrimination here appears to be personality scores, but two things stand out. First, having interviewed for an HYPS, I know our input doesn’t really matter (unless, perhaps, the candidate does something egregiously awful and there’s something else in the file that lends credibility to the interviewer’s account). Personality is assessed primarily through things like letters of rec and essays. Also, the most detailed account I’ve seen (Yang’s NYT op-ed) leads me to think that the disproportionately low personality scores get assigned to Asian applicants who aren’t in the running for admission anyway. Asian kids in the top decile academically are given high personality scores 20+% of the time. The difference is that whites, blacks, and Hispanics in other academic deciles get similar personality scores. It’s a weird stat (and one formulated by the plaintiff’s expert witness), so maybe there’s something more or better that I haven’t seen and I’d be happy to have it pointed out to me.

What I’d love to see is a negotiated settlement involving an experiment in which applications are sufficiently anonymized that applicants are in control of whether colleges know their race and/or gender. (I believe URMs should be recognized as adding diversity to the class, so race can be treated as a positive attribute). How/would that change Harvard College’s demographics?



Are you playing dumb? or are you totally clueless. The reason is that currently Harvard is around 22% Asians. The question is if we removed all traces of race from the application and forced the Admissions office to score these applications, using the same system that they use today, what would be the percentage of Asians in that pool? I am not asking Harvard to use only Academic rating. Let them use their current rating. If you anonymize the applications, would the percentage of Asians increase materially. Only a fool would think it wouldn't. It would be closer to 30% or maybe even 40%. So the only real factor keeping the number of Asians at 22% is racial animus towards them. Nothing else. Everything else is smoke and mirrors to justify the blatant bias.

Harvard can swat that charge away in a second. Agree to score an entire year of applications twice. First have a firm look at all applications and anonymize the applications completely. No trace of race, name or zipcode or anything that could suggest anything about an applicant race should be visible. Now they can pick the class. They then release the data. Lets see what the data shows.
Harvard WILL NEVER DO THIS. Because they know what will happen. Race at Harvard is not a factor. It is THE FACTOR



The assumption here is that the only correct way to assemble a class of students is using some arbitrary metric of school success (i.e. grades or standardized test results). Since these grades and test results can be strongly correlated with a culture of emphasizing education, and these cultures can be identified with ethnic/religious identities (e.g. Jews, Han Chinese, Japanese, SE Indians etc), then it is asserted that if these "groups" are not represented in the Harvard class that there is de facto "discrimination". It is also claimed that it is "just" to grant preferential admission to other "groups" (e.g. Latino, African American) who may or may not have cutural imperatives to emphasize educational attainment. Simple thought experiment - why should a "black student" whose mother is Jewish and who is raised in D.C. with a HHI of >$250k and attends the best schools and who works hard at school and extracurricular activities get preference in admissions to any other candidate - simply on the basis of skin color?? How is is correct to seek racial justice by discriminating against (or in favor of) a racial minority?
Anonymous
Anonymous wrote:


No - both of which is legal. Where in the world do you get that use of race in college admissions is illegal?


No. You are wrong. Read what I said carefully. The Supreme Court has said that using race as A FACTOR is legal. Racial Balancing is using race based quotas. That is illegal. and deliberately scoring Asians low on certain scores to bring their admission rates down so that you can balance your class racially is also illegal.


Perfectly legal. And balancing, i.e., making sure all races are more proportionally represented is not a quota.
You're talking out of the plaintiff's play-book here, and they will lose. The use of race in the fashion Harvard is doing it to have a balanced class of the type students they want is perfectly legal.
Anonymous
Here is a simple way to fix this. Harvard first sets a floor on what it will take to be considered "qualified" to attend Harvard. They can set this floor however they like, but it has to be the same for all races. Then have a third party firm, anonymize ALL Applications. They read through the application and remove anything that would help identify the race of the applicant, including zipcode. They also toss out any application that does not meet the floor.

