The shell game of Berkeley’s holistic admissions

By Razib Khan | August 3, 2013 3:36 am

The title refers to the basic thrust of a piece in The New York Times, Confessions of an Application Reader. The piece ends with a paragraph like so:

Underrepresented minorities still lag behind: about 92 percent of whites and Asians at Berkeley graduate within six years, compared with 81 percent of Hispanics and 71 percent of blacks. A study of the University of California system shows that 17 percent of underrepresented minority students who express interest in the sciences graduate with a science degree within five years, compared with 31 percent of white students.

You may or may not agree with this particular type of admissions policy (I do not, because I do not care if minorities are underrepresented at universities if that underrepresentation is due to transparent academic deficiencies, which I believe to be the case). Rather, I want to focus on the term ‘underrepresented minorities’ and ascertain how underrepresented minorities truly are at Berkeley. That’s easy enough to do. About ~80% of UC Berkeley undergraduates are California residents. The Census allows us to query the racial makeup of a range of age brackets for various localities. What I did was look for the percentage of individuals between the ages of 15-19 in the 2010 Census for California, approximately the source population of students who are freshman in the 2012 class.

Here are the percentages in the 15-19 age category in 2010:

30% Non-Hispanic White
6.5% Non-Hispanic Black
11% Non-Hispanic Asian
48% Hispanic

Comparing to UC Berkeley’s student data Non-Hispanic Whites are nearly perfectly represented (though they are underrepresented in freshman admits, making up for it in transfers). Hispanics are represented at 1/3 the proportion you’d expect, and Blacks at about 1/2. The unsurprising reality is individuals Asian ethnicity are overrepresented. The Berkeley website attempts to elide and obfuscate this obvious reality by disaggregating the Asian ethnicities. But it also allows one to compare to the Census data on these ethnicities.


CaliforniaUC BerkeleyRatio
South Asian1.48.15.8


A few notes. I assumed that Indian American is a good proxy for South Asian (in California it almost certainly is, as 90% or more South Asians probably are Indian). Second, I’m comparing against the whole population. Unlike Hispanics Asians are not that young-skewed (their fertility is lower than Non-Hispanic Whites), but because they are an immigrant community they are not as old-skewed as Non-Hispanic Whites. A quick scan shows that there is variation among Asian groups.

Finally, I decided to revisit the ethnic proportions at the various UC campuses. Taking all the undergraduates, and removing international students and those who refuse to provide racial information, here are the proportions on each campus:

InstitutionBlackNative AmericanAsian/Pac IsFilipinoLatinoWhite
Los Angeles513751836
San Diego215051528
Santa Barbara411732451
Santa Cruz312142249


The main thing that surprised me is that while Santa Barbara and Santa Cruz have reputations as schools for White kids, they actually have many Latinos as well. Though I have no idea if these are ‘White Hispanics,’ especially children of mixed-marriages who are putting down their more advantageous identity despite being functionally Non-Hispanic White in social contexts.


MORE ABOUT: Berkeley

Comments (52)

    Maybe the whole science degree *system* is culturally biased. Ever think of THAT, Razib? I haven’t seen a single Pygmy win a Nobel. What does that tell you?
    Seriously though, I wonder if they’re using “underrepresented minorities” to mean literally just that – only the minority groups that are underrepresented and so Asians don’t need to be included as they have been able to over come this bias via freestyle rap.

  2. Charles Nydorf

    Having to calculate extracurricular activities according to whether they help college admissions must be a drag.

    • Karl Zimmerman

      I wonder if any study has been done on how much extra-curricular activities are related to college performance, ultimate job placement, and adult income?

      My expectation is they would, but one you take out the class correlation (e.g., upper-middle class parents pressure smart kids to do these things now) there is no association at all.

  3. razibkhan

    but colloquially hispanic codes for non-white minority. even in cases where individuals could easily ‘pass’ as white because they are white. i have talked to spaniards who are convinced in the USA they would not be classified as white.

