Minorities’ Cognitive Performance In The UK

Here is data from the Cognitive Abilities Test for UK students in 2009/10 via Ambiguous.

Some interesting things to take away here:

(1) The sample is very large. Verbal IQ has the highest correlation with academic performance in most subjects, followed by Quantitative IQ, and then Non-Verbal Reasoning (recognizing patterns and such, I imagine).

(2) Indians do almost as well as Whites, although the structure of their cognitive abilities are a bit different: About 4 points lower than Whites in Verbal, but almost 2 points better in Quantitative. As rec1man said, “The Patels and Sikhs are Upper-Shudra / Vaishya and this is 80% of the diaspora in UK.” So this is highly encouraging for India’s eventual prospects; in indicates that the broad middle can in principle build a reasonably wealthy, middle-class society.

(3) The Pakistanis and Bangladeshis also don’t do too badly – certainly better than I would have expected (I visited a Pakistani school once in the UK and it was horrific).

(4) This might imply we are actually looking at the following average-IQ groups in India: Lower 40% – 93; Middle 40% – 99; Top 20% – 105, for an overall average of 98 (once Flynn Effect is done with them). That’s better than Greece today and certainly good enough  to have a developed society. But there’s tons of challenges: Malnutrition, slums, poor education, widespread vegetarianism (both voluntary and involuntary – due to poverty) that have to be sorted out for India to perform to its potential.

(5) As with most IQ tests, the Chinese do as well as Whites in Verbal, but massively better in Quantitative and Non-Verbal Reasoning.

(6) Blacks do surprisingly well, lagging Whites by less than 0.5 S.D., which is VERY encouraging considering that according to US data where they are almost always 0.8-1 S.D. behind Whites. Two issues to consider (and bear in mind) here:

  • To what extent are Caribbean Blacks admixed with Whites?
  • As regards African Blacks, they simultaneously benefit from the Flynn Effect (much better fed than parents) but also suffer from regression to the mean (African immigrants to the US are the most credentialed immigrant group and thus have IQ’s well above the African norm, and I assume to a certain extent this is the case in the UK also; logically, their children who take the CAT will have have lower genetic IQ’s). Which of these forces is stronger?

(7) Another curiosity is that British Blacks do better on Quantitative than on Verbal. In the US it the other way round.

(8) As is typically the case, boys do slightly better on Quantitative and girls do slightly better on Verbal; and girls have lower S.D.’s (i.e. have fewer morons and geniuses).

(9) Incidentally, as a matter of curiosity, I note that in this – what I take to be a fairly representative sample of Britain’s school-age population – the proportion of British Whites is 82%, and the share of overall Whites is slightly less than 85%; mixed people are about 3%. For comparison, British Whites constituted 86% of the population (in 2001), while only 64% of children born in 2005 where recorded as British Whites. Seems like a very fast rate of population replacement.

The other stats were all pretty much as I expected, except for one very, very big surprise – average rural scores considerably surpassed urban ones (2 points in Quantitative, 4 points in Verbal). Usually, it is the other way round.

I suspect this is because urban areas have been flooded with lower-IQ groups. This is backed by the observation that the difference is smaller in the Quantitative component. As we saw above, both Asians and Blacks are relatively better at Quantitative tasks than they are at Verbal ones.

Comments

  1. I don’t know, man, we’re dealing with school exams here. Some groups simply take those things more seriously than others.

    • Some groups presumably take IQ tests more seriously than others. :)

      But we’re talking of the CAT here, not GCSE’s.

      • The test was carried out to test 7th graders(12/13 yo), and IQ/Cognitive abilities gap between groups get larger as people age (Gradual gap appearance). If the test were carried out for younger students, the gaps between groups will become narrower still.

      • Ambiguous says:

        The CAT scores for groups at age 11 correlate very highly with the percentage of that group gaining 5 good GCSEs including English and Maths. I got a correlation of 0.92 for all groups in the above study. Some minority groups do significantly better than predicted though, my not so tentative prediction for Pakistanis and Bangladeshis is that their cognitive ability continued to rise relative to whites throughout their secondary schooling so that at age 16 the gap is trivial to non-existent. Maybe that of Indians rises too I’m not sure. Chinese do as well as expected though.

        • Tests have become easier over the years to satisfy the political dogma of “everyone has the same potential”. If the US has been doing it by throwing in massive $$$ primarily , the UK’s main method have been dumbing down tests.

          The quality of the UK exams such as GCSE, A levels etc ( just look at the stunning % that got straight A’s last year) is highly questionable for the last decade, I suspect, which could partially explain some of abnormalities. Even though under such a dumbed down system one still can more or less identify the smart and the less smart, the gaps have become blurring (one step closer to “everyone is the same” nonetheless) and in some cases such as the ones btw the Whites and varies South Asian groups are just a bit bizarre, at least to me, even considering the destructive impact of yob culture. Therefore IQs derived from the UK tests (particularly in teh last decade) could sometimes be very misleading.

          • Ambiguous says:

            The correlation between CAT scores at age 11 and a group getting 5 good GCSEs including Maths and English is 0.92. Even a modest difference of 5 points in CAT makes a big difference in the proportion of that group getting 5 good GCSEs. Chinese and Irish Travellers perform just as you would expect in GCSEs. If they’ve been dumbed down, they’ve been dumbed down equally for everyone it seems.

            “and in some cases such as the ones btw the Whites and varies South Asian groups are just a bit bizarre”

            I don’t see what’s so bizarre about it. Do you live in the UK?

            “Therefore IQs derived from the UK tests (particularly in teh last decade) could sometimes be very misleading”

            Are you saying the CAT is misleading? Or predicting IQs from GCSE results is misleading? As I said the two correlate very highly indeed.

            • “If they’ve been dumbed down, they’ve been dumbed down equally for everyone it seems. ”

              e.g. For 2+13= 15 type test questions, the smartest guy in the class could make several typos and ended up as the 2nd place or 3rd, while a not too bright guy could get em all correct and top the list, even though generally speaking we more or less still can identify the big picture. This is much more unlikely or almost impossible the harder the tests are.

              Yes, on and off I do live in the UK during weekdays.

              • To sum up:

                The easier a test is, the more blurring/confusing the gaps are (as we are witnessing here), amongst the brightest and the lesser ones. When a test is eventually dumbed down to what is 1+1, there will be no gap at all among everyone – everyone is the same thus.

                Easy tests, by default, disproportionably discriminate the brightest sections.

              • Ambiguous says:

                So what? The correlation between CAT and GCSEs is still 0.92 regardless of how dumbed down GCSEs may have become.

  2. Is there any difference for the structure of difficulty of CAT vs other IQ tests? As you know, the slope of performance can be changed by the organization of test questions. And when all questions are too easy (all passed) or too hard (all failed), you will not see any difference between groups. If CAT only stratified above white, you would see the result like this. If CAT stratified like other IQ test, then data would be real.

  3. Those Q and NV Chinese scores are simply astonishing—112 and almost 110. However, the numbers are tiny, so maybe they tend to be from elite professional immigrant backgrounds, like businessmen or engineers.

    I’ll admit I’m also a little surprised that all the students with any non-white ancestry amount to only about 13% of the total. To hear all the WNs complain on blogsites, you’d think they’d be over 30% of the younger age cohorts.

    • I doubt. The Chinese in the UK are by and large the descendants of the Cantonese settled long ago.

      Unlike some other ethnic groups such as “Asians”(Indians, Pakistanis, etc) whose cog elites could immigrant to the UK quite easily in the last decade under NuLabour(e.g. outsourced NHS doctors, engineers, IT wokers and school/uni teachers/researchers, financial analysts), Chinese professionals even with Oxbridge degrees mostly can not immigrante to the UK due to very strict regulations. Wierd enough that for decades the UK has put very high bar even on simple tourist Visa for the (rich)Chinese from the mainland – Telegraph’s Ed West wrote a recent piece on that. And from what I’s aware of the UK is not amongst most desired immigration destinations for the young mainland Chinese professionals.

