Us­ing genes and IQ in ad­mis­sions merely repli­cates class in­equal­i­ties, ar­gues Ken­neth Richard­son

THE (Times Higher Education) - - CONTENTS - Ken­neth Richard­son

Re­searchers have be­gun re­port­ing cor­re­la­tions be­tween dif­fer­ences in in­di­vid­u­als’ DNA and IQ scores, po­ten­tially open­ing the way to ge­netic test­ing of in­tel­li­gence and even suit­abil­ity for univer­sity. But IQ tests are, in re­al­ity, con­cep­tu­ally mud­dled mea­sures of learned knowl­edge, self-con­fi­dence and so­cial back­ground, says

The task of ad­mis­sions tu­tors is an un­en­vi­able one. On the one hand, they are de­rided for re­ly­ing on ap­pli­cants’ per­for­mance in school ex­am­i­na­tions, lend­ing a huge ad­van­tage to those whose par­ents can af­ford to send them to se­lec­tive schools and to top up that ad­van­tage with end­less sup­plies of pri­vate tu­ition. On the other hand, ad­mis­sions tu­tors are cas­ti­gated as so­cial en­gi­neers when they at­tempt to level the play­ing field a lit­tle by low­er­ing en­try stan­dards for those from less priv­i­leged back­grounds.

The lure of a test that could evade this so­cio-po­lit­i­cal mine­field and iden­tify raw in­tel­li­gence is clear. IQ test­ing is the ob­vi­ous so­lu­tion. But while the con­cept has been around for a cen­tury, it has never man­aged to con­vinc­ingly de­liver.

Re­cently, some psy­chol­o­gists have be­gun claim­ing to have found dif­fer­ences in DNA that cor­re­late with IQ test scores and ed­u­ca­tional at­tain­ments. There­fore, they say, it will soon be pos­si­ble to pre­dict an in­di­vid­ual’s adult in­tel­li­gence from a mouth swab or drop of blood taken at birth. The chief ad­vo­cate of the idea is Robert Plomin, the well-known pro­fes­sor of psy­chol­ogy at King’s Col­lege Lon­don. In a pa­per pub­lished in Na­ture Re­views Ge­net­ics ear­lier this year, called “The new ge­net­ics of in­tel­li­gence”, he ar­gues that par­ents will even­tu­ally use DNA tests to pre­dict their chil­dren’s men­tal abil­i­ties and plan their ed­u­ca­tion ac­cord­ingly.

But the idea that uni­ver­si­ties could aug­ment ad­mis­sions tu­tors with job­bing phar­ma­cists is based on many un­likely, dan­ger­ously mislead­ing as­sump­tions.

One prob­lem is that psy­chol­o­gists have no agreed the­ory of in­tel­li­gence. That’s why IQ tests have al­ways been only pre­tend tests of in­tel­li­gence. There is wide dis­agree­ment about what they re­ally mea­sure. They are not like blood tests, liver func­tion tests or breathal­yser tests, how­ever much IQ testers want peo­ple to be­lieve that.

The idea that IQ tests could re­place school at­tain­ment in univer­sity ad­mis­sions is sug­gested by testers’ ar­gu­ment that IQ scores must be valid tests of in­tel­li­gence pre­cisely be­cause they cor­re­late with school at­tain­ment. In­deed, what the ad­vo­cates don’t men­tion is that the tests are rigged to do just that. Lit­tle puz­zles and tests that make up the scales (called items) are de­vised and tried in ad­vance. Only those whose re­sults do cor­re­late with school at­tain­ment tend to be se­lected for in­clu­sion in the test.

Testers see, in score dif­fer­ences, the ex­pres­sion of a mys­te­ri­ous men­tal strength or en­ergy they call “g” (Bri­tish psy­chol­o­gist Charles Spear­man’s ab­bre­vi­a­tion of what he called the “gen­eral fac­tor” of in­tel­li­gence). But the test items them­selves are par­o­dies of the com­plex­ity of nearly ev­ery­one’s cog­ni­tive func­tions in the real world.

The logic as­sumes, in any case, that school per­for­mance is it­self a test of in­tel­li­gence. But there are many doubts about that. Although school per­for­mance is used to se­lect for oc­cu­pa­tional level, nei­ther IQ scores nor ed­u­ca­tional at­tain­ments have much as­so­ci­a­tion with men­tal per­for­mance in the real world. More­over, ap­peal­ing to IQ test re­sults’ cor­re­la­tion with school at­tain­ment does noth­ing to sug­gest what value might be added by us­ing them to se­lect stu­dents.

Iron­i­cally, it is be­cause of the poor pre­dictabil­ity of univer­sity per­for­mance from school at­tain­ment that uni­ver­si­ties and col­leges have looked to IQ-type tests for help. In the US, for ex­am­ple, SAT tests – which test lit­er­acy, nu­mer­acy and writ­ing skills – are widely used for univer­sity ad­mis­sions. But a study by the UK’s Na­tional Foun­da­tion for Ed­u­ca­tional Re­search re­ported in 2010 that the adop­tion of an SAT-style rea­son­ing test did “not add any ad­di­tional in­for­ma­tion, over

and above that of GCSEs and A lev­els”. Like­wise, a 2015 re­view of re­search in the US and the UK pub­lished in the jour­nal Psy­cho­log­i­cal Bul­letin found that school test scores were as­so­ci­ated with less than 10 per cent of the vari­a­tion in univer­sity per­for­mance. “Self-ef­fi­cacy” be­liefs, or con­fi­dence in per­sonal abil­ity, were found to be much more im­por­tant.

This is hardly sur­pris­ing, as exam re­sults seem to rely far more on swot­ting and re­gur­gi­ta­tion than true abil­ity. As Barn­aby Lenon, chair of the UK’s in­flu­en­tial In­de­pen­dent Schools Coun­cil, said re­cently: “The best GCSE and A-level re­sults don’t go to the clever­est stu­dents – they go to those who re­vised in the Easter hol­i­days.”

With up to 18 ap­pli­ca­tions per place at some UK med­i­cal schools, those as­pir­ing to be­come doc­tors are among the most se­lected stu­dents. Here, too, there has been con­cern about the pre­dictabil­ity prob­lem. For ex­am­ple, a Lon­don-based team did a very care­ful anal­y­sis of cor­re­la­tions be­tween A-level scores and per­for­mance at UK med­i­cal schools across a num­ber of co­horts, from the 1980s to about 2012. Even with tricky sta­tis­ti­cal cor­rec­tions, the cor­re­la­tions, pub­lished in BMC Medicine in 2013, turned out to be only low to mod­er­ate (around 0.35, with as­so­ci­a­tions with prac­ti­cal ex­ams tend­ing to be rather lower). Sur­pris­ingly, GCSE re­sults also turned out to be only mod­er­ate pre­dic­tors of A-level per­for­mances two years later.

More im­por­tantly, IQ tests were tried with some in­takes. But scores added no pre­dic­tive value at all. This gen­eral pic­ture has been con­firmed in a num­ber of other stud­ies since, most re­cently in a UK study, “Does the UKCAT pre­dict per­for­mance on exit from med­i­cal school? A na­tional co­hort study”, pub­lished in BMJ Open in 2016.

And a study pub­lished just last month in Sci­en­tific Re­ports, “The ge­net­ics of univer­sity suc­cess”, re­ports that A-level scores are as­so­ci­ated with only 4.4 per cent of in­di­vid­ual dif­fer­ences in fi­nal de­gree grades (cor­re­la­tion 0.22). More­over, se­quenced DNA vari­a­tions (called poly­genic scores) were as­so­ci­ated with only 0.7 per cent of such dif­fer­ences. The much-sought-af­ter g- fac­tor seems to evap­o­rate out­side the self-ful­fill­ing cor­re­la­tion be­tween IQ and school per­for­mance.

Of course, all of this says noth­ing about IQ tests’ abil­ity to pre­dict how peo­ple will fare in life be­yond univer­sity. But re­search over many decades has also found lit­tle re­la­tion be­tween IQ and job per­for­mance. Here, there have also been in­ten­sive ef­forts to boost the very low cor­re­la­tions by “cor­rect­ing” them sta­tis­ti­cally. But these ef­forts have been highly con­tro­ver­sial, in­volv­ing the pool­ing of re­sults from hun­dreds of dis­parate stud­ies (some from the 1920s), us­ing es­ti­mates of miss­ing data and adopt­ing a host of du­bi­ous as­sump­tions.

Like­wise, there is lit­tle ev­i­dence that mem­bers of high-IQ so­ci­eties like Mensa are dis­pro­por­tion­ately suc­cess­ful in their ca­reers. In any case, such cor­re­la­tions are eas­ily ex­plained by fac­tors such as cul­tural back­ground, and the above-men­tioned con­fi­dence and self-ef­fi­cacy be­liefs.

Sim­i­larly, sur­veys go­ing back to the 1960s have shown that nei­ther school nor univer­sity grades are good pre­dic­tors of job per­for­mance. A re­view by J. Scott Arm­strong in the En­cy­clo­pe­dia of the Sciences of Learn­ing in 2012 put the cor­re­la­tions at nearly zero six or more years af­ter grad­u­a­tion. Higher-per­form­ing pupils do not tend to be­come “high-per­form­ing” adults. Con­versely, the vast ma­jor­ity of high achiev­ers in the real world, as adults, did not stand out at school.

Em­ploy­ers are in­creas­ingly catch­ing on. In a New York Times in­ter­view in June 2013, Las­zlo Bock, who was then a vice-pres­i­dent at Google, said that “we’ve seen from all our data crunch­ing that [ed­u­ca­tional at­tain­ments] are worth­less as cri­te­ria for hir­ing, and test scores are worth­less…They don’t pre­dict any­thing”. Google is by no means the only com­pany that has re­cently said that it will dis­re­gard ed­u­ca­tional at­tain­ment – in­clud­ing at univer­sity – in its hir­ing.

If, in re­al­ity, IQ tests are re­ally just tests of cer­tain kinds of learned knowl­edge, along with self-con­fi­dence, you could also de­pict them as mea­sures of so­cial class back­ground.

Not sur­pris­ingly, mem­bers of dif­fer­ent classes, with dif­fer­ent so­cial and oc­cu­pa­tional roles, de­velop dif­fer­ent kinds of knowl­edge and cog­ni­tive pro­cesses. These be­come part of their so­cio-cul­tural “ecol­ogy”; the re­sult­ing dif­fer­ences in neu­ral net­works in the brain even show up on MRI scans. At one end of the class struc­ture, fam­ily wealth pro­motes health­ier life­styles, phys­i­cal growth and cog­ni­tive vi­tal­ity in chil­dren. Eco­nomic se­cu­rity, stable cir­cum­stances and pre­dictable fu­tures fos­ter strong be­liefs in per­sonal abil­i­ties, and boom­ing self-con­fi­dence. At the other end, work­ing­class par­ents suf­fer the grind­ing stress of money short­ages and in­se­cu­ri­ties of em­ploy­ment, in­come and hous­ing. A re­port in Sci­ence in Au­gust 2013, “The poor’s poor men­tal power”, notes how poverty alone de­presses cog­ni­tive func­tion­ing and drains men­tal re­serves in par­ents and chil­dren.

Phys­i­o­log­i­cal pro­cesses have been dis­cov­ered through which early life stress – or even that ex­pe­ri­enced by the mother be­fore birth – can sti­fle long-term cop­ing in chal­leng­ing sit­u­a­tions in later life.

Work­ing-class par­ents are also likely to have reached neg­a­tive con­clu­sions about their own abil­i­ties through their own school ex­pe­ri­ence (which may be re­in­forced by ar­ti­cles about genes and in­tel­li­gence). It’s dif­fi­cult to feel as­pi­ra­tional for self or chil­dren within a so­ci­ety that has cer­ti­fied you as de­fi­cient in brain­power. Yet a re­port in the Jour­nal of Eco­nomic Statis­tics in 2010, “Must try harder: eval­u­at­ing the role of ef­fort in ed­u­ca­tional at­tain­ment”, cited just such as­pi­ra­tions as the “key to a child’s ed­u­ca­tional per­for­mance”.