    (and fwiw, a much lower proportion of hispanics in california self-identify as white, as opposed to texas)

    • Yes, there’s definitely a lot of colloquial, lived-experience difference, but it is also true that White Hispanic has been a oft-used government classification for quite a while. And certainly a lot of regional variation too, although interesting about the California/Texas issue. But since you were using 2010 census categories, didn’t seem to be entirely a colloquial issue.

      • razibkhan

        i’m aware of all those details about the origins of hispanic. the point is that the public and the media rarely distingish hispanic’s ethnic vs. racial valence. the reason that g.z. was identified as ‘white hispanic’ is plausibly explained by the fact that the trayvon martin narrative gained heft if the antagonist was in some way connected to being white, because minority-on-minority violence 1) does not sell 2) does not fit the preferred narrative of whites vs. non-whites.

        • Perhaps. But as Chris Escalante has been tweeting, all the initial police reports seem to identify Zimmerman as “white.” Fox News was the first to say he was a “White Latino” and no one cared. But once the NY Times said “White Hispanic” everything broke loose, and the big “there’s no such thing as White Hispanic” meme took off.

          • razibkhan

            from a black perspective g.z. probably is ‘white’ (ie., a ‘cracka’). i’ve seen interviews of chicago activists where they refer to barack obama’s ‘white sister’ (she is of course half-indonesian). but from a white american perspective, he’s not, because white americans of all ideologies tend to be more punctilious in establishing boundaries of whiteness. basically this is complex stupidity on the part of the public and the media, easily hoked to different political/social narratives.

          • “White americans of all ideologies tend to be more punctilious in establishing boundaries of whiteness”
            Cheers, Razib, I’ll agree with you there! Always a pleasure! Happy Saturday…

          • Festivite

            Jason, if Wikipedia can be trusted, CNN got the ball rolling with “white Hispanic.” I would be surprised to learn Fox beat them to it. Have a link?

          • Hi, thanks, I almost always like Wikipedia. In this case it is a 3/20/2012 story on Fox News Latino, Feds To Investigate Killing of Black Teen by Florida Latino which uses “white Latino” in the first paragraph. It seems earlier than CNN, but like many things in this case, probably contested somewhere!

          • Antonio

            So if you are “hispanic” and kill a black it makes you whiter? By his appearance only, I doubt this guy does would ever pass as white in USA.

          • razibkhan

            there are several separate issues. for whatever reason on the order of half of american latinos identify as white. blanco. this is what shows up on the census forms. some of these individuals look white, and are clearly white in their self-identity and how people perceive them. for example, a cuban american who happens to be the grandchildren of german jews. a mexican american who lived in mexico until age 10, but moved to the USA, and has great-grandparents who are all eastern european jews (real example). OTOH, there are other latinos, perhaps the majority of ‘white’ latinos, who are visibly of partial non-european ancestry. within the latino community though they do not identify as black, or indigenous, and for whatever reason will not give ‘other’ as the response. the most reasonable course of action is then to offer up ‘white’ (i.e. spanish) as their racial identity. these people though are aware they are not white like a norwegian mexican is white, and are not perceived as such. in texas many tejanos identified as white because the legal framework of jim crow, which gave privileges to white people.

            in american society there is an understanding that even self-identified white hispanics will be identified by their hispanic identity without qualifier unless they are ethnically atypical (jewish latin americans fall into this class). hispanic identity has been racialized, and is implicitly non-white.

            the issue i’m alluding to is that in the g.z. case the tacit and near universal understanding that any hispanic with obvious likely non-white ancestry, as g.z. has, is not identified as white in the mainstream discourse, was broken. why? i think the most plausible rationale is simply that the narrative of a young black man being racial targeted by another ethnic minority has less power and valence than that where he was targeted by someone with some association with white privilege. ergo, making salient and highlighting the fact that g.z. is a ‘white hispanic’ (he is certainly as white as many of the 50% of hispanics who identify as such).

          • Antonio

            Thanks for your response!

          • Karl Zimmerman

            George Zimmerman (as his surname would suggest) is German-American from his father’s side (no relation to me, AFAIK). His mother’s side is Peruvian (ironically enough, he actually had one black great-grandmother).