    • Ambiguous says:

      The Chinese mean score of 107.5 is pretty much exactly what they get in developed places like HK and SIngapore. Nothing surprising about it.

  4. namae nanka says:

    googling for CAT, it’s given for 7th or 9th standard pupils and GCSEs derive a significant portion of their scores from coursework.

  5. “Any Other Asian” (865) Category : Vienamese, Laos, Malaysians(Malays and ethnic Chinese)…?

    • Ambiguous says:

      The largest group of “Other Asian” are Sri Lankans, there are probably some Nepalis and Afghans among them too. In Britain “Asian” has always primarily meant South Asian.

      From wiki:

      “South Asians who marked “Other Asian” as an ethnic group and then wrote in their specific ethnic group were mostly (23%) of Sri Lankan origin. This was followed by fill-ins of Middle Eastern (9%) origin. Due to a growing sense of affiliation with Britain, many third generation South Asians chose to not mark “Asian or British Asian” and instead marked “British Asian” in the “Other Asian” write in section.”

  6. Ambiguous says:

    “To what extent are Caribbean Blacks admixed with Whites?”

    I don’t have any real data for you off the top of my head, but they look considerably less mixed than US Blacks. Watching the London Olympics the Black Caribbeans looked a lot blacker than the US Blacks.

    “(African immigrants to the US are the most credentialed immigrant group and thus have IQ’s well above the African norm, and I assume to a certain extent this is the case in the UK also; logically, their children who take the CAT will have have lower genetic IQ’s).”

    African immigrants to the UK are a heterogeneous bunch, some are educated like many Africans from Britain’s old colonies, while some of them are Somali refugees who are even more disadvantaged than Bangladeshis. On average they’re somewhat more likely to have degrees than natives do. In any case, the CAT scores of the children seem no different to Black Caribbeans. Btw even Somalis who came over as refugees only score about 0;5 SD below whites in measures of academic achievement.

    • Ambiguous says:

      “Indians do almost as well as Whites, although the structure of their cognitive abilities are a bit different: About 4 points lower than Whites in Verbal”

      That’s almost certainly down to some Indian’s lack of English skills, though it’s not as big as the gap for Pakistanis and Bangladeshis. Look at the scores for EAL versus non-EAL, there’s a substantial gap between verbal and non-verbal scores for the EAL group. 95% and 99% of Pakistanis and Banglas are EAL, though even that doesn’t tell the whole story about their English skills because there’s a lot of variation in English skills among EAL learners.

    • People in the Olympics aren’t 11 but much older. There are also a lot of athletes that have been re-flagged to British. For 30 year old it is something like a not so small minority but now for Caribbean babies probably a majority have a white parent.

      • Ambiguous says:

        Uh yeah, you did notice a separate ethnicity in the table for Mixed White and Black Caribbean didn’t you?

  7. [q]. Seems like a very fast rate of population replacement.[/q]

    Not necessarily. In the last decade and a half it has become much more easier for 25 year old Western Europeans to work and live in Britain. The same is true for British in Western Europe. I wouldn’t call that replacement but mixing. Also immigrants and British are much more likely to get children with each other which also depresses the number of “pure” British whites

  8. ironrailsironweights says:

    The abysmal scores for Irish Travellers have some bearing on the nature-vs-nuture debate, as unless I’m quite mistaken the Travellers are physically identical to other Irish people. One thing that comes to mind is that inbreeding may play a factor.

    • Inbreeding plus a possible founder effect? It was probably the left tail of the Irish population that originally formed the Traveller culture.

    • Ambiguous says:

      Well what you have to understand is, the low score of Irish Travellers is probably down to poor environment and sporadic schooling, while that of Roma is obviously predominantly due to genetics and little to do with poverty and sporadic schooling. ;)

  9. This is a bit bizarre…..South Asians seem to be performing at about the European mean in Britain yet the claim by Richard Lynn is that they have an IQ of 92. The CAT seems to indicate an IQ in the high 90’s at least. Judging by the overall results, the CAT seems to be of a representative sample. I don’t know who to believe anymore: Richard Lynn and his colleagues, or this result. On a side note, why haven’t Rushton, Lynn, etc, commented on these results? They seem to be of theory shattering proportions.

    • Ambiguous says:

      “On a side note, why haven’t Rushton, Lynn, etc, commented on these results?”

      Bhargava, don’t hold your breath. Another blogger approached Lynn et al for their comment on this, and he got nothing back. Also note the relatively high scores for all Black groups, the difference is just 0.5 SD below whites which probably reduces by age 16.

      • I can’t help but suspect that they deliberately ignore contrary evidence…..I mean, these results were done on a MASSIVE data set and are undeniably a representative sample. Perhaps the HBD forum needs to make a larger splash with this dataset to get their attention.

  10. The UK Indian diaspora is about 80% from mid-level castes, 10% from upper castes and 10% from lower castes

    The ceiling for the lower 40% is definitely no more than 90

    Lower 40% ceiling = 85
    Mid 40% ceiling = 95
    Upper 20% ceiling = 105

    Based on UK results, there is a clear 10 IQ gap between muslims and Hindus / Sikhs

    There are 3 clear levels of reservation quota / affirmative action in India, due to a persistent huge gap between each level

  11. Ambiguous says:

    Oh great, recman again. The non-verbal/quantitative gap between Indians and Banglas/Pakistanis is 4 points for Banglas and 6 points for Pakistanis. I’ve no idea how you got 10 points from that. Nor is there is much of a reason to assume that gap is all or mostly genetic in origin. You also have no data concerning “caste” in British Indians, because no such data exists.

  12. “So what? The correlation between CAT and GCSEs is still 0.92 regardless of how dumbed down GCSEs may have become.” (ambiguous)

    Oh yeah? What does it mean to you even if the correlation is as high as 1, if the former is about questions of what is 2+13 and the latter is about what is 5*2+(4/2)?

    It means that “due to the happyness caused by the dumbed down, piggys can and will fly twice, valiantly, even without steroids”. Big deal. LoL

    • Ambiguous says:

      “Oh yeah? What does it mean to you even if the correlation is as high as 1, if the former is about questions of what is 2+13 and the latter is about what is 5*2+(4/2)?”

      That’s a strawman of your own making, You do know the correlation between GCSE and g right? Your tiresome, cliche, rationalisations are unconvincing. There are large differences between some groups, not just ethnic groups but also SES groups. Funny how groups as different as the Chinese and Irish Travellers perform just how the CAT predicts.

      But yeah, go ahead and believe whatever you want to believe, you’ve obviously made your mind up already, regardless of any evidence. The silence from Lynn et al about these results is kind of deafening, they can’t claim ignorance about it either.

      • Strawman? No. It’s an Ironman! LOL.

        Not every GCSE has a decent cor with g; even if they do all the time, as long as it ‘s not 1 there’s room begging for explanation.

        What’s your explanation for some seemingly “contradicting” results from some diff tests around the world, eh?

        I offered no strawman but a possible, and very likely, Ironman explanation of that, namely a problem of non-standardisation of all those tests. In this case the gradual dumping down of the tests themselves, being GCSE or CAT or A levels, etc. Even though they all have high mutual correlations and each will still have some decent correlation with g after the dumping down ( of course, those tests are not what is 1+1, yet.), they prove next to nothing when dealing with theoretical gaps – surely Richard Lynn ignores it and rightly so.

        Indeed the negative impact of the UK “feel good” super easy tests has been backed up by a wealth of complains from both studs and educationists and employers out there for over a decade.

        As I’ve demonstrated, more often than not the result will be nothing but confusing with blurring gaps whenever a test becomes easier, and vice versus – extremely straight forward, is it not? it’s pretty much the very definition of “test”. Else what’s it for if it’s so easy that almost everyone can become a brain surgeon with perfect scores?

        • Ambiguous says:

          Correlation between GCSE and g is 0.81, between SAT and g it’s 0.82. Scores for the CAT and the WISC predict each other. If you think the link between CAT, g and GCSE is spurious then you need some alternative explanation.

          Of course GCSE doesn’t have a correlation of 1 with g, no academic test in the world has that. Even mental tests themselves only have an average g loading of about 0.6 with about 0.8 for Ravens. So the RPM has the same g loading as the SAT and GCSEs, wonder why that is? If you have an alternative explanation other than hand waving, I’d like to hear it.

          Btw the high correlation of g and GCSEs well explains the high correlation between CAT and GCSE.

          On a related note, Flynn has a book coming out soon that claims to show a slight advantage for women in g in some countries. If the advantage exists at age 16 in the UK, that could go a long way to explaining why girls do better at GCSE, because even a small IQ advantage makes a big difference, as the graph shows.

          • School test are not supposed to test IQ but scholastic excellence. To be good in school you don’t need to only have a high IQ but also pay attention and do your homework. That is why a correlation with g of 1 would be proof that the test was done wrong

        • Dumbing down is always the claimed reason why people score better but an increase in IQ, effort and home environment are big reasons why you should expect better result now than 20 years ago.

  13. @Ambiguous

    There’s an old saying that “assumptions are the mother of all screw-ups”. LOL.

    Your “facts” seem correct at the surface, but with due respect your underlying assumptions appear not exactly true:

    1. e.g. GCSE’s 0.81 cor with g is derived from the data of 2006 documented in “Intelligence and educational achievement”, Deary, I; Strand, S; Smith, P; Fernandes, C (2007), yes? You assume that all GCSE tests remain the same g correlation every year, year after year, which is likely not correct for the last decade, if the widespread criticism that gradually the tests are becoming earsier under NuLabour rule is true (which is). Ditto CAT and A level.

    2. You assume that 0.81 is be all end all to rank the exact group order of the corresponding g and to explain the some discrancies as if it were 1.0, which isn’t. 0.89 or 0.92, for instance, sounds much better to my ears.

    Surely more data (such as exact correlations of A level , GCSE and CAT to g for each year) are needed to challenge my explanation which by far remains intact. BTW, intra-relations among the three doesn’t mean anything to the topic at hand.

  14. Ambiguous says:

    “You assume that all GCSE tests remain the same g correlation every year, year after year”

    1) Bear in mind that CAT correlated with GCSEs taken in 2010 at 0.92. That’s pretty damn high.

    2) I totally agree that grade inflation has shifted the curve to right, but it hasn’t appeared to change the shape of it otherwise we wouldn’t see such a strong correlation with IQ. The curve is still very steep between IQs of 90-110.

    3) CAT is an IQ test that correlates as highly as one would expect with other IQ tests like WISC.

    Anyway we’ve gotten far away from the topic at hand which was the CAT data itself.

    • Ambiguous says:

      2) I meant to say the curve has shifted to the left. Note how the curve for getting 5 A*-As is significantly shifted to the right. Those with IQs of 100 or less are extremely unlikely to achieve this.

      • Ambiguous says:

        Also, I didn’t mean to say we should use GCSEs as a proxy to determine IQ at age 16, just as a useful indicator, a predictor of what we would expect to find.

    • Have you proof of grade inflation?

  15. Ambiguous: LOL.. so it’s an IQ test? Given the motivation, long term policies, and means in place to have dumped down academic tests, what is the fat chance you think that the masters have left loose the mother goose of all, namely IQ test?

    Under such a scenario, we’re bound to see something that is mind boggling to say the least, such as:

    A. Mixed White & African > Indian (Patels, Sikhs /Upper-Shudra / Vaishya) in both verbal non-verbal reasoning. Yeeks?!

    B. Mixed White & Caribbean > Indian (Patels, Sikhs /Upper-Shudra / Vaishya) in verbal reasoning.

    But what’s IQ average of Mixed White & African/Caribbean in N America again?

    C. Mixed White & Asian (Indian, Pakistanis, etc South Asians) IQ > White British IQ on all accounts… somebody HELP! Indeed Mixed White & Asian is the 2nd highest IQ group after the Chinese. o-0

    D. A group called “unclassified” got 2nd highest non verbal reasoning IQ of 109! Who are these mystically intelligent people (with sample # 10x that of Chinese thus can not be Koreans and Japanese.) who are not non of above in the table?

    E. Black Africans and Black Caribbean top Traveller Irish on all account by almost half SD. WoW!

    Let alone the implications of Indian as AK wrote, actually this CAT table blows most of HBD theories out of the water.

  16. namae nanka says:
  17. Ambiguous says:

    Hi SP.

    YES it’s an IQ test with a correlation with g of 0.83, your conspiracy theories notwithstanding.;)

    “A. Mixed White & African > Indian”. I already explained about the verbal/non-verbal gap in South Asians. Indians have a slightly higher non-verbal score (average of quantitative and non-verbal) there’s hardly any difference in the non-verbal scores of these two.

    “B. Mixed White & Caribbean > Indian (Patels, Sikhs /Upper-Shudra / Vaishya) in verbal reasoning.”

    See above. And please don’t mention any of recman’s crap to me!

    “C. Mixed White & Asian (Indian, Pakistanis, etc South Asians) IQ > White British IQ on all accounts… somebody HELP!”

    It’s the more educated Asians (and whites) who are more likely to intermarry.

    “D. A group called “unclassified” got 2nd highest non verbal reasoning IQ of 109! Who are these mystically intelligent people”

    I’m not seeing this, the Unclassifieds got 99.2 and 100.5 in quantitative and non-verbal.

    “E. Black Africans and Black Caribbean top Traveller Irish on all account by almost half SD. WoW!”

    What’s so surprising about that? Irish Travellers do very badly academically.

  18. Doesnt the gap between different ethnicities become greater with age? Am I missing something here?

    • Amibiguous says:

      Not in the UK, for example Pakistanis and Bangladeshis and even Indians start out below whites at age 3, then they close the gap (or reverse it) as they get older. Pakistanis and Banglas start out WAY below everyone else.

      To say the gap between ethnicities increases as they get older kind of misses of the point. With increased age individuals and groups approach their genetic potential, (thought it still doesn’t erase the effects of environment).

  19. What I notice is mixed white & African is 100.8 (verbal), 99.6 (numerical), 99.6 (non-verbal). If the difference between whites and Africans in the sample is genetic, the mixed white/African scores should be somewhere in between. They really should be lower if the genetic hypothesis were true. But they are practically equal to the white scores, despite presumably a significant amount of African genes. I think this is definite evidence against the genetic hypothesis.

    I wonder what Chuck would say about this.

    I also notice a significant difference between Indians and Pakistanis, despite them being practically the same genetically. Indians more selected or is it Indian culture?

    I just really wish we had IQ data for all ages in Britain!

    • It’s on Chuck’s site that I found this.

      “I also notice a significant difference between Indians and Pakistanis, despite them being practically the same genetically. Indians more selected or is it Indian culture?”

      Hmm I dunno, could it be… SES factors? ;)

  20. prefers2banon says:

    The percentage of Black and white pupils achieving five or more A* to C grades in GCSE and equivalent by gender 2006 (sources DfES,2005,2006a)

    Caribbean boys 36%
    African boys 43%
    White boys 53%

    Caribbean girls 52%
    African girls 56%
    White girls 62%

    Note large differences between Caribbean blacks (not selected for intelligence); African blacks (around half from parents who selectively immigrated with much higher iq than typical Africans) and white kids.