There are even deeper con­se­quences. Mere per­cep­tion of in­fe­rior place in a so­cial or­der re­duces self-con­fi­dence and in­creases anx­i­ety in test sit­u­a­tions. Test anx­i­ety clouds the mind, dis­turb­ing at­ten­tion and fo­cus. Just be­ing told that it is a “test”, in­stead of, for ex­am­ple, a sur­vey for re­search, se­ri­ously de­presses the cog­ni­tive per­for­mance of mi­nor­ity and work­ing-class groups.

As Cardiff Univer­sity emer­i­tus pro­fes­sor Antony Manstead ex­plains in an ar­ti­cle, “The

Un­for­tu­nately, in the ab­sence of a gen­uine the­ory of in­tel­li­gence, mys­ti­cism pre­vails

psy­chol­ogy of so­cial class: how so­cioe­co­nomic sta­tus im­pacts thought, feel­ings, and be­hav­iour”, in a re­cent edi­tion of the Bri­tish Jour­nal of So­cial Psy­chol­ogy, “so­cial class dif­fer­ences in iden­tity, cog­ni­tion, feel­ings, and be­hav­iour make it less likely that work­ing-class in­di­vid­u­als can ben­e­fit from ed­u­ca­tional and oc­cu­pa­tional op­por­tu­ni­ties to im­prove their ma­te­rial cir­cum­stances”. Yet test scores are still treated as if they are read­outs from blood tests.

What IQ tests re­ally test is most clearly given away by the so-called Flynn ef­fect: the steep rise in av­er­age IQ scores over gen­er­a­tions in all de­vel­oped coun­tries. For ex­am­ple, av­er­age per­for­mances on a pop­u­lar test in the UK im­proved by 27.5 points be­tween 1947 and 2002 (the max­i­mum score is 60).

Such changes hardly re­flect the kind of fixed prop­erty of in­di­vid­ual in­tel­li­gence that testers ask us to be­lieve in. Psy­chol­o­gists have been ut­terly, and per­haps com­i­cally, mys­ti­fied by it, in­volv­ing them­selves in ar­cane de­bates about “bi­o­log­i­cal” or “en­vi­ron­men­tal” causes. They have been equally mys­ti­fied by more re­cent re­ports of a lev­el­ling-off, or even de­cline, in av­er­age IQs over the past 20 years.

The Flynn ef­fect cor­re­sponds al­most ex­actly with the ex­pan­sion of mid­dle-class jobs from the 1940s to the 1990s, re­sult­ing in the ef­fects on learn­ing and test-tak­ing con­fi­dence men­tioned above. Cor­re­spond­ingly, as so­cial mo­bil­ity has stalled over the past 20 years, so the ef­fect has tailed off.

The blind­ingly ob­vi­ous re­la­tion­ship be­tween IQ scores and class back­ground is also re­vealed by the fact that when chil­dren from de­prived back­grounds are fos­tered by mid­dle-class fam­i­lies, their IQs in­crease by up to 15 points.

Un­for­tu­nately, in the ab­sence of a gen­uine the­ory of in­tel­li­gence, mys­ti­cism pre­vails. It is prob­a­bly in­evitable that in­sti­tu­tions charged with grad­ing, sort­ing and plac­ing peo­ple in a class-struc­tured job mar­ket must re­sort to sim­ple metaphors such as “bright” or “dull”, “strong” or “weak”. But these are woe­fully self-ful­fill­ing ideas.

Politi­cians con­stantly prom­ise an ed­u­ca­tion sys­tem that “al­lows ev­ery­one to ful­fil their true po­ten­tial”. But this im­plies the ex­is­tence of a fixed, bi­o­log­i­cal lad­der of ap­ti­tude. It is what is called ge­netic de­ter­min­ism and it pre­vails even though ge­neti­cists warn that we must no longer think of the genome as a “blue­print”. The same ghost-in-the-ma­chine no­tion un­der­lies the con­cept of g. It does its job in re­pro­duc­ing, and seem­ing to le­git­imise, an il­lu­sory mer­i­toc­racy.

Ban­ish­ing such myths will, of course, have se­lec­tion im­pli­ca­tions for uni­ver­si­ties and col­leges. They are mak­ing com­mend­able ef­forts to widen ad­mis­sions cri­te­ria and pre­dictabil­ity, but a whole dif­fer­ent ap­pre­ci­a­tion of the na­ture and depth of the prob­lem is re­quired. The prob­lems are not those of a par­tic­u­lar so­cial class but of the class sys­tem as a whole.

Above all, we need to drop this tacit ge­netic de­ter­min­ism. Stan­ford Univer­sity psy­chol­o­gist Carol Dweck, in her best-sell­ing 2016 book Mind­set: The New Psy­chol­ogy of Suc­cess, has shown what hap­pens when we do. In her ex­per­i­ments, stu­dents and ed­u­ca­tors were en­cour­aged to re­place a “fixed mind­set” – the be­lief that peo­ple are ei­ther born smart or are not smart – with a “de­vel­op­ment mind­set” – with the im­pli­ca­tion that in­tel­li­gence and po­ten­tial are cre­ated from par­tic­i­pa­tion, not merely brought out or ful­filled. Apart from trans­form­ing the ed­u­ca­tional progress of stu­dents of all ages, there were leaps in self­con­fi­dence, an in­creased de­sire for chal­lenge and greater re­silience in the face of fail­ure.

At the Open Univer­sity, which has no for­mal en­try re­quire­ments, the re­mark­able trans­for­ma­tion of in­tel­lec­tual self-con­fi­dence among its stu­dents is well doc­u­mented. More widely, a sim­i­lar in­sti­tu­tional at­ti­tude change has helped to trans­form the gen­der im­bal­ance in sci­ence, tech­nol­ogy, en­gi­neer­ing and med­i­cal sub­jects.

All this should be enough to ques­tion the value of IQ test­ing, and to help ex­plain low as­so­ci­a­tions with later ed­u­ca­tional and re­al­world per­for­mance. But crit­ics also re­mind us of the dark, ide­o­log­i­cal side of IQ test­ing: its roots in the eu­gen­ics move­ment, its part in thou­sands of ster­il­i­sa­tions in the US in the 1930s, and the in­spi­ra­tion this pro­vided to the Nazis.

Ed­u­ca­tors should not be blinded by the ap­pli­ca­tion of amaz­ing tech­nol­ogy to crude ques­tions. As in the past, it is all based on moun­tains of sta­tis­ti­cal cor­re­la­tions: in this case be­tween mil­lions of tiny vari­a­tions in DNA and IQ scores. It is known that most such ge­netic vari­a­tions have no con­se­quences for de­vel­op­ment and func­tion. But waves of mi­grants from dif­fer­ent ge­netic back­grounds have en­tered the so­cial class struc­ture at dif­fer­ent lev­els. So the scope for spu­ri­ous as­so­ci­a­tions is enor­mous.

To­day’s would-be DNA-IQ testers de­clare more benev­o­lent aims than the IQ testers of the past did. But, through­out his­tory, sci­en­tists have of­ten be­come un­wit­ting bear­ers of ide­ol­ogy. And a pow­er­ful ploy has been to use cor­re­la­tions to turn ef­fects into seem­ing causes, to turn vic­tims into cul­prits. It is seen to­day in the del­uge of cor­re­la­tions be­ing de­scribed as “ge­netic ef­fects”, “genes for”, “ex­plain­ing”, “ac­count­ing for” and so on – when noth­ing of the kind is shown.

So the dire ef­fects of so­cial in­equal­ity – on health, de­vel­op­ment, age­ing, as well as psy­cho­log­i­cal and ed­u­ca­tional test­ing – have them­selves be­come por­trayed as the ef­fects of dif­fer­ences in IQ caused by dif­fer­ences in genes. In the process, IQ has be­come el­e­vated to tran­scen­den­tal sta­tus. To Univer­sity of Ed­in­burgh pro­fes­sor Ian Deary, it may mea­sure all round “bi­o­log­i­cal fit­ness”; to Plomin, it is “the om­nipo­tent vari­able” of hu­man ex­is­tence.

In re­al­ity, such ideas are only rhetor­i­cal re­descrip­tions of the class struc­ture of so­ci­ety, its priv­i­leges and de­pri­va­tions. We must protest about ef­forts to re­duce these so­cial causes to the ef­fects of in­ert se­quences of DNA.

Newspapers in English

Newspapers from UK

© PressReader. All rights reserved.