            Regardless, half-Latino people are usually considered as white in the U.S. context, particularly if they have Anglo names. He’s almost certainly significantly over 50% West Eurasian by ancestry.

          • Antonio

            Hi, you’re taking about “half-latino” as if it is a race. I don’t follow it. I think the issue will depend a lot on your overall apparent ancestry. Suppose GZ take something from you in the street and run: would you report him as a white on the run? Cheers

          • Karl Zimmerman

            If I saw him on the street, I might presume he was Latino. Or from somewhere in the Middle East. If he introduced himself and his name was George Zimmerman, I wouldn’t think “oh, hey, he’s Latino” though. I’d think he was biracial something.

            Really, I think his name is what it’s all about. If his name was Jorge Carpintero, and he had identical looks, no one would have identified him as being white in the press. Having a very “Anglo” name, particularly surname, just confused the initial reports enough to muddy people’s responses.

          • razibkhan

            name does matter. many middle eastern people ‘look latino,’ but christian arabs from syria-lebanon often ‘pass’ (tony shaloub plays italians often).

          • Antonio

            Again, in you first paragraph you’re treating latino as a ‘race’. That’s wrong but ‘fine’: I think most americans do it anyway. I am only try to stressed that point. I agree with your second paragraph. It is all very interesting for a foreigner like myself: you are also identifying a very germanic name by anglo, which again would only happen in USA, I think. Cheers

          • razibkhan

            i hate the term ‘anglo.’ a lot of the time it obviously is used to mean white european, so just say it. my command of english is fine enough, and in fact my family has been educated in english at least back to my great-grandparents’ generation, so i’m OK with anglo, but i don’t think that that’s what people mean….

          • Antonio

            I am western european genetically and culturally but I am not fine with ‘anglo’ at all!

          • Karl Zimmerman

            I put “Anglo” in quotes for a reason, since Zimmerman is obviously a German, not an English name. Within the U.S. context, Anglo roughly means non-Hispanic white.

            As to the question of Hispanic as a race, you were the one who brought said question up to begin with when you said “by his appearance only, I doubt this guy does would ever pass as white in USA.”

            If you are saying GZ doesn’t “pass” for white, what do you think he passes as?

          • Antonio

            for me he “looks” “hispanic” or middle eastern.
            I don’t know what to think of his name vey well, because I am not from US. I kind of get that spanish names come to the US mostly from non-whites and therefore are associated with non-whites. I am also aware of the ignorance of the US mass public here regarding the rest of the world so that people don’t know – and don’t care – about the differences between, say, Iberia peninsula and Guatemala. But on the other hand I’ve see lots of people with ‘anglo’ names who are clearly brown or black. So I am a little confused with that. Btw, would italian names be “anglo” as well?

            I thought ‘white hispanic’ would be restricted to europeans types only that come from latin america. By now I am seeing that the issue is even more confusing.

            I guess in practice is not so confusing because the vast majorities of hispanics are actually quite different from the more typical USA whites. But I am digressing.

          • Karl Zimmerman

            In the U.S. context, I think “anglo” refers to Anglo-American culture, not an ethnicity. I could be wrong, but I would presume that among Mexicans (which is where the term originated I believe) blacks would be called anglo as well. The term has be reapropriated by white American culture, but in by experience it’s really only widely used on the west coast.

            I’d say that generally speaking Americans understand that Spaniards and Portuguese are white within the U.S. However, I think people have a false idea of what Iberians look like, as they use “white” Latinos (like many Cubans) as a baseline, when they are already a bit admixed. Hence people tend to think Iberians are as dark as Italians or Greeks, when in truth they tend to be significantly lighter.

          • Antonio

            I am not sure who is lighter or darker: I’ve seen very light italians as well and also not so light germans and english people so that I don’t know. I brazil people think portuguese are darker as well but I think that is because a lot of people that claim to be portuguese are already highly mixed.

            Anyway the irony of the hispanic is that the term is mixing the colonizers white people with the very people they oppressed! And the ignorance here is not only the americans who don’t know the difference between europeans indians. Hispanics themselves are very confused by thinking, say, that european and christian names that they use are non-white while they reality are that these names are “white” much before the english or german *versions* of them.

            But again in practice none of it matter because most people fit the stereotypes in the US context. Cheers

          • razibkhan

            italians are lighter, on average.

            1) i know the allele frequencies for light pigmentation genes, and italy north of rome is much more like northern europe than any part of iberia

            2) wife lived in italy (north), and has visited spain and portugal. has told me her blondism more atypical in iberia. though the southern party of italy is probably similar to spain, that’s only ~1/3 of the country

    • “and fwiw, a much lower proportion of hispanics in california self-identify as white, as opposed to texas”

      Unless a higher proportion of California’s Hispanics come from heavily Indio regions such as Chiapas or Oaxaca, or Central America, which AFAIK isn’t the case, this is more likely a result of greater ethnic consciousness in California. It’s certainly not irrelevant that California is much more politically liberal than Texas.

    • Anthony_A

      I’m a mostly-white half-hispanic. (23andMe tells me that I’m about 6% American Indian; someone who did a more detailed breakdown says the American Indian component is closer to Maya than Pima. My mother immigrated from South America, and has ancestors from Mexico and South America.)

      White people in the U.S. all see me as white – I have black hair and brown eyes, but I’ve never tanned that darkly, and I speak English natively, with a General American accent. I’ve had people make racist comments about Mexicans or Hispanics to me. (Stupid ones, too! If Mexicans were lazy, they wouldn’t be stealing your jobs!)

      However, back when I worked outdoors more, and was more tanned, mestizo and indio Mexican construction laborers on jobsites would usually address me in Spanish first.

      • razibkhan

        i have a friend who is of spanish (spain) background. mother from the south, father from the north. he is obviously white. but he has the same issue with latino laborers, etc. but i think the reason isn’t that he looks non-white, but that his facial features look like a lot of mexicans, because they’re about ~50% european. that is, a lot of mexican who are mestizo still have spanish faces, so to speak. so they recognize that.

        • Antonio

          I think there are lots of variance in the iberian peninsula, more so than in the north of Europe perhaps. Or this is at least my impression from a few trips.

          Even though I am vastly iberian I am never addressed in spanish by no one. Even when I say that one of my first names is Antonio the still think I am not spanish or latino. They think I am from Italy. When I say nothing about my name I am automatically associated with northern Europe, often as German or Russian but sometimes further north. It is quite interesting. I think this is a function of (1) look white – in the USA sense -, (2) have a strong accent and (3) being somewhat big and perhaps too masculine 🙂 I was never confused as french in the USA either.

          Btw, my actual portuguese family that I have record of is actually quite nordic except perhaps they are shorter. Thus I never have associated iberians with non-white types.

          • razibkhan

            i think you are getting confused by he unfortunate american tendency to think in typological binaries. e.g. white vs. black. anglo vs. hispanic. as someone who is not white, black, hispanic, or anglo as it is implicitly understood (white european who is non-hispanic) i can empathize 🙂

          • Antonio

            I don’t see my confusion which may mean I am either very confused or not at all. Either case I would be interesting to know. Cheers

  4. Riordan

    In lieu of budgetary issues in the last 2-3 years, I wonder how much “holistic review”

  5. Would you elaborate on “transparent academic deficiencies”?

    • razibkhan

      incoming lower test scores, lower GPAs, lower fraction of realized graduates, especially in difficult fields (e.g., engineering, not political science). also outgoing lower test scores (GREs).

      • bstew34864

        You are kidding right? You really believe that university entrance requirements are fair? Surely you realize most test scores are due to students having access to good schools, tutors, and a complex system of other advantages most poor students don’t have.

        More importantly, university entry requirements are primarily based upon little research and usefulness. Why have stringent requirements to get into a school? When much research has suggested that most universities are over hyped as esteemed institutions of learning.

        • razibkhan

          You are kidding right?</i.


          You really believe that university entrance requirements are fair?


          Surely you realize most test scores are due to students having access to good schools, tutors, and a complex system of other advantages most poor students don’t have.

          your ideas are simplistic. it is look well known that higher income black students perform more poor than poorer white (and asian) stuents.

          white test takers in 2003 who came from families with incomes less than $10,000 had verbal and math scores of 478 and 480. black test takers from families with incomes $80-$100,000 have scores of 461 and 468. Above $100,000 black test takers finally overtake white test takers from families with incomes below $10,000. using wealth, and not income, does tend to control the gap.

          the latest social science is that socioeconomic affirmative action would probably benefit mostly poor whites. it is known that at many elite universities most ethnic minorities are actually not socially or economically representative of their ‘communities’ (e.g., at harvard in the early 2000s 1/4 of black americans where biracial, and on the order of 1/2 were of some west indian background).

          When much research has suggested that most universities are over hyped as esteemed institutions of learning.

          probably. i’m not discounting education reform. i’m just saying that if you score low on your math SAT/GRE i do not want you to be a civil engineer for my locality. others are free to take a different though, though i would appreciate if those individuals would live in a different town, because i am a bit more risk averge.

          • bstew34864

            Razib, I love your blog though I sometimes strongly disagree with some of your views and conclusions. So, I am not going to enter a verbal battle with you on this subject. But, let me make a couple of brief points…

            1. It is widely known that there are several possibilities for the lower test scores of blacks regardless of income. For instance, there is a distinct difference between “income” and “wealth”. Though blacks may have more income there wealth and access to opportunities can be very much constrained. Even in comparison to poor whites.

            2. There is no good research that I am aware of that shows any predictive capability between SAT scores and future job success. As a matter of fact there is a lot of debate about the validity of ost “intelligence” test in many areas of life to include college entrance.

            3, Finally, your comment about the civil engineer is important. But not for the reason you stated. I’m not advocating that people should not be trained for their respective specialties. However, what I am saying is that the traditional way engineering is accomplished today is done differently that 20 years ago. Today, most calculations are automated and much work is done through teams and through others in the form of intergrated groups. So, instead of trying to create the smartest civil engineer we should be focused on developing a group of diverse engineers who will develop more innovative ways to build stronger bridges. Read Scott Page’s book called The Difference. His book proves a theorem he calls the diversity trumps ability theorem. It will change the way you view intelligence.

          • razibkhan

            For instance, there is a distinct difference between “income” and “wealth”.

            i stipulated that in my comment above specifically to head off this response. why are you making this point?

            There is no good research that I am aware of that shows any predictive capability between SAT scores and future job success.

            i’m pretty sure that beyond a particular cognitive threshold the correlation between test scores and performance is weak. the key is that most people in most professions don’t get to a specific stage without a minimum level of academic competency, which test scores are a reasonable predictor for (or more precisely, as good or better than others we’ve got). e.g. there has been some research (though disputed by others) that medical doctors’ performance has no correlation to their grades i medical school. i think this is easily explainable by the fact that the filtering stage for doctors is moderately stringent for medical school.

            in regards to this discussion the key point is a much lower proportion of underrepresented minorities who intend to complete STEM degrees do so. i am aware that in technical fields a lot of the old fashioned computation is done by computers. but i’m using a bayesian mcmc package it is often useful to know the outlines of how a metropolis-hastings algorithm works at some point.

          • botti

            @ bstwew34864,

            You should read some of Steve Hsu’s slide show for the BGI Cognitive Genomics Project. Hsu provides some data suggesting there are cognitive thresholds for certain subjects, such that mastery is very unlikely below a certain SAT scores.


            In terms of group differences, perhaps see Pesta & Poznanski’s paper in Intelligence on the possible mediating role of elementary cognitive tasks (eg reaction time, processing time measures rather than pencil and paper tests).


            Also, note there is a large body of research in industrial/occupational psychology on the predictive value of psychometrics for selection. See Frank Schmidt & John Hunter’s work.

  6. Antonio

    “i have talked to spaniards who are convinced in the USA they would not be classified as white.” It is just a digression, but I tend to have a cynical view of these attitudes. These are the “white man burden” in Latin American and elsewhere in the world, way before other europeans, but now they want special fellowships? If I recall correctly, these are the people – or perhaps the portuguese – who actually come up with the modern white/black distinction to champion the slave trade. It is quit an irony that now they don’t want to be “white” but instead by mistaken by the very people the colonize and oppressed. But anyway, I think US is the only country that they can try to do something like that, so I would still blame some of it to the US local ethnic confusion. Cheers

    • hb2

      Yes, they are the only white people in the world who are not white in the USA, except other white europeans (and asians) who immigrate to south America, where they become latinos and hence minorities.

      • Antonio

        In the past, USA also have other whites that were not white: Italians, Irish, germans, etc. The differences is that all those groups struggle to be white while these spaniards don’t want to be.

        • Cimon Alexander

          That’s not entirely true. Italians, Irish, and Germans certainly didn’t drink out of the “colored” fountains during the era of segregation. They may have faced discrimination, but there was little doubt they were “white”.

          • Antonio

            disagree. they were not ethically white and therefore exclude from the main group. Not as much as the blacks but neither the white hispanics today are as excluded as the blacks in the past.

          • razibkhan

            you need to separate groups. e.g.,non-catholic germans assimilated rather well and quickly.


            catholic germans OTOH less so.

            a lot of the issues re: irish had to do with roman catholic revivalism. the racial issue became more salient as racial ideology and identity became more crystallized in 19th century america. by the time large numbers of italians showed up they were much more explicitly marginally white, because the ideology of white anglo-saxon nordic whiteness was more well developed by the late 1800s.

            i the assimilationist model does have some value. but people need to be more careful about not forgetting historical context. america was much more racist in 1880s than 1830s (despite slavery in 1830s!), for example.

          • Antonio

            Or perhaps because of the end of the slavery plus more immigrants then the old white groups felt more threated than ever before. I know some models in which for racism to become salient you need some level of heterogeneity in the society as otherwise there is no apparent threat.

            I think beer production and consumption among the germans was also an issue for some groups in US. The only groups I haven’t seen much documentation about discrimination against were the Nordics (Danes, etc).

            Thanks for your reply.

  7. John Belmont

    someone with your expertise in population genetics must realize that these categories represent something beyond individual qualities and performance

  8. Thank you for the link and pointing out the simultaneity for the earlier reporting. This all of course got replayed after the verdict.

    Interestingly, via Razib’s census link above, in 2010 there were about 3.2 million Hispanics in Florida who identified as “white alone” or 76% of the Florida Hispanic population checking in on the Census as “white Hispanic”

    • razibkhan

      many white cubans in florida have strong race consciousness. plenty of cuban friends have mentioned this. cuba was, and to a great extent still is, a caste society despite the admixture. the diaspora in south florida is not representative of the island’s racial demographics. secondarily, many white cubans are descendants of early 20th century migrants from spain, so their european identity may be very strong. gloria estefan’s maternal grandparents were both born in northwest spain, for example.

  9. hb2

    White hispanics can be just white, such as Charlie Sheen. I question your assertion that White is the advantageous identity. In college, it is without question a disadvantage, as white students are held to different standards. Also, white identity is often looked on with shame, with shame. Hispanics have pride. The white hispanics I know (including an Italian guy who was born in Argentina) would be offended on being called white.


Discover's Newsletter

Sign up to get the latest science news delivered weekly right to your inbox!

Gene Expression

This blog is about evolution, genetics, genomics and their interstices. Please beware that comments are aggressively moderated. Uncivil or churlish comments will likely get you banned immediately, so make any contribution count!

About Razib Khan

I have degrees in biology and biochemistry, a passion for genetics, history, and philosophy, and shrimp is my favorite food. In relation to nationality I'm a American Northwesterner, in politics I'm a reactionary, and as for religion I have none (I'm an atheist). If you want to know more, see the links at


See More


RSS Razib’s Pinboard

Edifying books

Collapse bottom bar