    Note difference between boys and girls is roughly the same as whites and blacks. This shows how performance is loaded to behaviour rather than g.

    I don’t see much evidence for black and white convergence in uk

    • Amibiguous says:

      If selection played such a large role you’d expect to see Africans having higher IQs than Caribbeans, they don’t. You’d also expect to see Chinese having higher IQs than they do in other developed countries instead they score just the same.

  21. prefers2banon says:

    If we up the age and difficulty of exams differences get more marked

    The A level is taken at around 18. Around one third of students take A levels: on average 3. The highest grade is A – around 25% of students achieve this. Fewer achieve 3 As or more.
    8% of whites achieve at least 3 As but only 1% of blacks

    11% of white takers get a first class degree but only 3% of Blacks.

    Chinese and Indian students perform above British White at GCSE but below at first class degree (9% and 7% respectively).

    It is not clear why this should be, although one explanation is innate difference

    • Amibiguous says:

      For God’s sake, if you’d looked into this in any detail at all you’d know white A level candidates are a select group, AND you need to take into account that ethnic groups don’t have the same participation rates for different subjects. Whites are more likely to do easier A levels while minorities are more likely to do harder ones.

      At university it’s a similar difference. Only 38% of whites go compared to 60% of minorities, and they’re more likely to do harder courses such as STEM.

      So actually it’s very clear why this should be.

      • prefers2banon says:

        not true. Chinese and Indians (who do better than whites) are more likely to do STEM (because they want to do medicine, law and engineering). Blacks who do worse than whites tend to do easy A levels like sports and media. So higher iq is associated with harder A levels and better results duh.

        • prefers2banon says:

          Its true that 38% of whites are in higher education and 61% of blacks. Correcting for participation in education: 4% of the white population get a first class degree and 2% of the black. Indian and Chinese — pretty similar to white

        • prefers2banon says:

          for the stats: 3% of whites have a science degree, 1.5% of black Caribbeans, 7.8% of Chinese.

          It’s rather disingenuous of you to compare whites with minorities: Chinese tend to perform better than whites at high levels, Indian indians (selective immigration) better or the same, Africans (mostly selective immigration) somewhat worse, Pakistanis and Bangladeshis a lot worse, and Caribbeans worst of all.

          Black African, Indian and Chinese students are over-represented in science, whereas Black Caribbean (especially men) and Bangladeshi (especially women) students are under-represented. (google royal society report)

  22. prefers2banon says:

    And some more:
    this time the total score achieved in A levels. The top rank is >360 points. This is achieved by:
    Chinese 32%
    White 28%
    Indian 20%
    Pakistan 11%
    Black 9%

    http://www.hefce.ac.uk/pubs/year/2010/201013/

    The data for the first class degree (summa laude) comes from https://www.education.gov.uk/publications/eOrderingDownload/RW92.pdf

    • Amibiguous says:

      Figure 4 in your second link actually controls for some of the factors I mentioned like different entry qualifications and subject choice.

      • prefers2banon says:

        yes one of the factors they control for is “A level route to university” versus “dodgy-non-g-loaded” entry route. Sure once they control for that the difference goes down. They are controlling for intelligence.

  23. prefers2banon says:

    more from uk government data

    Whites 38% get an A level of some kind, and 19% get a STEM A Level, and 4% get 3 or more STEM A levels
    Caribbeans 28% get an A level of some kind, and 9% get a STEM A level, and <1% get 3 or more STEM A levels

    As we up the level of difficulty from: "any A level < STEM A level < several STEM" the ratio of white over-performance relative to Caribbeans goes from 1.3, to 2.1, to 4.4.

    There is no doubt that British Whites perform better at science than Black Caribbeans in Uk or that Chinese perform better than British Whites. The harder the exams are, the greater the difference in black versus white performance. What we don't know or the reasons.

    A smaller proportion of the Caribbean population is studying science than the proportion of the white population studying science. Yet 13.4 of Whites get a first class degree in STEM compared to 2.2 of Caribbeans. Case closed.