They then forward the apps to Harvard. Harvard can use any criteria it wants to pick its class. They assign ranks to all their applicants. and pick the top 2000. After this is done, only the URM applicant pool (but not names) is revealed to them. So they will know that applicant #31938 is black from North Dakota. and 20,987 is Hispanic from Texas.

Now lets say that the resulting class is 2% black and they want to increase that to 6%. They now have a choice.

Eject one non URM applicant out of the pool and get one URM in or balance between one URM and another. They don't know who they are ejecting. They only know their ranking in the pool. They stop when their diversity goal is reached.

That should be the final admitted class.
Anonymous
What if the anonymous assessment idea results in blacks and Hispanics being 6 or 7 percent of the class? Then wont all hell break loose in the other direction? And how many first generation kids will get accepted? The plaintiffs here really look like they're against inclusivity and diversity.
Anonymous
The bias is not used to determine the development candidates. The incredibly rich kids get in, with much lower standards, and that’s life. The reason the bias against Asians exist is because they want to preserve their culture. That’s also why they use legacy so much. Legacy is also another way to favor their longstanding culture and yes can race is a part of culture. Do you know why MIT’s demographics are really different than Harvard? Because they aren’t interested in culture. They are interested in have the best math and science candidates. The culture of institutions like Harvard and Yale are built on wealthy white elite culture. It’s still amazing that a supposedly liberal community like the DC metro area doesn’t understand the idea of white privilege. When the top of the pyramid is white- CEOs, University Deans, even just the overwhelmingly white alumni reading apps, the bias is just simply there. Harvard’s own internal study found it. Unfortunately, instead of just addressing it, they buried it.

Anonymous
What about Yale?
Anonymous
Anonymous wrote:Here is a simple way to fix this. Harvard first sets a floor on what it will take to be considered "qualified" to attend Harvard. They can set this floor however they like, but it has to be the same for all races. Then have a third party firm, anonymize ALL Applications. They read through the application and remove anything that would help identify the race of the applicant, including zipcode. They also toss out any application that does not meet the floor.

They then forward the apps to Harvard. Harvard can use any criteria it wants to pick its class. They assign ranks to all their applicants. and pick the top 2000. After this is done, only the URM applicant pool (but not names) is revealed to them. So they will know that applicant #31938 is black from North Dakota. and 20,987 is Hispanic from Texas.

Now lets say that the resulting class is 2% black and they want to increase that to 6%. They now have a choice.

Eject one non URM applicant out of the pool and get one URM in or balance between one URM and another. They don't know who they are ejecting. They only know their ranking in the pool. They stop when their diversity goal is reached.

That should be the final admitted class.


you're kidding? It's 15% now. Can you imagine the criticism Harvard will get if it's incoming freshmen class from now on is only 6% black?
Anonymous
I’m not a fool or naive or clueless or playing dumb. I’m someone who understands the elite college admissions process, constitutional law, and statistical work in the social sciences. With this background, I can tell you that the evidence presented here is not compelling. And it’s really hard to argue racial animus wrt a group that has been very successful in the admissions process.

Unconscious bias is certainly possible, which is why I proposed anonymization as an experimental method. At a minimum, it would be interesting to give essays and letters of rec to Harvard admissions officers in an anonymized form and see if Asians get higher scores or whites lower scores when those inputs get translated into a number representing personal attributes.

But it’s entirely possible, for example, that fewer Asian applicants write essays/tell stories about themselves that display the traits admissions officers are looking for (humor, empathy, rebounding from failure, and an abundance of self-confidence) because they don’t expect those attributes to be relevant in this context. Or that Asian students disproportionately go to public schools with under-resourced college counseling offices and have teachers, especially in STEM fields, who don’t highlight those attributes.

I also suspect that Asian-Americans are disadvantaged by Harvard’s interest in geographical diversity and informal caps on the number of kids taken from each HS.
post reply Forum Index » College and University Discussion
Message Quick Reply
Go to: