DOCUMENT RESUME - ERIC · 2014. 1. 2. · DOCUMENT RESUME ED 084 318 UD 013 879 AUTHOR TITLE. SPONS...

DOCUMENT RESUME

ED 084 318 UD 013 879

AUTHORTITLE

SPONS AGENCYPUB DATENOTE

EDRS PRICEDESCRIPTORS

Kamin, Leon; Green, WinifredTranscripts From the Southern Regional Council'sSymposium on Human Intelligence, Social Science, andSocial Policy. Text of Di. Kamin's PresentationDenying That Proof Exists That IQ Test Score areHereditary; How Testing Harms Children; IQ ,sts asInstruments of Oppression--From Immigratioi ,uotas toWelfare.Southern Regional Council, Atlanta, Ga.7316p.; A reprint from "South Today"

MF-$0.65 HC-$3.29*Ability Grouping; Educational Diagnosis;Environmental Influences; Genetics; *Heredity;*Intelligence Differences; *Intelligence Tests;Racial Differences; *Racial Discrimination; ResearchMethodology; Research Problems; Social Differences;Social Discrimination; Social Sciences; StudentGrouping

ABSTRACTThis document comprises three presentations made on

March 23 at a symposium sponsored by the Southern Regional Councilfocusing on Human Intelligence, Social Science and Social Policy. Thefirst of the three parts of the document is the toxt of the principalpresentation, made by Dr. Leon Kamin, Chariman of the Department ofPsychology at Princeton University. The presentation first summarizesfindings of his extensive research into the original studies on whichsome American social scientists have based writings which at leastquestion whether environment has any effect on IQ testscores--suggesting that heredity may be the determinant. The secondpart of the document is the text of a presentation by Ms. WinifredGreen, director of the Atlanta-based Southeastern Public EducationProgram of the American Friends Service Committee. The presentationargues that qr011DIng reinforces the effects of years ofdiscriminatory treatment in the education of black children--lockingthem into classroom situations where curriculum, materials, teacherexpectation and the resulting stigmas and hopelessness are the Fameas, or scale predict worse than, the days of separate but unequalschools. The third part is the text of Dr. Kamin's afternoonpresentation on the hi.story of the use in the United States tosupport repressive public policy. (Author/JM)

Transcripts FromThe Southern Regional Council's

SymposiumOn Human Intelligence,

Social Science,And Social Policy

U.S. DEPARTMENT OF HEALTH,EDUCATION A WELFARE

NATIONAL INSTITUTE OFEDUCATION

THIS DOCUMENT HAS BEEN REPROOUCED EXACTLY AS RECEIVED FROMTHE PERSON OR ORGANIZATION ORIGINAT MG IT POINTS OF VIEW OR OPINIONSSTATED DO NOT NECESSARILY REPRESENT OFFICIAL NATIONAL INSTITUTE OFEDUCATION POSITION OR POLICY

A Reprint From

TODA.A Digest of SoutherreAffairs

Price: Twenty-five Cents

May-June, and July, 1973 Issues Vol. 4, No. 8 and 9

FILMED FROM BEST AVAILABLE COPY

Transcripts Fromle Southern Regional Council's

SymposiumOn Human Intelligence,

Social Science,IN.

J F ROMO R G N

And Social PolicyREPRO

N IONSREPRE

VOTE OF

A Reprint From

tues

TORAYA Digest of Southern Affairs

Price: Twenty-five Cents

Vol. 4, No. 8 and 9 Atlanta, Georgia

FILMED FROM BEST AVAILABLE COPY

Text of Dr. Kamin's PresentationDenying That Proof Exists

That 10 Test Scores Are HereditaMy current impressionthough 1 re-

main open to persuasion by thedatais that, upon detailed analysis,there is no evidence sufficiently strong toconvince a reasonably prudent person toabandon the null hypothesis that there isno hereditary determination whatsoeverof intelligence test scores. In arriving atthat conclusion, I think it is important foryou to know that it is not based on sec-ondary sources or review articles, eventhough they appear in such distinguishedjournals as the Atlantic Monthly. If onegoes back and looks at the actual experi-mental data which psychologists haveaccumulated on heritability of human in-telligence, the evidence seems to crumbleunder one's probing finger so to speak.

Weight of Twins

Let me begin with an example of aparticular study. Not because it is aterribly important one, but because itillustrates, to my mind at least, the wayin which imperfect data are amplified inpublic journals which are currently madeavailable to policy makers. If one readsProf. Richard Herrnstein's celebratedarticle in the Atlantic Monthly, what hedoes there is first to review the evidencewhich claims to show that about 80 per-cent of the variation in intelligence isattributable to heredity. That leaves about20 percent of the variance still to beaccounted for.

And how do we account for that 20

On March 23, the Southern RegionalCouncil sponsored a symposium with far-reaching implications on Human Intelli-gence, Social Science and Social Policy.The principal presentation was made byDr. Leon Kantin, Chairman of the De-partment of Psychology at Princeton Uni-versity.

Dr. Kamin first presented findings ofhis extensive research into the originalstudies on which some American socialscientists have based writings which atleast question whether environment hasany effect on IQ test scoressuggestingthat heredity may he the determinant.Dr. Kamin presented his findings in aquiet, almost off-hand manner. But whatIre said was startling and highly signifi-

cantin effect that there is no valid evi-dence at all to support the heredity as-sumption.

Since controversy over the heredity-environment question is one of the mostimportant intellectual debates of ourtimes, fraught with social policy implica-tions and influence, we consider the fol-lowing transcript of his presentationamong the most important materialsSOUTH TODAY has ever published.(Dr. Kant:n spoke from notes; the tran-script was tape-recorded. The transcriptof his second presentation. on the historyof the use of testing in America to sup-port repressive public policy, will appearin a subsequent issue of SOUTHTODAY.)

percept which is not genetic? Well, saidProf. Herrnstein and I quote him: "Theusual assumption that education and cul-ture are critical is running into evidencethat the physical environment, forexample, early diet, might be more im-portant. In fact, the twin studies that[Prof. Arthur] Jenson [educational psy-chologist at the University of Californiaat Berkeley] surveyed showed that thesingle most important influence on IQwas not education or social environment,but something prenatal, as shown by thefact that the twin heavier at birth usuallygrew up with a higher IQ."

That's a rather astonishing finding.And it has clear policy implications. What

the statement says is that insofar as thereis an environmental effect at all on humanintelligence, it's only the intrauterine en-vironment during pregnancy which hasany major effect. Therefore, perhaps weought to feed welfare mothers with aspecial protein supplement during preg-nancy. That might help the IQ's of theirchildren, but there's no sense in wastingmoney on the cultural, social, or educa-tional environment of these children. Bythe way, there's also no sense in feedingall of their fathers, if we are going to takethis seriously, at least not to improve thekids' IQ's.

But what's the basis for this statement?Well, what Prof. Herrnstein has done in

his articleearlier aJensen, wlowing: "genetic vaand otherthan diffelogical enfavorablement areweight betin birth wIQ differe

The sadoes Profcites a sinhe presenScarr, whyof a numsented infitwins we,that theinvariablyThe basicmust havein the woblood sup

The fiJensen'stake'. frobreak twiFor somethat eachFor othermore th.weighed I

xt of Dr. Kamin's PresentationDenying That Proof Exists

t 10 Test Scores Are Hereditaryis

1S.

toIS

Crat

c-cn

:dIlcri-veri-de

a

ait

ayinde;dscc!he,ce

IS

uthe

20

On March 23, the Southern RegionalCouncil sponsored a symposium with far-reaching implications on Human Intelli-gence. Social Science and Social Policy.The principal presentation was made byDr. Leon Kamin, Chairman of the De-partment of Psychology at Princeton Uni-versity.

Dr. Kamin first presented findings ofhis extensive research into the originalstudies on which some American socialscientists have based writings which atleast question whether environment hasany effect on IQ test scoressuggestingthat heredity may be the determinant.Dr. Kamin presented his findings in aquiet, almost off-hand manner. But whathe said was startling and highly signifi-

rantin effect that there is no valid evi-dence at all to support the heredity as-sumption.

Since controversy over the heredity-environment question is one of the mostimportant intellectual debates of ourtimes, fraught with social policy implica-tions and influence, we consider the fol-lowing transcript of his presentationamong the most important materialsSOUTH TODAY has ever published.(Dr. Kamin spoke from notes; the tran-script was tape-recorded. The transcriptof his serond presentation. on the historyof the use of testing in America to sup-port repressive public policy, will appearin a subsequent issue of SOUTH.TODAY.)

percent which is not genetic? Well, saidProf. Herrnstein and I quote him: "Theusual assumption that education and cul-ture are critical is running into evidencethat the physical environment, forexample, early diet, might be more im-portant. In fact, the twin studies that[Prof. Arthur] Jenson [educational psy-chologist at the University of Californiaat Berkeley) surveyed showed that thesingle most important influence on IQwas not education or social environment,but something prenatal, as shown by thefact that the twin heavier at birth usuallygrew up with a higher IQ."

That's a rather astonishing finding.And it has clear policy implications. What

the statement says is that insofar as thereis an environmen.al effect at all on humanintelligence, its only the intrauterine en-vironment during pregnancy which hasany major effect. Therefore, perhaps weought to feed welfare mothers with aspecial protein supplement during preg-nancy. That might help the IQ's of theirchildren, but there's no sense in wastingmoney on the cultural, social, or educa-tional environment of these children. Bythe way, there's also no sense in feedingall of their fathers, if we are going to takethis seriously, at least not to improve thekids' IQ's.

But what's the basis for this statement?Well, what Prof. Herrnstein has done in

his article is basically to paraphrase anearlier article by Prof. Jensen. Prof.Jensen, writing in 1970, had said the fol-lowing: "A major proportion of the non-genetic variance is attributable to prenataland other biological differences, ratherthan differences in the social or psycho-logical environments. Differences in thefavorableness of the intrauterine environ-ment are reflected in differences in birthweight between twins. And the differencesin birth weight are known to be related toIQ differences in twins."

The same thingbut on what basisdoes Prof. Jensen make this claim? Hecites a single article by Sandra Scarr andhe presents in his paper some data fromScarr, who had measured the intelligenceof a number of identical twins and pre-sented information about how much thosetwins weighed at birth. Scarr concludesthat the heavier twin turns out almostinvariably to grow up with a higher IQ.The basic idea being, the heavier twinmust have had an easier and happier timein the womb, gotten more of the maternalblood supply, etc.

The first table is taken from Prof.Jensen's article, where he quotes datataken from Scarr. What Scarr did was tobreak twins up into three basic groups.For some twin pairs, it was true at birththat each weighed more than 2,500 grams.For other twin pairs, one twin weighedmore than 2,500, but the other twinweighed less than 2,500 grams. And for

To page 2

Weight of Twins Studies Do Not PrAs Claimed, Lack of InfluenceEnvironment on a Child's 10 Sc

From page 1the third type of twin, each twin weighedless than 2,500. Now the only data thatJensen presents are those which you seein Table I.

Table IMean IQ Difference in Favor of Heavier MZ Twins

Both TwinsGreater Than2500 grams

One TwinLess Than2500 grams

4.9 13.3

Both TwinsLess Than2500 grams

6.4

And what that data show is that in thecase where one twin weighed more than2,500 grams and the other less, theheavier twin on the average has an IQ13.3 points higher than the lighter twin.But notice that if both are reasonably fatand healthy, the heavier twin still has ahigher IQ, but not that much higher, only4.9 points; and similarity, if both twinsare light in weight, the difference betweenthem in IQ is not so large.

What Scarr concluded in her paperwas, "the greater discrepancy in thegroup with one twin below 2,500 gramsresulted from large differences in birthweight. Six of the eight pairs had birthdifferences of more than 500 grams. Thislarge discrepancy may be caused by trans-fusion syndrome. One monozygotic twinbleeds into the other. The size of the IQdifferences between co-twins was notaffected by the absolute weight of thetwins, but by the size of the weight dif-ferences between them."

Now she talks about single births, andconcludes from these data: "One mightconclude with justice, the bigger thebetter." The fatter the baby, presumably,the fatter the head, the fatter the brain,the higher the IQ. Strong biological rea-soning.

But there are some problems withScares data that I would like to call toyour attention. First, it makes very goodsense that she chose to work with twins,for the following reason. It is well knownthat there are social class and race differ-ences in birth weight, in the United Statesat least. The average white child weighsmore at birth than the average blackchild. The average upper-class childweighs more than the average lower-classchild. So the advantage of working with

- -I nro nnlv

27 pairs, the average difference is only6.3 IQ points. And that is a statisticallysignificant difference.

In short, if one had reasonable caution,one would not have put together in asingle analysis these two sets of twins,tested with two different so-called intelli-gence tests. That's a relatively small point.What about those three numbers in TableI that Prof. Jensen cites-4.9, 13.3, 6.something-or-other? If one does a tradi-tional, simple, standard analysis forwhether they differ significantly in astatistical sense, the answer is an un-qualified no. There is no reason to con-chide that the differences between thosethree numbers are due to anything otherthen chance. The differences betweenthem are not, to use a technical term,statistically significant. That's a relativelyminor point.

Now let's recall her logic, that if onetwin is more than 2,500 grams and theother is less than 2,500 grams, that meansthere is a big birth weight difference be-tween them. That means there's beensomething called transfusion syndrome.So, that in turn means that when there isa big birth weight difference, there willhe a big IQ difference.

It is interesting to look at the eighttwin pairs of whom it is true that oneweighed more than 2,500 grams and theother less. As it happens, if you look ather data, for six of those pairs the dif-ference in birth weight was in fact morethan 500 grams. So that's "transfusionsyndrone." And those twins differ on theaverage by 11 IQ points. But for theother two of her pairs, in that group, thedifference in weight was much less than500 grams. But with that small differencein birth weight, where presumably therehas been no "bleeding" of one twin intothe other in the womb, the difference inIQ turns out to be 20 points, not II!

As you will see from Table I, no mat-ter what category of twin you look at, thedifference in IQ actually turns out to belarger for the twins who are close togetherin birth weight than for twins sufferingfrom this hypothetical and never-observedbleeding and transfusion syndrome.

Scarr's logic says that the greater thedifference in birth weight between a pair

Other Contrary FindingsBut I don't think we ought to take it

all that seriously, for the following rea-sons. Prof. Jensen brings Scarr's data intothe very end of an article, during whichhe has been reviewing the evidence onseparated identical twins. There are foursuch studies which he reviewed in thatarticle. The only evidence that he giveshis readers about birth weight in twinscomes from this extra article by Scarrwhich he imports at the very end of hisreview article. That is peculiar. Because,as it happens, if one looks at the fourstudies which Prof. Jensen was reviewing,two of them had information on birthweight in twins. One of those studies wasin England, by Shields, and Shields, for61 pairs of twins, a larger sample thanScarr's, had asked the question doesthe heavier twin at birth have the higherIQ? He concluded, "Birth weight is notassociated with test score differences."Another of '1 four papers reviewed byProf. Jensen was by Juel-Nielsen. Shc alsoraised the question whether there was anyassociation between birth weight and IQ,itnd presented evidence to show no.

Those two studies that said no, Prof.Jensen did not choose to tell his readersabout. The one thz.: said yes, Prof. Jensenchose to cite. Now it's on the basis of thiskind of empirical evidence that the read-ers of Atlantic Monthly are assured thatpsychological science has demonstratedthat the major environmental variableaffecting IQ is something prenatal. I thinkif we want to make policy on the basisof this kind of science, we had better notuse science in making policy. I'm sorry tohave taken so much time on that singleexample.

Identical Twin StudiesThere is no possibility of reviewing this

morning all the different types of evidencethat psychologists have tried to bringtogether about the possible inheritance ofintelligence. I would like to restrict myremarks to what by consensus is thestrongest form of evidence, and also con-ceptually the simplest.

Primarily, I want to talk about thecelebrated studies on separated identical

anobromenashavdoge nhasThepare

AthetwipenUndonetwewastwezero

TtwinviesProt197peoNew(4)theytestscorrrear

Stu

BurtShieki

N.F.Juel-

Bwhictherstudforfirstneveidenanotteste

Tcti it

t of Twins Studies Do Not Prove,!aimed, Lack of Influence of

ronment on a Child's 10 Score

ahedthatsee

rwins

insanums---

i thethan

then IQtwin.:y fathas aonlytwinsween

paperthe

ramsbirthbirthThis

trans-twin

he IQs not,1 theIt dif-

s, andmightr thenably,brain,

rea-

withall togood

twins,nown

differ-StatesNeighsblackchild

r-classwith

27 pairs, the average difference is only6.3 IQ points. And that is a statisticallysignificant difference.

In short, if one had reasonable caution,one would not have put together in asingle analysis these two sets of twins,tested with two different so-called intelli-gence tests. That's a relatively small point.What about those three numbers in TableI that Prof. Jensen cites-4.9, 13.3, 6.something-or-other? If one does a tradi-tional, simple, standard analysis forwhether they differ significantly in astatistical sense, the answer is an un-qualified no. There is no reason to con-clude that the differences between thosethree numbers are due to anything otherthen chance. The differences betweenthem are not, to use a technical term,statistically significant. That's a relativelyminor point.

Now let's recall her logic, that if onetwin is more than 2,500 grams and theother is less than 2,500 grams, that meansthere is a big birth weight difference be-tween them. That means there's beensomething called transfusion syndrome.So, that in turn means that when there isa big birth weight difference, there willbe a big 10 difference.

It is interesting to look at the eighttwin pairs of whom it is true that oneweighed more than 2,500 grams and theother less. As it happens, if you look ather data, for six of those pairs the dif-ference in birth weight was in fact morethan 500 grams. So that's "transfusionsyndrone." And those twins differ on theaverage by 11 ,IQ points. But for theother two of her pairs, in that group, thedifference in weight was much less than500 grams. But with that small differencein birth weight, where presumably therehas been no "bleeding" of one twin intothe other in the womb, the difference inIQ turns out to be 20 points, not 11!

As you will see from Table I, no mat-ter what category of twin you look at, thedifference in IQ actually turns out to belarger for the twins who arc close togetherin birth weight than for twins sufferingfrom this hypothetical and never-observedbleeding and transfusion syndrome.

Scarr's logic says that the greater the_betwe n a pairIR

Other Contrary FindingsBut I don't think we ought to take it

all that seriously, for the following rea-sons. Prof. Jensen brings Scarr's data intothe very end of an articl, during whichhe has been reviewing the evidence onseparated identical twins. There are foursuch stud,. which he reviewed in thatarticle. The only evidence that he giveshis readers about birth weight in twinscomes from this extra article by Scarrwhich he imports at the very end of hisreview article. That is peculiar. Because,as it happens, if one looks at the fourstudies which Prof. Jensen was reviewing,two of them had information on birthweight in twins. One of those studies wasin England, by Shields, and Shields, for61 pairs of twins, a larger sample thanScarr's, had asked the question doesthe heavier twin at birth have the higherIQ? He concluded, "Birth weight is notassociated with test score differences."Another of the four papers reviewed byProf. Jensen was by Juel-Nielsen. She alsoraised the question whether there was anyassociation between birth weight and IQ,Ind presented evidence to show no.

Those two studies that said no, Prof.Jensen did not choose to tell his readersabout. The one that said yes, Prof. Jensenchose to cite. Now it's on the basis of thiskind of empirical evidence that the read-ers of Atlantic Monthly are assured thatpsychological science has demonstratedthat the major environmental variableaffecting IQ is something prenatal. l thinkif we want to make policy on the basisof this kind of science, we had better notuse science in making policy. I'm sorry tohave taken so much time on that singleexample.


morning all the different types of evidencethat psychologists have tried to bringtogether about the possible inheritance ofintelligence. I would like to restrict myremarks to what by consensus is thestrongest form of evidence, and also con-ceptually the simplest

Primarily, I want to talk about thecelebrated studies on separated identical

another a great deal, if they have beenbrought up in entirely different environ-ments, then that can be seen reasonablyas a consequence of the fact that theyhave the same heredity. Whereas, if theydo not resemble one another in intelli-gence at all, it is clear that environmenthas largely determined their intelligence.The logic of this experiment is trans-parently simple to anybody.

As it happens, there are four studies inthe psychological literature on separatedtwins. Everything I want to say now de-pends upon the correlation coefficient.Understand that if pairs of twins resembleone another perfectly, the correlation be-tween them w ill be plus 1.00. If therewas nothing but chance resemblance be-tween twins, the correlation would bezero.

There arc four studies of separatedtwins, and they have been graciously re-viewed for us by no less an authority thanProf. Jensen in Behavior Genetics in1970. The four studies were done bypeople named (I) Burt, (2) Shields, (3)Newman, Freeman and Holzinger, and(4) Juel-Nielsen. As you see in Table IIthey used rather different intelligencetests. The next column shows you thecorrelation coefficients for identical twinsreared apart,

Study

BurtShields

Table IIIQ Correl ttions For MZ Twins

Test Reared Apart

N.F. & H.Juel-Nielsen

"Individual Test"Dominoes

.77 + 2 x Mill HillStanford-Binet IQWechsler IQ

.86

.67

.62

But if you look at the third column,which is the critical one, you will see thatthere are some disagreements between thestudies. The correlations that are reportedfor separated twins vary from .86 in thefirst case to .62 in Juel-Nielsen's. Butnevertheless, it is clear that separatedidentical twins do indeed resemble oneanother in intelligence, when they aretested, to a very considerable degree.

Scrutiny of Burt's DataThe most important of these four

studies is by Cyril Burt, for several rea-

conclude with justice, the bigger thebetter." The fatter the baby, presumably,the fatter the head, the fatter the brain,the higher the IQ. Strong biological rea-soning.

But there are some problems withScarr's data that I would like to call toyour attention. First, it makes very goodsense that she chose to work with twins,for the following reason. It is well knownthat there are social class and race differ-ences in birth weight, in the United Statesat least. The average white child weighsmore at birth than the average blackchild. The average upper-class childweighs more than the average lower-classchild. So the advantage of working withidentical twins is that when you are onlycomparing twins you have an absolutecontrol for social class and for race. Thepa of twins is obviously born in thesame race and the same social class. Sothe basic logic makes sense.

So what Scarr had done was to take 25pairs of twins she had studied herself, andcombine those data with data from 27other pairs of twins who had been studiedby other people, namely Willerman andChurchill. And if one looks at the tableof raw numbers in Scarr's paper, one dis-covers such facts as the following: Thetest which Scarr used to measure "1Q"was something called the GoodenoughDraw-a-Person test. The test used byWillerman and Churchill was WechslerIntelligence Scale for Children, the Per-forn,ance Scale. Both had been combinedin Scarr's study, which may sound fairenough, but there are some problems.

If you look, for example, at the averagedifference between a pair of twins, forScarr's twenty-five pairs the average dif-ference is 14.4 IQ points. For the other

2 SOUTH TODAY

ay /1.1.4 6..13,,J, , U111.11.1 Loi.

average by 11 IC points But for theother two of her pairs, in that group, thedifference in weight was much less than500 grams. But with that small differencein birth weight, where presumably therehas been no "bleeding" of one twin intothe other in the womb, the difference inIQ turns out to be 20 points, not i 1!

As you will see from Table I, no mat-ter what category of twin you look at, thedifference in IQ actually turns out to belarger for the twins who are close togetherin birth weight than for twins sufferingfrom this hypothetical and never-observedbleeding and transfusion syndrome.

Scarr's logic says that the greater thedifference in birth weight between a pairof twins, the greater should be the differ-ence in their IQ's. But when calculated,the correlation between the difference inbirth weight and the difference in IQturns out to be .03. There is no correla-tion whatsoever between the differencein birth weight and the difference in IQnone whatsoever. But I don't want to beunfair to Prof. Scarr. There is, in fact,one statistically significant effect in herdata. That is the following: Suppose youdon't ask, what is the difference in birthweight, but ask a very much simplier andblunter question. For any pair of twins,one is heavier and the other lighter.Forget how much heavier or how muchlighter, and just ask the question, is ittrue that the heavier twin a:. birthevenif he's only one gram heavierhas thehigher IQ, even if that IQ was 100 pointshigher? The answer is, yes. For her 52pairs of twins, it is true that in 37 casesthe heavier twin had a higher IQ than thelighter. For eight cases, the lighter twinhad a higher IQ. In seven cases therewere no differences. That is a statisticallysignificant effect.

if we want to make policy on the basisof this kind of science, we had better notuse science in making policy. I'm sorry tohave taken so much time on that singleexample.


morning all the different types of evidencethat psychologists have tried to bringtogether about the possible inheritance ofintelligence. I would !.KC to restrict myremarks to what by consensus is thestrongest form of evidence, and also con-ceptually the simplest.

Primarily, I want to talk about thecelebrated studies on separated identicaltwins. The point of these studies is reallyvery simple. Identical twins, who arequite rare, are the only individuals in theworld who have literally the same genes.As far as their heredity is concerned, theyare identical. Not all twins, of course, areidentical twins. But it can he determinedwith reasonable accuracy, by blood test-ing and various other procedures, whethertwins are identical or not. If they are,then literally they have the same genes.

So in a way, nature provides thepsychologist with a kind of experiment.Suppose you have a pair of identicaltwins, who for one reason or another areseparated very early in childhood. Per-haps the mother dies, perhaps the familycan't afford to bring up twins, whateverreason, one of these identical twins isreared in one home, in one environment,and the other twin is reared in a differenthome, a different environment.

If one can find such separated twins,and give them intelligence tests, then theinteresting question to ask is, do theyresemble one another in intelligence? Tothe extent that they do resemble one

N.F. & H.JucI -N IC Ise

But itwhich isthere arstudies.for sepafirst ca.nevertheidenticalanothertested, t

Thestudiessons. Finumberpairs,correlatiboth Preout, Bupresentseconomdifferenportant,it is theis placeanotherworkingbe placeate hornIf the mdies, anhomes,sional hfact thein diffportantkinds oonly in

r thenably,brain,tl rea-

with._all to

goodtwins,nown

differ -States,neighsblackchild

'r-classe withc onlynsolutee. Thein the,Iss. So

ake 25'If, and.gym 27studiedan ande table,ne dis-g: Thec "IQ"enoughsee, by'easierhe Per-mbinednd fairterns.

averagens, for

age dif-e other

ay 11U10.1114 (111%.1 111.111,3 %.111141 tn. .11,

average by 11 IQ points. But for theother two of her pairs, in that group, thedifference in weight was much less than500 grams. But with that small differencein birth weight, where presumably therehas been no "bleeding" of one twin intothe other in the womb, the difference inIQ turns out to be 20 points, not I I!

As you will see from Table I, no mat-ter what category of twin you look at, thedifference in IQ actually turns out to helarger for the twins who are close togetherin birth weight than for twins sufferingfrom this hypothetical and never-observedbleeding and transfusion syndrome.

Scarr's logic says that the greater thedifference in birth weight between a pairof twins, the greater should he the differ-ence in their IQ's. But when calculated,the correlation between the difference inbirth weight and the difference in 1Qturns out to be .03. There is no correla-tion whatsoever between the differencein birth weight and the difference in IQnone whatsoever. But I don't want to heunfair to Prof. Scarr. Th-re is, in fact,one statistically significant effect in herdata. That is the following: Suppose youdon't ask, what is the difference in birthweight, but as a very much simplier andblunter question. For any pair of twins,one is heavier and the other lighter.Forget how much heavier or how muchlighter, and just ask the question, is ittrue that the heavier twin at birthevenif he's only one gram heavierhas thehigher IQ, even if that IQ was 100 pointshigher? The answer is, yes. For her 52pairs of twins, it is true that in 37 casesthe heavier twin had a higher IQ than thelighter. For eight cases, the lighter twinhad a higher IQ. In seven cases therewere no differences. That is a statisticallysignificant effect.

if we want to make policy on the basisof this kind of science, we had better notuse science in making policy. I'm sorry tohave taken so much time on that singleexample.

Identical Twin SfudiesThere is no possibility of reviewing this

morning all the different types of evidencethat psychologists have tried to bringtogether about the possible inheritance ofintelligence. I would like to restrict myremarks to what by consensus is thestrongest form of evidence, and also con-ceptually the simplest.

Primarily, I want to talk about thecelebrated studies on separated identicaltwins. The point of these studies is reallyvery simple. Identical twins, who arequite rare, are the only individuals in theworld who have literally the same genes.As far as their heredity is concerned, theyare identical. Not all twins, of course, areidentical twins. But it can he determinedwith reasonable accuracy, by blood test-ing and various other procedures, whethertwins are identical or not. If they are,then literally they have the same genes.

So in a way, nature provides thepsychologist with a kind of experiment.Suppose you have a pair of identicaltwins, who for one reason or another areseparated very early in childhood. Per-haps the mother dies, perhaps the familycan't afford to bring up twins, whateverreason, one of these identical twins isreared in one home, in one environment,and the other twin is reared in a differenthome, a different environment.

If one can find such separated twins,and give them intelligence tests, then theinteresting question to ask is, do theyresemble one another in intelligence? Tothe extent that they do resemble one

1/0111111oeS

.77 -I- 2 x Mill Hin& If. Sonford-Binet 10

Juel-Nielsen Wechsler IQ.67.62

But if you look at the third column,which is the critical one, you will see thatthere are some disagreements between thestudies. The correlations that are reportedfor separated twins vary from .86 in thefirst case to .62 in Juel-Nielsen's. Butnevertheless, it is clear that separatedidentical twins do indeed resemble oneanother in intelligence, when they aretested, to a very considerable degree.

Scrutiny of Burt's DataThe most important of these four

studies is by Cyril Burt, for several rea-sons. First of all, he studied the largestnumber of pairs of separated twins 53pairs. Secondly, he reported the highestcorrelation of any study. And thirdly, asboth Profs. Herrnstein and Jensen pointedout. Burt is the only investigator whopresents quantitative data on the social oreconomic class of the homes in which thedifferent twins were reared. That is im-portant, for the following reason. Supposeit is the case that a pair of identical twinsis placed in homes that resembled oneanother very closely. If the mother is aworking-class mother, both twins mighthe placed in working class homes, separ-ate homes but both working class homes.If the mother is a professional person anddies, and the twins go to two separatehomes, both might be placed in profes-sional homes. One must be certain that infact the separated twins have been placedin different types of homes. It's im-portant to have this information on whatkinds of homes they were placed in, It'sonly in the Burt study that we were given

such information. So that makes the Burtstudy theoretically more important thanany other.

Burt died a couple of years ago. Hewas English. He certainly was the majorcontributor by for to the data on heritabil-ity of human intelligence; a tremendousproportion of the data in this whole area,on how much relatives of different de-grees of biological relatedness resembleeach other in IQ, is due to the work ofBurt and his students. He is the monu-mental figure in the area.

So I want to spend a little time lookingat Prof. Burt. Let me say at once thatwhenever one reads Burt's papers, there

hygienic conditions." And what he meansby cultural status is "educational andmotivation background."

Now look at that .315. It now turns outthat the correlation of .315 (or .32) isbetween cultural status and intelligence!Whereas we had been assured in 1943that the correlation was between economicstatus and intelligence-definitely notcultural status. And we weren't told in1943 that there had been any measure-ment at all of cultural status. There is asimple flat contradiction here, and this isnot unrepresentative of Burt. I think thequestion has to be asked in all seriousness,

the intelligence correlation. And if youlook at unrelated children reared to-gether, there were more children againin 1966 than in 1955. Once again onediscovers that most correlation coeffi-cients haven't been changed appreciablyby the additions of extra cases. The prob-ability of this occurring is so astronomi-cally small that one must conclude thatthere is no reasonable interpretation left-something is wrong. It just doesn't happenthat way. You don't add 20 new pairs oftwins and find all those correlations re-maining identical.

Consider the case of non-identicaltwins. Burt reported on 172 pairs in

.944. By thewith 95 pairs.944.

That's realof intelligenting such staUnless youintelligence I

can't apprecidata are.

The testingtions quotedgives us theand child,grandchild,and so on. T

"...Whenever one reads Burt's paperthere are certain ambiguities in t

are certain ambiguities in them. It's oftenvery difficult to know precisely what hehas done to whom and when. He presentscorrelation coefficients, but he doesn'tgive us very much information about whattest he used, to whom he gave it, and soon.

I will give you some examples. In thefirst paper that he published which re-ported correlation coefficients for differ-ent types of relatives in 1943, Burt tellsus: "Some of the inquiries have been pub-lished in London County Council reportsor elsewhere. But the majority remainburied in typed memoranda or degreetheses." It is going to be hard to getprecise details about many of his studies,and often one wishes one had precisedetails for reasons such as the following.

If one starts in 1943, in that originalpaper, Burt reported a correlation be-tween what he then called intelligenceand "economic status." The higher theeconomic status, the higher the intelli-gence. The correlation was reported as.32. In that paper, Burt made a big pointof stating that what he had measured (hedoesn't tell us how he measured it) waseconomic status, not cultural status. Hewants to make that point because, as heindicates, other investigators have showna higher correlation than .32 betweencultural status and intelligence. There-fore, to explain the fact that his correla-tion was only .32, he very specifically in-forms us that he measured economicstatus and not cultural status. ,

In 1956, with Burt referring to exactlythe same study, we discover something.Now we get different sets of correlations.One set of correlation's is for somethingthat Burt now calls the "crude test re-sults," and the ether is something that he

would you buy a used car from a sales-man with this kind of record?

What Burt did was to go through lifecollecting cases. He got interested in thisproblem relatively early in his life. Andfor example, at one point he would writea paper in which he might have a hi'ndredpairs of twins, and over the next fiveyears he might discover 25 more pairs,and five years later he would write a newpaper, now based on 125 pairs of twins.He simply kept on adding cases during avery lengthy career. He died in his eight-ies, and was still working away.

Consider the case of siblings rearedapart. In 1955, Burt had reported on 131pairs of siblings reared apart. By 1966,the number of pairs had increased to 151.There were some remarkable coinci-dences, as shown in Table 111. The addi-

Table HICorrelations From Burt

SiblingsReared Apart

DZTwinsReared

Together1955 1966 1955 1966

IntelligenceGroup Test .441 .412 .542 .552Individual Test .463 .423 .526 .527Assessment .517 .438 .551 .453

School AttainmentReading. Spelling .490 .490 .915 .919Arithmetic .563 .563 .748 .748General .526 .526 .831 .831

PhysicalHeight .536 ,.536 .472 .472Weight .427 .427 .586 .586

tion of 20 new pairs of twins didn't changeto the third decimal place the correlationseither for weight or height or generalschool attainment or arithmetic or En-glish or spelling. The addition of 20 newpairs of twins didn't affect one of thosecorrelation coefficients even to the third

1965. By 1966, 45 of those pairs haddisappeared, We now have only 127. Butthat's all right; it doesn't change thecorrelation coefficient for height andweight and general school attainment andarithmetic. It makes no difference, inProf. Burt's data, whether you add pairsor lose pairs, you get the same thing allthe time. That's rather surprising, and itdoesn't inspire confidence.

Now, this has to do with the separatedidentical twin studies. Prof. Burt evi-dently had a marvelous test, a grouptest of intelligence, which produced somereally remarkable data. Concern yourselfwith only the first column and the fourthcolumn of Table IV. The first column

Separated

Table IVCorrelations For MZ Twins

Reared Totr:.bur

Group Test of Intelligence Group Trot diIntel ligtmcc

1955: .771 N=21 .944 N.-,871958(A): .771 N="over 30" .944 N-----.

1958(B): .778 N=42 .936 N ?1966: .771 N=53 .944 W:95

First and fourth columns referred to kr, tr::

gives you the correlation coeffi6en4zi..separated identical twins. At ,c1;:ifroi.:,r,tpoints in Burt's career he reportedprogressively larger and larger relmbersof such twins. In his 1955 pape7, when hehad only 21 pairs of twins, correla-tion for this marvelous group test of intel-ligence was .771. By early 3.953, he hadfound at least nine more pairs. Novi, hehad over 30, but the correlationficient was still .771. And for $3 .c.,!airs in1966, it is still .771!

Now this same marvelous gro,,,lp in-telligence test was also ,givru it non-separated identical twins and notice thatwit 83 airs he had .94,4__A,xm._irk, an

i0417wsrarAtatamiza

those eoTrelagenetic stnshe ritedt Burtand it..:!; ',cryto clOe.venscIticelroom :an y wwhin .yrai gi

,owof ac

'Y-57 tell usof &via' sury

attats:pkiple, of pOf 5:findardivve." Rene.vs. Most

rental intelliwere gotten-alit to the'5C:SSment ofk,ence is. \meet, you4:tild's IQwhat kind oreassures usrepresentatiintelligence

Burt ancorrelationschild, andand grandcus in that pthe intelligeprocedureshave alreadpublications.ious publicareads that pa footnoteinformation:parents, weinterviews. Bcases, an opemployed."

rt conditions." And what he meansin by cultural status is "educational and

motiva!,:,-,n background."kt.ok at that .315. It now turns out

the rite ,2orrelation of .315 (or .32) isbetv.rt.ken. cultural status and intelligence!onl,r,;:.as- we had been assured in 1943

:II-At Cue correlation was between economicsr.:ms. :mid intelligence-definitely notz-,.;ritrtal status. And we weren't told inr94t3 that there had been any measure-

:mei-it at all of cultural status. There is asimple, flat contradiction here, and this isnot unrepresentative of Burt. I think thequestion has to be asked in all seriousness,

mtmaf Arrcff

or11-

usa.

e-le,f

LI-

the intelligence correlation. And if youlook at unrelated children reared to-gether, there were more children againin 1966 than in 1955. Once again onediscovers that most correlation coeffi-cients haven't been changed appreciablyby the additions of extra cases. The prob-ability of this occurring is so astronomi-cally small that one must conclude thatthere is no reasonable interpretation left-something is wrong. It just doesn't happenthat way. You don't add 20 new pairs oftwins and find all those correlations re-maining identical.

Consider the case of non-identicaltwins. Burt reported on 172 pairs in

.944. By the time he concludes his careerwith 95 pairs of those twins, he's still at.944.

That's really a remarkable group testof intelligence that's capable of provid-ing such stable data. Truly remarkable.Unless you have worked yourself withintelligence tests and correlations, youcan't appreciate how remarkable Burt'sdata are.

The testing literature is full of comia-tions quoted from Burt, in wl...iCh hegives us the correlation bevy, een parentand child, between grandparent andgrandchild, between Linde and nephew,and so on. The mint is to show that all

4/aver one reads Burt's papers'are certain ambigtOes in them."

e

ts

at0

IC

e-

Is

is

would you buy a used car from a sales-man with this kind of record?

What Burt did was to go through lifecollecting cases. He got interested in thisproblem relatively early in his life. Andfor example, at one point he would writea paper in which he might have a hundredpairs of twins, and over the next fiveyearS he might discover 25 more pairs,and five years later he would write a newpaper, now based on 125 pairs of twins.He simply kept on adding cases during avery lengthy career. He died in his eight-ics, and was still working away.n

Consider the case of siblings rearedt apart. In 1955, Burt had reported on 131

s, pairs of siblings reared apart. By 1966,e the number of pairs had increased to 151.

There were some remarkable coinci-dences, as shown in Table III. The addi-

Table IIICorrelations From Burt

IntelligenceGroup TestIndividual TestAssessment

School AttainmentReading, SpellingArithmeticGeneral

Physicalre Height

Weight

lyg.s.lg

he

SiblingsReared Apart1955 1966

.441.463.517

.490

.563

.526

.412.423.438

.490.563.526

.536 , .536

.427 .427

DZTwinsReared

Together1955 1966

.542.526.551

.915

.748

.831

.552.527.453

.919

.748

.831

.472 .472

.586 .586

tion of 20 new pairs of twins didn't changeto the third decimal place the correlationseither for weight or height or generalschool attainment or arithmetic or En-glish or spelling. The addition of 20 newpairs of twins didn't affect one of those

ci s even to_

1955, By 1966, 45 of those pairs haddisappeared. We now have only 127. Butthat's all right; it doesn't change thecorrelation coefficient for height andweight and general school attainment andarithmetic. It makes no difference, inProf. Burt's data, whether you add pairsor lose pairs, you get the same thing allthe time. That's rather surprising, and itdoesn't inspire confidence.

Now, this has to do with the separatedidentical twin studies. Prof. Burt evi-dently had a marvelous test, a grouptest of intelligence, which produced somereally remarkable data. Concern yourselfwith only the first column and the fourthcolumn of Table IV. The first column

Table IVCorrelations For MZ Twins

Separated Reared Together

Group Test of Intelligence

1955:1958(A):1958(B):1966:

.771

.771

.778

.771

N=21N="over 30"N=42N=53

Group Test ofIntelligence

.944 N=83.944 N= ?.936 N.-- ?.944 N=95

First and fourth columns referred to in text

gives you the correlation coefficients forseparated identical twins. At differentpoints in Burt's career he reported onprogressively larger and larger numbersof such twins. In his 1955 paper, when hehad only 21 pairs of twins, the correla-tion for this marvelous group test of intel-ligence was .771. By early 1958, he hadfound at least nine more pairs. Now hehad over 30, but the correlation coef-ficient was still .771. And for 53 pairs in1966, it is still .711!

Now this same marvelous group in-telligence test was also given to non-separated identical twins and notice that

1

those correlations make perfect biologicalgenetic sense, that intelligence is in-herited. Burt was a school psychologist,and it's very easy to give intelligence teststo children, particularly if you are aschool psychologist. They are in a class-room anyway, they're a captive audiencewhen you give them a test.

But how do you measure the intelli-gence of adults? Burt and Howard in1957 tell us the following: "But in eachof our surveys assessments were individ-ually attained for a representativesample of parents, checked for purposesof standardization by tests of the usualtype." Reflect for a moment on what thatsays. Most of the measurements for pa-rental intelligence, or adult intelligence,were gotten by assessment. That is, youtalk to the adult and you make an as-sessment of what you think his intelli-gence is. When you make that assess-ment, you probably know what hischild's IQ is, and you certainly knowwhat kind of home he lives in. Prof. Burtreassures us that he does give, at least to arepresentative sample of adults, a realintelligence test of the usual type.

Burt and Howard in 1966 reportedcorrelations for 963 pairs of parent andchild, and for 321 pairs of grandparentand grandchild. The only thing they tellus in that paper about how they measuredthe intelligence of the adults was, "Theprocedures employed and results obtainedhave already been described in previouspublications." And they cite one prev-ious publication. If one goes back andreads that previous publication, one findsa footnote which gives us this bit ofinformation: "For the assessment of theparents, we relied chiefly on personalinterviews. But in doubtful and borderlinecases, an open or a camouflaged test wasemployed." Now, the picture of Prof.

or stating mat WIlal ne nau 111CitNtlit-u A H.;

doesn't tell us how he measured it) waseconomic status, not cultural status. Hewants to make that point because, as heindicates, other investigators have showna higher correlation than .32 betweencultural status and intelligence. There-fore, to explain the fact that his correla-tion was only .32, he very specifically in-forms us that he measured economicstatus and not cultural status.

In 1956, with Burt referring to exactlythe same study, we discover something.Now we get different sets of correlations.One set of correlation's is for somethingthat Burt now calls the "crude test re-sults," and the other is something that hecalls "adjusted assessments." In the 1943paper we diun't know much about thisdistinction. What it means is this. Yougive someone an intelligence test and youscore it, that's the crude test mark. Butif you feel there is some inaccuracy inthe test, it didn't give a perfect measure,then you adjust it. You take the test scoreand if it doesn't seem accurate to you,you adjust it.

It turns out, as you will see, that thecorrelation which in 1943 had been re-ported as between intelligence and eco-nomic status was really the correlationbetween the adjusted assessment (aboutwhich we hear for the first time in 1956)and "socio-economic status." But now in1957, Burt has another paper where herefers once again to exactly the samestudy, and look at what has happened. Itturns out that he evidently had all alongnot only a measure of economic status,but a measure of cultural status. Burt'suse of words is getting loose now, and itturns ou. what he means by economicstatus is "material, i.e. financial and

117Ja 1171.113 1733 711t1Intelligence

Group Test .441 .412 .542 .552Individual Test .463 .423 .526 .527Assessment .517 .418 .551 .451

School AttainmentReading. Spelling .400 .490 .915 .919Arithmetic .563 .561 .748 .748General .526 .526 .811 .831

Physicalflciyht .536 ..516 .472 .472Weight .427 .427 .586 .586

Lion of 20 new pairs of twins didn't changeto the third decimal place the correlationseither for weight or height or generalschool attainment or arithmetic or En-glish or spelling. The addition of 20 newpairs of twins didn't affect one of thosecorrelation coefficients even to the thirddecimal place! It did slightly change

41111Mmbe

Post and fourth columns relerred to in text


Now this same marvelous group in-telligence test was also given to non-separated identical twins and notice thatwith 83 pairs he had .944. Again, for anunspecified number of pairs in 1958,

George L. Walker III

Dr. Komin Discusses Presentation With SRC Staffer Alexis Herman

Burt acorrelatiochild, anand granus in thatthe intelliprocedurhave alrepublicanions publreads thaa footnoinformatiparents,interviewcases, anemploye.Burt givigence tohaps aninspireminds.

And ntion. Requotesstudy. Inline casememory,tive samwere "o1957, inries, theytyPe."deal of c

Neverttextbooksarticles i

learned jtions resion thdemonstrtelligenceand that

S

isIntelligence

£703 171)0 1730 171)0

Group Test .441 .412 .542 .552Individual Test .463 .423 .526 .527

n Assessment .517 .438 .551 .453

nSchool Attainment

Reading, Spelling .490 .490 .915 .919C- Arithmetic .563 .563 .748 .748

General .526 .526 .831 .831

PhysicalC Height .536 , .536 .472 .472

Weight .427 .427 .586 .586

.2.

S.

3is

ut

utInc,CC

u,

hec-o-onut

6)in

he

Itng

US,

isitIC

tion of 20 new pairs of twins didn't changeto the third decimal place the correlationseither for weight or height or generalschool attainment or arithmetic or En-glish or spelling. The addition of 20 newpairs of twins didn't affect one of thosecorrelation coefficients even to the thirddecimal place! It did slightly change

First and fourth columns referred to in teal


Now this same marvelous group in-telligence test was also given to non-separated identical twins and notice thatwith 83 pairs he had .944. Again, for anunspecified number of pairs in 1958,

George L Walker Illnd Dr. Kamin Discusses Presentation With SRC Staffer Alexis Herman

al 11:616.9.011..... 11194. 441.441. y Fs..Burt and 'coward in 1966 reported

correlations frr 963 pairs of parent andchild, and 321 pairs of grandparentand grandchild The only thing they tellus in that paper about how they measuredthe intelligence of the adults was, "Theprocedures emrloycd and results obtainedhave already been described in previouspublications." And they cite one prev-ious publication. if one goes back andreads that previous publication, one findsa footnote which gives us this bit ofinformation: "For the assessment of theparents, we relied chiefly on personalinterviews. But in doubtful and borderlinecases, an open or a camouflaged test wasemployed." Now, the picture of Prof.Burt giving a camouflaged test of intelli-gence to p grandparent in London is per-haps ut entertaining one, but it does notinspire scientific confidence in mostminds.

And note also the magical transmuta-tion. Remember that each of the followingquotes refers to precisely the samestudy. In 1955, it is "doubtful and border-line cases," but by 1957 in Prof. Burt'smemory, this has become "a representa-tive sample of parents." In 1955 the testswere "open or camouflaged." But by1957, in the mist of Prof. Burt's memo-ries, they had become "tests of the usualtype." This doesn't really inspire a greatdeal of confidence.

Nevertheless, all of our introductorytextbooks in psychology, all our reviewarticles in Atlantic Moneily and otherlearned journals will quote the correla-tions reported by Burt, giving the impres-sion that psychological science hasdemonstrated that the correlations in in-telligence test scores for relatives is thisand that and that. The poverty of the

To page 4

SOUTH TODAY 3

Four Studies of Separated Identical TwiAre Shot Through With rif q:hodological

mown page 3empirical and procedural backgrounddoesn't come through when you read re-,views of what is alleged to be shown bythese data.

How about this question of usingassessments rather than using test scores?Let me read a couple of quotations, fromProf. Burt. Burt writes in 1958, "The datawere secured in the course of surveys.Thz final assessments for the childrenwere obtained by submitting the marksfrom the group tests to the judgment ofthe teachers. Where the teacher disagreedwith the verdict of the marks, the childwas interviewed personally and subjectedto further tests, often on several succes-sive occasions. The assessments for theadult members of the families werenaturally far less accurate." That is, if'aparticular child gets an intelligence tsstscore which doesn't seem to fit, he ca.)be tested and has been tested by Prof.Burt a second time and a third time, andif necessary, a fourth time. How muchcredibility are we to attach to correlationcoefficients based not upon the standard-ized administration of an intelligence testbut upon a repeatedly adjusted assess-ment made by teachers or by Prof, Burt?The most detailed explanation that Burtmakes of this assessment procedure is asfollows: "We are perfectly willing to ad-mit [Ns Ls Burt in 1956] that as a meansof estimating genotypic differences, [thatis, differences due to the genes] tests arehighly fallible instruments, far less trust-worthy than the judgments of teachers.An officially appointed psychologist pos-sesses the authority to extend or repeathis tests and his interviews, and to requirefrom teachers further assistance. The datahave been secured in this way and havingsatisfied ourselves that by these means wecan reduce the disturbing effects of en-vironment to relatively slight proportions,we have gone on . . ," etc.

So what Prof. Burt is doing is de-liberately removing from his numbers"the disturbing effect of environment."Because he is interested only in the al-leged effect of the genes, he removes thedisturbing effect of environment on thetest scores by adjusting the scores interms of comments and criticisms madeby the teachers. What's very peculiarabout this is that these adjusted scoreswere then maned by Prof. Burt to, among

Now, anybody is entitled to changehis mind, but it is instructive to ask inwhat context Prof. Burt made that earlystatement. The earlier quotation was in1943, when the debate was already on inEngland about who was to be allowed toattend university, and how children wereto be streamed in the public schools. Atthat point, Prof. Burt was arguing essen-tially the following: We should segregateearly on in their careers the geneticallybright. They are the ones who will heallowed to go on to universities. And tofind out who is genetically bright, tofind out who is going to he allowed to gointo what classroom, he advocated useof the test score because it's a better esti-mator of genes. Don't allow the teacherto influence student placement!

When it suited Prof. Burt's purpose tokeep the teacher out of socially importantdecisions, he argued that the test was amarvelous measure of hereditary intel-ligence and the teacher was not to betrusted. But later in his career, Prof.Burt had test data which he tried to fit toa theoretical model of inheritance, andthe data didn't fit that well. Then heturned to the teacher to help him out.Now the teacher was the better estimatorof the genotype.

Twin Study Inconsistencies

We finally get to the separated twinsstudy itself. It would be reasonable toask what intelligence test did Burt give tothe separated twins? The answer is, I'mafraid, we do not know. Prof. Jensendoes know. He informed us very confi-dently in 1970, talking about Burt's data,"their IQ's were obtained from an indi-vidual test, the English adaptation of theStanford-Binet." They were not, by theway.

Burt in 1966 had said the following:"The tests employed have been fully de-scribed elsewhere. They consisted of (1)a group test of intelligence containingboth non -ver' .1 and verbal items; (2) anindividual test, the London revision of theTerman-Binet scale, used primarily forstandardization, and for doubtful cases,(3) a set of performance tests standard-ized by Miss Gaw in 1925." Of coursethe test results were submitted to teach-ers, and so on. Now the interesting thing

one of those seven, if any, he used. Infact he couldn't have used any of thoseseven, because he says clearly, "All thegroup tests in the present category arelinguistic in form." But elsewhere he hasinformed us that his marvelous intelli-gence test had both verbal and non-verbalitems.

Now, about those seven group testswhich he gives us, he tells us such thingsas the following: "Complete tables of agenorms would he unnecessary or evenmisleading. I give only rough averages,calculated regardless of sex. I have notthought it worth the necessary time andspace to elaborate or print a set ofstandardized instructions as to proceduresor marking."

Nevertheless, that's the test whichrepeatedly gives us that remarkably stablecorrelation of .771, no matter how largethe number of twins. This is really a ter-ribly important question, for the follow-ing reason: Intelligence tests in theoryhave been constructed in such a waythat, for each sex the average IQ is 100,and for each age, the average IQ is 100.To the degree that the test might possiblynot be perfect, we are going to find

. serious problems. Consider the following:Any pair of identical twins is necessarilyof the same sex and of the same age.Suppose you are using a test where, forexample, men score higher than women,or where older children score higher thanyounger children. That means that whenyou compare twins, they may resembleone another simply because they arcthe same sex or simply because they arethe same age. And not because theyhave the same genes. Unless one is as-sured that the test which one is using inthese studies does not give differentscores for different ages, or for differentsexes, the studies simply hold no water.1 have looked into Miss Gaw's perfor-mance test, standardized in 1925. Sheinforms us that there are "striking" dif-ferences between the sexes in the perfor-mance test that Prof, Burt used.

Recall that Burt's is the only studythat gives us data on the social and eco-nomic status of the homes in whichseparated twins were placed. By 1958Burt had reported on 42 pairs of sepa-rated twins. In 1958 we were told that

logo fonr phirlren of hinh nrofecsinn-41

of thepsycholcontemogy, asfashionninetecpreachthreeworkform h

Burtwhat ;also dsecond40 palus witlvidcs twe cotsomdetails

FirsShieldsgists nidenticAppen"separwerebrouglage,grandgethersameyears-ttheirShieldresem

Jessat thifew htractebut mthat tcoverone aquitenings.fred.to sitseparaftercover

Buat birtakeamial-

es of Separated Identical Twinshrough With Methodological Flaws

oundad re-m by

using:ores?fromdataveys.Idrennarksnt ofgreedchild

metedicces-ir thewere

. if* atest

. canProf.. andnuch

lationdard-e testssess-

Burt?Burtis as) ad-leans[that

is aretrust-:hers.

pos-epeatquiredata

avingIs weI en-:ions,

de-bers

lit."e al-s the

thes innade'liar

cores,one

Now, anybody is entitled to changehis mind, but it is instructive to ask inwhat context Prof. Burt made that earlystatement. The earlier quotation was in1943, when the debate was already on inEngland about who was to be allowed toattend university, and how children wereto he 'streamed in the public schools. Atthat point, Prof. Burt was arguing essen-tially the following: We should segregateearly on in their careers the geneticallybright. They are the ones who will beallowed to go on to universities. And tofind out who is genetically bright, tofind out who is going to he allowed to gointo what classroom, he advocated useof the test score because it's a better esti-mator of genes. Don't allow the teacherto influence student placement!

When it suited Prof. Burt's purpose tokeep the teacher out of socially importantdecisions, he argued that the test was amarvelous measure of hereditary intel-ligence and the teacher was not to betrusted. But later in his career, Prof.Burt had test data which he tried to fit toa theoretical model of inheritance, andthe data didn't fit that well. Then heturned to the teacher to help him out.Now the teacher was the better estimatorof the genotype.

Twin Study Inconsistencies

We finally get to the separated twinsstudy itself. It would be reasonable toask what intelligence test did Burt give tothe separated twins? The answer is, I'mafraid, we do not know. Prof. Jensendoes know. He informed us very confi-dently in 1970, talking about Burt's data,"their IQ's were obtained from an indi-vidual test, the English adaptation of theStanford-Binet." They were not, by theway.

Burt in 4 966 had said the following:"The tests employed have been fully de-scribed elsewhere. They consisted of (I)a group test of intelligence containingboth non-verbal and verbal items; (2) anindividual test, the London revision of theTerman-Binet scale, used primarily forstandardization, and for doubtful cases,(3) a set of performance tests standard-ized by Miss Gaw in 1925." Of coursethe test results were submitted to teach-ers, and so on. Now the interesting thing

one of those seven, if any, he used. Infact he couldn't have used any of thoseseven, because he says clearly, "All thegroup tests in the present category arelinguistic in form." But elsewhere he hasinformed us the his marvelous intelli-gence test had both verbal and non- vLrhalitems.

Now, about those seven group testswhich he gives us, he tells us such thingsas the following: "Complete tables of agenorms would he unnecessary or evenmisleading. I give only rough averages,calculated regardless of sex. I have notthought it worth the necessary time andspace to elaborate or print a set ofstandardized instructions as to proceduresor marking."

Neverthe.:ss, that's the test whichrepeatedly gives us that remarkably stablecorrelation of .771, no matter how largethe number of twins. This is really a ter-ribly important question, for the follow-ing reason: Intelligence tests in theoryhave been constructed in such a waythat, for each sex the average IQ is 100,and for each age, the average IQ is 100.To the degree that the test might possiblynot be perfect, we are going to find

. serious problems. Consider the following:Any pair of identical twins is necessarilyof the same sex and of the same age.Suppose you are using a test where, forexample, men score higher than women,or where older children score higher thanyounger children. That means that whenyou compare twins, they may resembleone another simply because they arethe same sex or simply because they arethe same age. And not because theyhave the same genes. Unless one is as-sured that the test which one is using inthese studies does not give differentscores for different ages, or for differentsexes, the studies simply hold no water.I have looked into Miss Gaw's perfor-mance test, standardized in 1925. Sheinforms us that there are "striking" dif-ferences between the sexes in the perfor-mance test that Prof. Burt used.

Recall that Burt's is the only studythat gives us data on the social and eco-nomic status of the homes in whichseparated twins were placed. By 1958Burt had reported on 42 pairs of sepa-rated twins. In 1958 we were told thatII AC 111

of the relations between politics andpsychology than some of his Americancontemporaries. He wrote. "In psychol-ogy, as in politics, the pendulum of thefashion swings to and fro. During thenineteenth century, the associationistspreached an egalitarian doctrine, andthree reform bills were passed." Burt'swork is not likely to lead to any new re-form bills.

Other Twin StudiesBurt's is only one of the four studies;

what about the others? The second wasalso done in England, by Shields. It is thesecond largest study, with approximately40 pairs of twins. Where Burt providesus with no procedural details, Shields pro-vides us with all the possible informationwe could ask for. I'm going to skip oversome of the more technical proceduraldetails.

First of all, we can get from Prof.Shields an indication of what psycholo-gists mean when they talk about separatedidentical twins. I'm quoting from hisAppendix: These three cases were"separated twins." Benjamin and Ronaldwere separated at nine months. "Bothbrought up in the same fruit-growing vil-lage, Ben by the parents, Ron by thegrandmother. They were at school to-gether, they have continued to live in thesame village." Ben and Ron were 52-years -old when they were brought fromtheir village to London, to be tested byShields, who then found out that theyresembled one another.

Jessica and Winifred were separatedat three months. "Brought up within afew hundred yards of one another. At-tracted to each other at the age of two,but meetings were not encouraged. Toldthat they were twins, after the girls dis-covered it for themselves, gravitated toone another at school. They play togetherquite a lot at school and during the eve-nings. Jessica often goes to tea with Wini-fred. They were never apart and wantedto sit at the same desk." This is a pair ofseparated twins, eight-years-old when,after this terrific separation, Shields dis-covered that they resembled one another.

Burton and Christopher were separatedat birth. "The paternal aunts decided totake one twin each and brought them upamiably livi,ng next door to one another

from teachers further assistance. The datahave been secured in this way and havingsatisfied ourselves that by these means wecan reduce the disturbing effects of en-vironment to relatively slight proportions,we have gone on ... ," etc.

So what Prof. Burt is doing is de-liberately removing from his numbers"the disturbing effect of environment."Because he is interested only in the al-leged effect of the genes, he removes thedisturbing effect of environment on thetest scores by adjusting the scores interms of comments and criticisms madeby the teachers. What's very peculiarabout this is that these adjusted scoreswere then mailed by Prof. Burt to, amongothers, Prof. Jensen and Prof. Shockleyin the United States, who had a greatdeal of fun with those numbers. Prof.Jensen has used those numbers fromwhich the disturbing effects of environ-mental influence had been systematicallyremoved to estimate'the proportion ofvariance in intelligence attributable toheredity vs. environment. It is a kind ofcomedy, but the possible effects on socialpolicy are not so funny.

Prof. Burt was not always so confidentin the teacher as the best estimator of thegenetic portion of intelligence. Burt saysin 1956 that the test is far less trust-worthy than the teacher. Thirteen yearsearlier he had written, "But in regard toinnate general ability there can be noquestion. The unaided judgments even ofthe most experienced teachers, shrewd asthey are in many cases, are neverthelessfar less trustworthy in the long run thanresults obtained with properly adminis-tered intelligence tests." This is in abso-lute flat contradiction to what he thenwrote 15 years later.

4 SOUTH TODAY

dently in 1970, talking about Burt's data."their IQ's wee obtained from an indi-vidual test, the English adaptation of theStanford-Binet." They were not, by theway.

Burt in 1966 had said the following:"The tests employed have been fully de-scribed elsewhere. They consisted of ( I )a group test of intelligence containingboth non-verbal and verbal items: (2) anindividual test, the London revision of theTerman-Binet scale, used primarily forstandardization, and for doubtful cases,(3) a set of performance tests standard-ized by Miss Gaw in 1925." Of coursethe test results were submitted to teach-ers, and so on. Now the interesting thingis that in 1958 Prof. Burt had repliedto an earlier critic and said the follow-ing about the "individual test" used onthe twins: "The figures he quotes frommy 'wn research were based on a non-verh_LI test of performance type."

So in 1958, Prof. Burt blandly informsus that the individual test that he usedwith the twins was a non-verbal test ofthe performance type. In 1966, heblandly informs us that it was the Londonrevision of the Terman-Binet, which is averbal test not of the performance type.Another flat contradiction.

We simply cannot discover from Prof.Burt what tests were given by whom towhom and under what circumstances. Hedoes, as you see, refer us to other sourcesto find out about this marvelous grouptest of intelligence which he's given. Hegives two sources, and in one source thereare no group intelligence tests at all. Inthe other source there are no fewer thanseven group intelligence tests which hehad presented. We can't figure out which

tne same age. And not because theyhave the same genes. Unless one is as-sured that the test which one is using inthese studies does not give differentscores for different ages, or for differentsexes, the studies simply hold no water.I have looked imo Miss Gaw's perfor-mance test, standardized in 1925. Sheinforms us that there are "striking" dif-ferences between the sexes in the perfor-mance test that Prof. Burt used.

Recall that Burt's is the only studythat gives us data on the social and eco-nomic status of the homes in whichseparated twins were placed. By 1958Burt had reported on 42 pairs of sepa-rated twins. In 1958 we were told thatat least four children of high professionalfamilies"class I" in his study hadbeen reared in orphanages. But in 1966,reporting on 53 pairs, zero children ofclass I professionals had been rearedin orphanages and only two of clericalworkers, the next class down, had beenreared in orphanages. How did four outof 42 become zero out of 53? Flat con-tradiction. Furthermore, Prof. Burt after1966 mailed this marvelous data on socialclass to Prof. Jensen. Due to the kindnessof Prof. Jensen I have a copy of the datawhich Prof. Burt mailed. Those data inturn contradict 1966 data, which in turncontradict the 1958 data.

If you look for mathematical impossi-bilities in Burt's work they assail youfrom every side. I really have to con-clude, and I think I have presentedenough evidence so that most reasonablepeople can follow me, that one simplycannot take Prof. Burt's numbers serious-ly.

I will say one thing for Prof. Burt: Hewas much more conscious and aware

at infewtract'butthatcoveone ;quiteningsfred.to Sitseparaftercove

Buat bitakealmain t

Theyotherfairtwinsradicwhat

Onthatstandwhiclteststest.aboutof twher twhicShiel(undergiveSo th

Hescore.Shiel

ie data

ins weof en-rtions,

is de-imbersent."he al-es thein theres :nrr.de

eculiarscores

amongockley

greatProf.from

nviron-aticallytion ofble toind ofsocial

nfidentof the

irt saystrust-years

Bard tobe no

;yen ofewd asrthelessin thandminis-n abso-le then

dently in 1970, talking about Burt's data,"their IQ's were obtained from an indi-vidual test, the English adaptation of theStanford-Binet." They were not, by theway.

Burt in 4966 had said the following:"The tests employed have been fully de-scribed elsewhere. They consisted of (1)a group test of intelligence containingboth non-verbal and verbal items; (2) anindividual test, the London revision of theTerman-Binet scale, used primarily forstandardization, and for doubtful cases,(3) a set of performance tests standard-ized by Miss Gaw in 1925." Of coursethe test results were submitted to teach-ers, and so on. Now the interesting thingis that in 1958 Prof. Burt had repliedto an earlier critic and said the follow-ing about the "individual test" used onthe twins: "The figures he quotes frommy own research were based on a non-verbal test of performance type."

So in 1958, Prof. Burt blandly informsus that the individual test that he usedwith the twins was a non-verbal test ofthe performance type. In 1966, heblandly informs us that it was the Londonrevision of the Terman-Binet, which is averbal test ,e,ot of the performance type.Another flat contradiction.

We simply cannot discover from Prof.Burt what tests were given by whom towhom and under what circumstances. Hedoes, as you see, refer us to other sourcesto find out about this marvelous grouptest of intelligence which he's given. Hegives two sources, and in one source thereare no group intelligence tests at all. Inthe other source there are no fewer thanseven group intelligence tests which hehad presented. We can't figure out which

the same age. And not because theyhave the same genes. Unless one is as-sured that the test which one is using inthese studies does not give differentscores for different ages, or for differentsexes, the studies simply hold no water.I have looked into Miss Gaw's perfor-mance test, standardized in 1925. Sheinforms us that there are "striking" dif-ferences between the sexes in the perfor-mance test that Prof. Burt used.

Recall that Burt's is the only studythat gives us data on the social and eco-nomic status of the homes in whichseparated twins were placed. By 1958Burt had reported on 42 pairs of sepa-rated twins. In 1958 we were told thatat least four children of high professionalfamilies"class 1" in his study -- hadbeen reared in orphanages. But in 1966,reporting on 53 pairs, zero children ofclass 1 professionals had been rearedin orphanages and only two of clericalworkers, the next class down, had beenreared in orphanages. How did four outof 42 become zero out of 53? Flat con-tradiction. Furthermore, Prof. Burt after1966 mailed this marvelous data on socialclass to Prof. Jensen. Due to the kindnessof Prof. Jensen I have a copy of the datawhich Prof. Burt mailed. Those data inturn contradict 1966 data, which in turncontradict the 1958 data.

If you look for mathematical impossi-bilities in Burt's work they assail youfrom every side. I really have to con-chide, and I think I have presentedenough evidence so that most reasonablepeople can follow me, that one simplycannot take Prof. Burt's numbers serious-ly.

will say one thing for Prof, Burt: Hewas much more conscious arid aware

at uiree montns. tirougnt up mum' afew hundred yards of one another. At-tracted to each other at the age of two,but meetings were not encouraged. Toldthat they were twins, after the girls dis-covered it for themselves, gravitated toone another at school. They play togetherquite a lot at school and during the eve-nings. Jessica often goes to tea with Wini-fred. They were never apart and wantedto sit at the same desk." This is a pair ofseparated twins, eight-years-old when,after this terrific separation, Shields dis-covered that they resembled one another.

Burton and Christopher were separatedat birth. "The paternal aunts decided totake one twin each and brought them upamiably living next door to one anotherin the same midlands collier!, village.They were constantly in and out of eachothers' houses." That gives you a not un-fair picture of what's meant by separatedtwins. Separation doesn't seem terriblyradical in these cases: I'm sure its notwhat you had visualized.

One can also see in Shields's appendixthat sometimes psychologists, for under-standable reasons, have to make decisionswhich may affect their data. One of thetests that Shields used was the Dominoestest. The average score on this test isabout 27. So he tells us about one pairof twins: Valerie got a score of 27, buther twin, Joyce, got a score of only onewhich was a fantastically low score. SoShields writes, "Joyce evidently did notunderstand the instructions; she did notgive an impression of low intelligence."So the data for this pair were discarded.

Here we have a case where one twinscores 27, the other scores one, andShields decides that the twin who scored

To page 5 of SOUTH TODAY(following Annual Report)

The Conclusions Which Have Been Drawn BPsychologists From These Data Do Not Stan

From page 4one didn't understand the instructions.But consider what happens in anothertwin case, Olwen and Gwladys. Olwenhad a score of four, and Gwladys had ascore of two. As it happens, Olwen andGwladys were Welsh. English was nottheir native language. But Shields writes,"Both twins are naturally Welsh speak-ing, but this does not explain their lowscore on the Dominos, which is in keepingwith the impression on interview of lowintelligence." So if you have two scoreswhich are both abominably low, evenwhen you know you are speaking to twinsnot in their native language, you concludethat they are both stupid and you keeptheir data. But if one scores high, and theother scores abominably low, then youuse your impression to say oh, gee, onedidn't understand the instructions.

If you look at the 40 pairs of separatedtwins who had been given the Dominoestest, for 35 pairs it is the case that Shieldshimself tested each member. But for theother five pairs, as it happens, two dif-ferent psychologists, usually one of thembeing Shields, tested the two members ofthe twin set. Suppose we now ask, whatwas the average difference in score forthose twins, both of whom were testedby Shields, compared with the averagedifference in score for those twins, eachof whom was tested by a different psy-chologist? For twins both tested byShields, the average difference is only4.9. But when different psychologiststested the twins, the average differenceis 13.2, and if one analyzes that differ-ence, it turns out to be statistically signifi-cant.

What that tells us is that the degree towhich identical twins resemble oneanother in the intelligence test dependsupon whether the same psychologist is

testing both twins or not. We know thatintelligence tests are not precise instru-mcnts which give mechanical scores. Apsychologist would be less than human ifhe is studying a pair of twins, and believesthat twins ought to resemble one another,if he didn't exhibit this kind of biasingeffect. But in all of our studies of identi-cal twins the elementary precaution ofthe test being given by different testers tothe two members of the twin pair has

markedly? Do people always have troubletelling you apart? Do you regard your-selves as being very much alike? So theyonly included in their study twins whosaid they were alike.

They used the Stanford-Binet Intelli-gence Test. They had in their study agr. ip of 100 individuals who were non-separated identical twins, brought up inthe normal family fashion. Their mono-graph indicates that the correlation be-tween age and IQ for non-separatedidentical twins is.49. What that says is,the older the child is, the stupider he is!Now that's obviously nonsense. What itmeans is that the test had not beenproperly standardized so as to give amean IQ of 100 at all ages.

If you now look at the separatedidentical twins, as it turns out there are19 pairs of separated twins and sevenwere males, so those seven male pairsgive us 14 individuals. And we can ask,what's the correlation for those 14 malesbetween IQ and age? The answer is .78,an enormously significant correlation.The older you are, the stupider you are.

Now it looks as if it's different forwomen. For if one takes the remaining12 pairs of separated twins that werefemales, the correlation between age andIQ was only.28. But there's a joker hid-den there. As in all of these studies, therewas a surplus of elderly women. When-ever you try to get volunteer twins, youget more women than men, and in turnyou get a disproportion of elderly wo-men. Therefore, if one discards the olderwomen and looks just at the eight sepa-rated females within the same age rangeas the 14 separated males, for them thecorrelation between age and IQ was .85.Not a bad match to the .78 for themales. The point of this is the apparentresemblance between the separated twinsin IQ may be due to nothing more thanthe fact that the twins are of the sameage. Indeed, by some fairly complicatedstatistical procedures, this point can bemade quite elegantly. Notice that you canpredict a person's IQ in this study betterby knowing his age than by knowing theIQ of his identical twin!

You have to look at the apparentrPeamt-danena csf twinA in-10. and take out

than Danish men; it means that the testis not standardized.

What about age? In Denmark, it turnsout if you look at her 18 female separatedtwins, the correlation between age andIQ on the Wechsler is +.60, which isstatistically significant. Whereas for hersix males it is .82, which is also statisti-cally significant, but in the other direc-tion. So if one looks at the correlationcoefficients reported by Juel-Nielsen,without taking age ir. o account, it looksas if among the nine female pairs oftwins the IQ correlation is .59. But whenone takes out of that statistically, as onemust, the effect of age, there is essentiallyno IQ correlation at all between the sepa-rated twins. This is with a different testthan that used by Newman, Freeman,and Holzinger, but the conclusions arevery much the same.

IQ Heredify ConclusionsUnsupported

There are some obvious problems withthe standardization of the IQ tests used inthe twin studies, which gravely affecttheir interpretation. The conclusionswhich many psychologists have drawnfrom these data do not, in my opinion,stand up.

It would be interesting to make thesame kind of analysis of Prof. Burt's data,but that data, of course, has never beenmade available. We know nothing at allabout the sex and ages of his twins. Andwe know that the data he reported abouttheir socio-economic class was inconsis-tent and unreliable.

I am sorry that I have not been ableto speak this morning about other kindsof IQ data used by some psychologiststo support the idea that IQ tests scoresare inheritable. I believe that these otherkinds of data stand up even less well thanthe twins data, and I hope in the nearfuture to have a detailed written analysisavailable.

Answers to QuestionsFrom Audience

kids acrolooked atthis expernew test, t

a small bdifference.males didor vice-yecause it dithere shouligence. SoThey wendividual itekind of tofollowingitem to fisix-year-olproportionought toof six-yeasix-year-olof five-yeto look atwhether theach sex.items ofmales, thehe was towasn't truesome iteperfectly rthose itemout of theany sex dtween mal

On thesocial claout that tple had athe childrIQ's of abto take thsee whethefor the chithe childredidn't occfound reasdid not diswould notas an intelli

A: If you mean by "intelligence" what- What wever intelligence Jests happen to measure, tpst

sions Which Have Been Drawn By SomeIs From These Data Do Not Stand Up ---

ons.etherwenid aandnot

rites,eak-lowpinglow

;ores.2venwinstide

,ecpi theyouone

atednoesoldsthedif-

hems ofvhat

forsted-ageach)sy-by

fitlyistsncefer-ifi-

to,mendst isthattru-. An ifvesher,singcnti-

ofN to

1-Ina

markedly? Do people always have troubletelling you apart? Do you regard your-selves as being very much alike? So theyonly included in their study twins whosaid they were alike.

They used the Stanford-Binet Intelli-gence Test. They had in their study rgroup of 100 individuals who were non-separated identical twins, brought up inthe normal family fashion. Their mono-graph indicates that the correlation be-tween age and IQ for non-separatedidentical twins is.49. What that says is,the older the child is, the stupider he is!Now that's obviously nonsense. What itmeans is that the test had not beenproperly standardized so as to give amean IQ of 100 at all ages.

If you now look at the separatedidentical twins, as it turns out there are19 pairs of separated twins and sevenwere males, so those seven male pairsgive us 14 individuals. And we can ask,what's the correlation for those 14 malesbetween IQ and age? The answer is .78,an enormously significant correlation.The older you are, the stupider you are.

Now it looks as if it's different forwomen. For if one takes the remaining12 pairs of separated twins that werefemales, the correlation between age andIQ was only.28. But there's a joker hid-den there. As in all of these studies, therewas a surplus of elderly women. When-ever you try to get volunteer twins, youget more women than men, and in turnyou get a disproportion of elderly wo-men. Therefore, if one discards the olderwomen and looks just at the eight sepa-rated females within the same age rangeas the 14 separated males, for them thecorrelation between age and IQ was .85.Not a bad match to the .78 for themales. The point of this is the apparentresemblance between the separated twinsin IQ may be due to nothing more thanthe fact that the twins are of the sameage. Indeed, by some fairly complicatedstatistical procedures, this point can bemade quite elegantly. Notice that you canpredict a person's IQ in this study betterby knowing his age than by knowing theIQ of his identical twin!

You have to look at the apparent

than Danish men; it means that the testis not standardized.

What about age? In Denmark, it turnsout if you look at her 18 female separatedtwins, the correlation between age andIQ on the Wechsler is +.60, which isstatistically significant. Whereas for hersix males it is .82, which is also statisti-cally significant, but in the other direc-tion. So if one looks at the correlationcoefficients reported by Juel-Nielsen,without taking age into account, it looksas if among the nine female pairs oftwins the IQ correlation is .59. But whenone takes out of that statistically, as onemust, the effect of age, there is essentiallyno IQ correlation at all between the sepa-rated twins. This is with a different testthan that used by Newman, Freeman,and Holzinger, but the conclusions arevery much the same.

IQ Heredity ConclusionsUnsupported

There are some obvious problems withthe standardization of the IQ tests used inthe twin studies, which gravely affecttheir interpretation. The conclusionswhich many psychologists have drawnfrom these data do not, in my opinion,stand up.

It would be interesting to make thesame kind of analysis of Prof. Burt's data,but that data, of course, has never beenmade available. We know nothing at allabout the sex and ages of his twins. Andwe know that the data he reported abouttheir socio-economic class was inconsis-tent and unreliable.



kids across the country. When theylooked at the results that they got withthis experimental pilot version of thenew test, they discovered that there wasa small but statistically significant sexdifference. In their sample, either themales did a little better than the females,or vice-versa. That disturbed them be-cause it didn't make sense to them thatthere should be a sex difference in intel-ligence. So they told us what they did.They went back and looked at the in-dividual items in the Stanford-Binet. Thatkind of test for children is built on thefollowing assumption. If you give a testitem to five-year-old kids, and then tosix-year-old kids, and seven-year-olds, theproportion of seven-year-olds passing itought to be greater than the proportionof six-year-olds. And the proportion ofsix-year-olds ought to be greater than thatof five-year-olds. So, what they did wasto look at the individual items and seewhether they were equally age-graded foreach sex. They found some individualitems of which it was true that, withinmales, the older the kid the more likelyhe was to give the right answer. Thatwasn't true with females. And they foundsome items vice-versa. So they, then,perfectly rationally, threw out of the testthose items. When they threw those itemsout of the test, then there was no longerany sex difference in the test scores be-tween males and females.

On the next page, they tell us aboutsocial class differences. Thus it turnedout that the children of professional peo-ple had average IQ's of about 125. Andthe children of day laborers had averageIQ's of about 90. It did not occur to themto take the individual items and look tosee whether they were equally age-gradedfor the children of professionals and forthe children of the day laborers. And itdidn't occur to them for a simple, pro-found reason. An intelligence test whichdid not discriminate between social classeswould not be accepted by psychologistsas an intelligence test.

A: If you mean by "intelligence" what- What we mean by an intelligencf,1 testver intelli en ,;igigAr erip6)

ence, it turns out to be statistically signifi-cant.

What that tells us is that the degree towhich identical twins resemble oneanother in the intelligence test dependsupon whether the same psychologist istesting both twins or not. We know thatintelligence tests are not precise instru-ments which give mechanical scores. Apsychologist would be less than human ifhe is studying a pair of twins, and believesthat twins ought to resemble one another,if he didn't exhibit this kind of biasingeffect. But in all of our studies of identi-cal twins the elementary precaution ofthe test being given by different testers tothe two members of the twin pair hasnever been taken.

The third study, which is the oneAmerican study in the series, was byNewman, Freeman and Holzinger. Insome ways I think its the best of the fouravailable studies. For one thing, weknow about how they got their twins.They were in Chicago and it was the De-pression; they conducted a radio andnewspaper campaign asking for separatedtwins to volunteer in the name of science.Separated identical twins. They did nothave much money with which to do thisresearch. Sometimes they got letters fromBoston, sometimes from San Francisco.They did not want to take the chance ofpaying the transportation expenses of apossible pair of subjects to Chicago onlyto discover when they arrived and wereblood tested that they were not reallyidentical twins. What they did, as theyvery candidly and honestly reported, wasthis: Whenever people volunteered theywere sent a letter and a questionnaireasking do you really resemble one another

men. Ineretore, it one cusearus bit owe,women and looks just at the eight sepa-rated females within the same age rangeas the 14 separated males, for them thecorrelation between age and IQ was .85.Not a bad match to the .78 for themales. The point of this is the apparentresemblance between the separated twinsin IQ may be due to nothing more thanthe fact that the twins are of the sameage. Indeed, by some fairly complicatedstatistical procedures, this point can bemade quite elegantly. Notice that you canpredict a person's IQ in this study betterby knowing his age than by knowing theIQ of his identical twin!

You have to look at the apparentresemblance of twins in IQ, and take outof that apparent resemblance the effectof age, because two people of the sameage will have similar IQ's. That's the waythese IQ tests are built.

Juel-Nielsen's is the fourth separatedtwin study. Prof. Jensen in his reviewarticle informs us that the Wechsler,which is the test Juel-Nielsen used, hadbeen standardized on the Danish popula-tion to give a mean of 100 and a standarddeviation of 15. The problem is whenyou read Juel-Nielsen, she apologizes forthe fact that there has never been aDanish standardization of the Wechslertest and therefore some of her data maybe questionable in terms of age and sexof her subjects.

If you look at her data, it turns out forexample that Danish men are brighterthan Danish women. For 18 females inher sample, the mean IQ was 102; for sixmales it was 113. That is a statisticallysignificant difference. It of course doesn'tmean that Danish women are brighter

we Know that the data he reported abouttheir socio-economic class was inconsis-tent and unreliable.

I am sorry that I have not been ableto speak this morning about other kindsof IQ data used l'y some psychologiststo support the idea that IQ tests scoresare inheritable. I believe that these otherkinds of data stand up even less well thanthe twins data, and I hope in the nearfuture to have a detailed written analysisavailable.

Answers to Question:From Audience

A: If you mean by "intelligence" what-ever intelligence tests happen to measure,by that definition, I am saying it does notappear to be that genetics are involved inthat kind of intelligence. That dies notdeny the possibility that there are somecentral cognitive capacities and abilitieswhich we have not yet been clever enoughto measure which, if we could measure,might turn out to be largely geneticallydetermined. They might also turn out tobe not genetically determined. What I amsaying is that intelligence tests scores donot appear to be the least bit related togenetic factors. That doesn't say neces-sarily that it might not be possible tosome day measure people as being"bright" and others "dull" and findingout that indeed this is genetically deter-mined. The present IQ tests don't do that.I really think that d stinction is important.

A: Let me cite examples from twoconsecutive pages from the book byTerman and Merrill in 1937, which pre-sented the standardization of the Stan-ford-Binet test. What they had done wasto give on an experimental basis the testto a reasonably large sample of school

social classout that thepie had averthe childrenIQ's of abouto take the isee whetherfor the childthe childrendidn't ocr':rfonnzi reasoidid not disci.'would notas an intellig

What weis a test onsity professwill score hneighbors. Imales, sometbecause wewhat "intelliwell as profwhites, theretest. The teplacement ocupational sto use the tesomebody wsuccessfula good test.

A: Savincause I wanto talk abhave been ubeen somethat they hapolicies. I wplications onday.

is

is

at13-

Aif

es

r.ngti-oftoas

ne

byInur.ve

)e-

lided

.e.

,othis'pm

co.ofa

lilyere

lly

icyas

leyire

her

men. l nerelore, it one oiscalus Inc (noelwomen and looks just at the eight sepa-rated females within the same age rangeas the 14 separated males, for them thecorrelation between age and IQ was .85.Not a bad match to the .78 for themales. The point of this is the apparentresemblance between the separated twinsin IQ may be due to nothing more thanthe fact that the twins are of the sameage. Indeed, by some fairly complicatedstatistical procedures, this point can bemade quite elegantly. Notice that you canpredict a person's IQ in this study betterby knowing his age than by knowing theIQ of his identical twin!

You have to look at the apparentresemblance of twins in IQ, and take outof that apparent resemblance the effectof age, because two people of the sameage will have similar IQ's. That's the waythese IC tests are built.

Juel-NielFen's is the fourth separatedtwin study. Pt...f. Jensen in his reviewarticle informs us that the Wechsler,which is the test Juel-Nielsen used, hadbeen standardized on the Danish popula-tion to give a mean of 100 and a standarddeviation of 15. The problem is whenyou read Juel-Nielsen, she apologizes forthe fact that there has never been aDanish standardization of the Wechslertest and therefore some of her data maybe questionable in terms of age and sexof her subjects.

If you look at her data, it turns out forexample that Danish men are brighterthan Danish women. For 18 females inher sample, the mean IQ was 102; for sixmales it was 113. That is a statisticallysignificant difference. It of course doesn'tmean that Danish women are brighter

we Know that the data he reported abouttheir socio-economic class was inconsis-tent and unreliable.



A: If you mean by "intelligence" what-ever intelligence tests happen to measure,by that definition, I am saying it does notappear to be that genetics are involved inthat kind of intelligence. That does notdeny the possibility that there are somecentral cognitive capacities and abilitieswhich we have not yet been clever enoughto measure which, if we could measure,might turn out to be largely geneticallydetermined. They might also turn out tobe not genetically determined. What I amsaying is that intelligence tests scores donot appear to be the least bit related togenetic factors. That doesn't say neces-sarily that it might not be possible tosome day measure people as being"bright" and others "dull" and findingout that indeed this is genetically deter-mined. The present IQ tests don't do that.I really think that distinction is important.

A: Let me cite examples from twoconsecutive pages from the book byTerman and Merrill in 1937, which pre-sented the standardization of the Stan-ford-Binet test. What they had done wasto give on an experimental basis the testto a reasonably large sample of school

Vss

social class differences. Thus it turnedout that the children of professional peo-ple had average IQ's of about 125. Andthe children of day laborers had averageIQ's of about 90. It did not occur to themto take the individual items and look tosee whether they were equally age-gradedfor the children of professionals and forthe children of the day laborers. And itdidn't occur to them for a simple, pro-found reason. An intelligence test whichdid not discriminate between social classeswould not be accepted by psychologistsas an intelligence test.

What we mean by an intelligence testis a test on which the children of univer-sity professors, doctors, lawyers, etc.,will score higher ths.n :bildren of theirneighbors. If males sere higher than fe-males, something is wrung with our testsbecause we have pre-conceived ideas ofwhat "intelligence" is. If workers score aswell as professors, or blacks as well aswhites, there's something wrong with thetest. The tests are used to predict theplacement of people in a particular oc-cupational structure. And if you wantto use the test as a prediction of whethersomebody will grow up to be a wealthy,successful white middle-class male, itsa good test. It predicts that pretty well.

A: Saving that for this afternoon, because I want to do precisely that. I wantto talk about how intelligence testshave been used in the past. There havebeen some incredible ways, to my mind,that they have been used to affect socialpolicies. I will point to some of the im-plications on how they are being used to-day.

SOUTH TODAY 5

How TestingHarms Children

Grouping According TDiscriminatory--No

By Winifred

Following is the text of a presentationat the symposium on Human Intelligence..Social Science and Social Policy hr Ills.Green. director of the Atlanta-basedSoutheastern Public Education Programof the American Friends .Serrice Com-mittee.

Sgt. Moses Williams, a member of thepolice force in Tallulah, La.. who is

a hlack parent long active in the strugglefor civil rights, told me two weeks ago:"Giving those tests and then placing chil-dren in groups is for one purpose and onepurpose only and that is to segregate.Once you put a child in that low groupthere is no way for him to advance him-self, You know some of our children mayhe a little slow, but they all can learn."

And two days ago. talking to a middle-class white mother in a suburban schooldistrict in the urban South, I was told:"You know Winifred, John is in hiseleventh year of school and just this year.hecause he has two teachers who haveconfidence in him, is he heginning to getout of the hox he was put in in elementaryschool because he was tracked." She wenton to say that she had no worries abouther own child hecause of the support hereceived at home and that he (her child)would he ok. But what, she asked, aboutthe ones who don't get the support athome?

'Some Are Better . .

The effects of testing and the inevi-table result of classification cut across raceand class. It affects all our children, black.white, middle-class, rich and poor. Ofcourse it is the minorities and the poorwho suffer most from this device whichsays, no matter how sophisticated thelanguage. SOME ARE BETTER THANOTHERS.

would like to spend a few minutesthis afternoon telling you about some ofthe children, the human beings. withwhom we have worked and how the cur-rent definition of intelligence and theuse of testing have affected their livesand then to suggest some of the ways Ithink private groups (civic cluhs, founda-tions, churches and others) can he in-volved in solving the problem.

We are a nation that gives lip-serviceto how much we value our children. Howoften have you heard that children areour nation's greatest resource? If weexamine the facts, the statistics, and thechildren that we know who together m.ikeun these stotistics, it is not true. We are

_10.14k,,Children of All Races, Classes Penalized

mapannrinummil.

AA--

George L. Walker Ill

well qualified to talk about what I haveseen as .the results in many Southernschool systems and to tell you what thestudents themselves, the so-called "testusers." say ahout them. Conversationswith students in special education classesin four Georgia school districts that range

Q. Why were you assigned to SpecialEducation classes?

A. They gave all or us a test andthe dumb students was put in these kindof classes.

I ain't dumb! They gave me a test butthat don't mean nothing. I just mark them

the

SCI

hi:ratLitt

galS.

disthcde!thc

darat.

for

the

onmimestugwint

indehia lstugr:ortivth

ni7It'sHi

edmichitea

priifgras.

ledo

thw

sttin

tingildren

stationnee.

by Ms.a-basedrogram

C0171-

of thewho is,truggle:s ago:

chil-nd oneregatc.group

e him -cn

middleschool

ts told:in his

is year.o have

to getientarylc wents aboutport hechild)about

port at

e inevi-oss race1. black.)or. Ofis poor

whichted theTI-IAN

min utesome of

s. withhe cur-ed theit livesways I

unda-he in-

-servicen. Howten areIf we

ind ther maker

s

,

4).

.r+

Grouping According To Tests Scores IsDiscriminatory--No Matter How It's Done

By Winifred Green

George L. Walker illChildren of All Races, Classes Penalized

4 :" j

V 41" :

well qualified to talk about what I haveseen as the results in many Southernschool systems and to tell you what thestudents themselves, the so-called "testusers," say about them. Conversationswith students in special education classes

cur or . school ist icts hat an !e

4-


A. They gave all of us a test andthe dumb students was put in these kindof classes.

I ain't dumb! They gave me a test butat 't can nothin I *ust mark them

than, the days of separate but unequalschools, What we have now for manyblack, Chicano and poor children is sepa-rate and very unequal classrooms withlittle if any hope of escape.

Grouping In AikenAFSC provided the resources for liti-

gation against ability grouping in Aiken.S. C., in a suit now pending in federaldistrict court. Aiken illustrates many ofthe absurdities of grouping, as well as itsdesign to keep black and poor children intheir place.

Grouping in Aiken is done by stan-dardized test grades and teacher recom-mendations. Two teachers reported to at-torneys that they had been informed bytheir department heads at different sec-ondary schools to give their low groupsminimum scores on the teacher recom-mendation form, and to give high-groupstudents maximum scores. This couldguarantee that a student would be lockedinto his group, whatever its level.

Another insidious aspect of groupingin Aiken is the altering of grades on stu-dents' permanent records. Students inhigh sections have their averages raiseda letter grade on their permanent record;students in average sections retain thegrades they make, and low group studentsare lowered a grade. Thus no one is mo-tivated: The high groups can't lose, andthe low groups can't win.

In case students in Aiken can't recog-nize themselves as dummies, or geniuses,it's marked for them on their report cards.High group students' cards are clearlymarked "plus," and low groups are mark-ed "minus." Thus the terms plus andminus become common references tochildren in those groups among bothteachers and students.

Elementary students are not allowed theprivilege of being smart in language artsif they are dumb in arithmetic. They aregrouped at the same level in all subjects,assuring that their real skills go unchal-lenged at a critical period in their educa-tion.

During the first year of desegregation,the lower groups at one junior high schoolwere segregated by sex.

Finally, there is no policy for movingstudents upward, and little evidence thatin fact there ever is any upward move-ment.

Grouping Is DiscriminatoryI have never seen trouping operate

I r-tr.11.3.I would like to spend a few minutes

this afternoon telli g you about some ofthe children, the human beings. withwhom we have worked and how the cur-rent definition of intelligence and theuse of testing have affected their livesand then to suggest some of the ways Ithink private groups (civic clubs, founda-tions. churches and others) can he in-volved in solving the problem.

We are a nation that gives lip-serviceto how much we value our children. Howoften have you heard that children areour nation's greatest resource? If weexamine the facts, the statistics, and thechildren that we know who together makeup these statistics, it is not true. We arenot a child-oriented nation. Children arethe unrecognized, neglected and mis-treated minority in our country. We needto start the process of becoming awareof how our children are mishandled.

The long struggle for equal educationalopportunity has led those concerned withchange in this country. and in our regionparticularly, to start to examine what wemean when we speak of quality educa-tion. A staff member of the AFSC oncedefined quality education as the processthat "gives the student the basic skillswhich he needs for the great range ofopportunities before him.... It (a qualityeducation) equips him with the tools sothat he . . not our society, can deter-mine what he wants to do with his lifeand will be well prepared to take the nextstep toward that goal. . .

What Students SayAre testing and classification proce-

dures consistent with the aim of increas-ing these options for children? While I amnot qualified to speak of the validity, orthe fairness of testing, I am, I believe.

6 SOUTH TODAY

"Ayr

,

well qualified to talk about what I haveseen as the results in many Southernschool systems and to tell you what thestudents themselves, the so-called "testusers," say about them. Conversation'with students in special education classein four Georgia school districts that rangefrom 40 to 65 percent black indicar.dover and over that they, the students, f. Itthat they would function better in hek7o-genous classesor as they put it, 1, ',*

other students not on their "level."Here are a few of their comments:Q. Why were you placed in Sreca

Education?A. I took a test and she (my teatime r)

said I did real bad and was assigner! othis class.

Q. Do you think you have impi..1 cosince being in this class?

A. No! They give us some qocialways to do our work but they only CY- In-

fuse me.Q. Have you told your teats: r. or

counselor how you feel?A. What good would that do? ','hey

don't care. All they gonna do is Vow /alk-ing that stuff about a test. . . . no-body care about us.

And here are comments tq otherstudents who seemed to fee; (7, _n morepowerless, one of whom :Ner viewedherself as inferior as a result -le tests:

Elemenprivilegeif they argroupedassuringlenged attion.

Dunnthe lowerwere seg

Ct., Way were you assigned to Special Finallylr c..rinn classes? students

.1. They gave all of us a test and in fact t

domb students was put in these kind ment.of . lasses.

; ain't dumb! They gave me a test butthat don't mean nothing. I just mark themlittle dots as soon as I can.

The third group of students seems to!eject its status in these classes. A senior

rural Georgia says:.`These classes are no good and a trick.They just trying to make us look dumb.Now when I go to class (math) we don'tdo that much of nothing. The teacherseems to think we are no good. She givesus a lot of what I call busy work. Inorder to force her to work with us we,some of us, bother her by going to herdesk for help. At the end of the period weplay games. Most of these games aresilly; so we do a lot of complaining. She'salways telling us we have a disciplineproblem."

I cannot be in support of testing andgrouping for students such as these be-cause our experience leads us to believethat grouping reinforces the effects ofyears of discriminatory treatment in theeducation of black childrenlockingthem into classroom situations where cur-riculum, materials, teacher expectationand the resulting stigmas and hopelessnessare the same as, or some predict worse

GroI hay

withoutcriminatorganizat

Schoolwe moniSouth, 3.60 perceroom segon the bfor theHigher tand hadpredomiers.

We amay callkids callones," oclass wh

Groupreally toknow onwhich tstereotyhere thadon't msocio-ec

itinutesome of r,. withhe cur-lid theit livesways I

toil nda-he in-

-service well qualified to talk about what I haven. How seen as the results in many Southernren are school systems and to tell you what the

If we students themselves, the so-called "testand the users," say about them. Conversationser make with students in special education classesWe are in four Georgia school districts that range

siren are from 40 to 65 percent black indicatedid mis- over and over that they, the students, felt,Ve need that they would function better in hetero-g aware genous classesor as they put it, withdled. other students not on their "level."icational Here are a few of their comments:ied withit regionwhat wev educaSC once

processsic skillsrange ofa qualitytools so

an deter-his life

the next

Q. Why were you placed in SpecialEducation?

A. 1 took a test and she (my teacher)said 1 did real bad and was assigned tothis class.

Q. Do you think you have improvedsince being in this class?

A. No! They give us some specialways to do our work but they only con-fuse me.

Q. Have you told your teacher, orcounselor how you feel?

A. What good would that do? Theydon't care. All they gonna do is start talk-

V ing that stuff about a test.. . . Don't no-n prose- body care about us.t increas- And here are comments of two otherihile I am students who seemed to fee! even moreilidity, or powerless, one of whom even viewedI believe, herself as inferior as a result of the tests:

,


A. They gave all of us a test andthe dumb students was put in these kindof classes.

I ain't dumb! They gave me a test butthat don't mean nothing. 1 just mark themlittle dots as soon as 1 can.

The third group of students seems toreject its status in these classes. A seniorin rural Georgia says:"These classes are no good and a trick.They just trying to make us look dumb.Now when I go to class (math) we don'tdo that much of nothing. The teacherseems to think we are no good. She givesus a lot of what I call busy work. Inorder to force her to work with us we,some of us, bother her by going to herdesk for help. At the end of the period weplay games. Most of these games aresilly; so we do a lot of complaining. She'salways telling us we have a disciplineproblem."

I cannot be in support of testing andgrouping for students such as these be-cause our experience leads us to believethat grouping reinforces the effects ofyears of discriminatory treatment in theeducation of black childrenlockingthem into classroom situations where cur-riculum, materials, teacher expectationand the resulting stigmas and hopelessnessare the same as, or some predict worse

Elementary students are not allowed theprivilege of being smart in language artsif they are dumb in arithmetic. They aregrouped at the same level in all subjects,assuring that their real skills go unchal-lenged at a critical period in their educa-tion.

During the first year of desegregation.the lower groups at one junior high schoolwere segregated by sex.

Finally, there is no policy for movingstudents upward, and little evidence thatin fact there ever is any upward move-ment.

Grouping Is DiscriminatoryI have never seen grouping operate

without students feeling that it was dis-criminatory. In a report we did with otherorganizations in 1970 ("The Status ofSchool Desegregation, 1970"), in whichwe monitored 467 school districts in theSouth, 35 percent of the high-school and60 percent of the elementary-school class-room segregation we found was defendedon the basis of tests, usually administeredfor the first time with desegregation.Higher tracks were predominantly whiteand had white teachers; lower tracks werepredominantly black and had black teach-ers.

We as parents, teachers and citizensmay call it "special education," but thekids call it "the dumb class," or "the slowones," or "those idiots in Mrs. Smith'sclass who can't read."

Grouping means that integration neverreally takes place. Students do not get toknow one another, and in the context inwhich they attend school together, oldstereotypes are reinforced. I might addhere that when I speak of integration, Idon't mean just black and white, butsocio-economic integration as well.

IQ Tests As Instruments of OpprFrom Immigration Quotas to W

On March 23, the Southern RegionalCouncil sponsored a one-day symposiumon the history of intelligence testing andits public policy implications. The prin-cipal presentation was made by Dr. LeonKamin, chairman of the Department ofPsychology at Princeton University.

During the morning session, Dr. Kaminpresented findings of his extensive re-search into the original studies on whichsome American social scientists havebased writings on the inheritability ofintelligence.

Dr. Kam ncluded that there is novalid evidence ..1 all to support the hered-ity assumption. A transcript of this por-tion of his presentation was published inthe May -June issue of SOUTH TODAY.

The following transcript is Dr. Kamin'safternoon presentation on the history ofthe use of testing in the United Statesto support repressive public policy.

he pioneers of the intelligence test-ing movement in the United States

were three men who, around 1910, im-ported from France and translated thefirst so-called intelligence test.

This test had been constructed by theFrench psychologist Alfred Binet, inter-estingly enough, for the purpose of detect-ing learning disabilities among schoolchildren, and not for the purpose ofmeasuring a person's fixed intelligencequotient.

The three American pioneers wereLewis Terman at Stanford, Henry God-dard at Princeton, and Robert Yerkes atHarvard.

These men had some pretty well de-fined ideas about society and about race.Whether or not these ideas influencedtheir interpretations of intelligence testscores I will leave for you to decide. I'dlike to read a few quotations from theirwritings.

Terman, in his book published in 1916which presented the famous Stanford -Binet test, described the poor test scoresof a couple of Portuguese children andthen wrote the following: "The dullnessseems to be racial, or at least inherentin the family stocks from which theycome. The fact that one meets this typewith such extraordinary frequency amongIndians, Mexicans and Negroes suggestsquite forcibly the whole question of racial

A Talk by DR. LEON KAMIN

for the majority of cases of chronic andsemi-chronic pauperism. The feeble-minded continue to multiply. Organizedcharities often contribute to the survivalof individuals who would otherwise notbe able to live and reproduce. Whencharity organizations help the feeble-minded float along in the social and in-dustrial world and to produce and rearchildren after their kind, a doubtful serv-ice is rendered. A little psychological re-search would aid the united charities ofany city to direct their expenditures intomore profitable channels. If we wouldpreserve our state for a class of peopleworthy to possess it, we must preventas far as possible the propagation ofmental degenerates. We must curtail theincreasing spawn of degeneracy." Youwill detect perhaps a certain social andpolitical ideology there.

Henry Goddard was invited to Prince-ton University in 1920 to give lectures.He had devised his own version of theBinet scale, and this is what he told thePrinceton audience in 1920, after firstexplaining that this marvelous new testmeasured genetically determined intelli-gence. "We, may perhaps be permittedto apply the principle to another problemthat looms up rather large at the presenttime, namely socialism. Especially its ex-treme form of Bolshevism. Most of thearguments used by the more intelligentmembers of these groups are fallacious,because they ignore the mental levels(meaning IQ]. These men in their ultrahumane attitudes, their desire to he fairto the workman, maintain that the great'nequalities of social life are wrong andunjust.

"For example, here is a man who says,I am wearing $12 shoes, there is a laborerwho is wearing $3 shoes. Why should Ispend $12 while he could only afford $3?I live in a home that is artisticallydecorated--carpets, high priced furniture,expensive pictures, etc., and the laborerlives in a hovel. As we have said theargument is fallacious. The fact is, thatworkman may have a 10 year intelligencewhile you have a 20. To demand forhim such a home as you enjoy is as

absurd as it would be to insist that everylaborer receive a graduate fellowship.How can there be such a thing as socialequality, with this wide range of mental

ogy and Prof. Terman's.Prof. Terman should not have been

so pessimistic as he was about the possi-bility of convincing society that those withlow IQ scores should not reproduce. Overabout a 10-year period about 35 statespassed compulsory sterilization laws, ap-plied only to people with the misfortuneto end up in state-supported institutions.

Robert Yerkes at Harvard entered intoa small argument with Terman. Termanhad pointed out that the virtue of the testwas that it provided a hard scientific num-ber which enabled one to decide who wasand who wasn't feeble-minded. Well,Yerks said no, the test score certainlyshould be respected but, "never should adiagnosis of feeble-mindedness be madeon the IQ alone. We must inquire furtherinto the subject's economic history. Whatis his occupation and his pay? We mustlearn what we can about his immediatefamily. What is the economic status andoccupation of the parents? When all thisinformation has been collected, the psy-chologist may be of great value in gettingthe subject into the most suitable place insociety." If his IQ was low enough, thatsuitable place was a public institutionwhere he might be sterilized.

History of Immigration LawsI think the first immediate major prac-

tical effect of intelligence testing was inan area which most people don't knowabout. I didn't know about it until a yearago. It turns out that intelligence testshad a great deal to do with the framingof the immigration laws of the UnitedStates. And I would like to run throughbriefly some of the history of that period.

As early as 1912 Prof, Goddard, thePrinceton coal miner speaker, was invitedto Ellis Island by the United States PublicHealth Service and invited to apply thenew science of intelligence testing toimmigrants from Europe. Goddard re-ported in a scientific paper that he hadgiven the intelligence test to a representa-tive sample of European immigrants. Hediscovered that 83 percent of the Jewstrying to enter the U. S. were feeble-minded, that 80 percent of the Hungar-ians were feeble-minded, that 79 percentof the Italians were feeble-minded, andthat 87 percent of the Russians were

chided,grants

Thenwind.manychologition csociatiimprovting offeeble -Journaogist nas 191headin"Recendevelothat brto theexcellequalitiexistenknownbe exptest fmeasu

Wo

Thelargerto grU. S.first tindivitestining pRobe

Inof pseastercalledthe Ain Nesisteddistinas ato psiful hi

Influxthe Eat thtremeperimin the

s As Instruments of Oppression--mmigration Quotas to Welfare

'gional)osiurtiig and

prin-. Leonhem ofty.Warnin've re-which

havelily of

is nohered-is por-lied inDAY.

rory ofStates

c test-States

U. im-ed the

ny theinter-

letect-

Ise ofigence

wereGod-

kcs at

II de-race.

fencedtest

e. I'dtheir

1916nford-sco resn andillnessiherenth theys typeamongggestsracial

A Talk by DR. LEON KAMIN

for the majority of cases of chronic and

semi-chronic pauperism. The feeble-minded continue to multiply. Organizedcharities often contribute to the survivalof individuals who would otherwise notbe able to live and reproduce. Whencharity organizations help the feeble-minded float along in the social and in-dustrial world and to produce and rearchildren after their kind, a doubtful serv-ice is rendered. A little psychological re-search would aid the united charities ofany city to direct their expenditures intomore profitable channels. If we wouldpreserve our state for a class of peopleworthy to possess it, we must preventas far as possible the propagation ofmental degenerates. We must curtail theincreasing spawn of degeneracy." Youwill detect perhaps a certain social andpolitical ideology there.

Henry Goddard was invited to Prince-ton University in 1920 to give lectures.He had devised his own version of theBinct scale, and this is what he told thePrinceton audience in 1920, after firstexplaining that this marvelous new testmeasured genetically determined intelli-gence. "We may perhaps be permittedto apply the principle to another problemthat looms up rather large at the presenttime, namely socialism. Especially its ex-treme form of Bolshevism. Most of thearguments used by the more intelligentmembers of these groups are fallacious,because they ignore the mental levels[meaning IQ]. These men in their ultrahumane attitudes, their desire to be fairto the workman, maintain that the greatinequalities of social life are wrong andunjust.

"For example, here is a man who says,I am wearing $12 shoes, there is a laborerwho is wearing $3 shoes. Why should Ispend $12 while he could only afford $3?I live in a home that is artisticallydecoratedcarpets, high priced furniture,expensive pictures, etc., and the laborerlives in a hovel. As we have said theargument is fallacious. The fact is, thatworkman may have a 10 year intelligencewhile you have a 20. To demand forhim such a home as you enjoy is asabsurd as it would be to insist that everylaborer receive a graduate fellowship.How can there be such a thing as socialequality, with this wide range of mental

ogy and Prof. Terman's.Prof. Terman should not have been

so pessimistic as he was about the possi-bility of convincing society that those withlow IQ scores should not reproduce. Overabout a 10-year period about 35 statespassed compulsory sterilization laws, ap-plied only to people with the misfortuneto end up in state-supported institutions.

Robert Yerkes at Harvard entered intoa small argument with Terman. Termanhad pointed out that the virtue of the testvas that it provided a hard scientific num-

ber which enabled one to decide who wasand who wasn't feeble-minded. Well,Yerks said no, the test score certainlyshould be respected but, "never should adiagnosis of feeble-mindedness be madeon the IQ alone. We must inquire furtherinto the subject's economic history. Whatis his occupation and his pay? We mustlearn what we can about his immediatefamily. What is the economic status andoccupation of the parents? When all thisinformation has been collected, the psy-chologist may be of great value in gettingthe subject into the most suitable place insociety." If his IQ was low enough, thatsuitable place was a public institutionwhere he might be sterilized.

History of Immigration Laws

I think the first immediate major prac-tical effect of intelligence testing was inan area which most people don't knowabout. I didn't know about it until a yearago. It turns out that intelligence testshad a great deal to do with the framingof the immigration laws of the UnitedStates. And I would like to run throughbriefly some of the history of that period.

As early as 1912 Prof. Goddard, thePrinceton coal miner speaker, was invitedto Ellis Island by the United States PublicHealth Service and invited to apply thenew science of intelligence testing toimmigrants from Europe. Goddard re-ported in a scientific paper that he hadgiven the intelligence test to a representa-tive sample of European immigrants. Hediscovered that 83 percent of the Jewstrying to enter the U. S. were feeble-minded, that 80 percent of the Hungar-ians were feeble-minded, that 79 percentof the Italians were feeble-minded, andthat 87 percent of the Russians were

chided, but the total number of immi-grants was still wide open.

Then one sees the first straw in thewind. Terman, Yerkes and Goddard andmany of the leading biologists and psy-chologists were members of an organiza-tion called the Eugenics Research As-sociation. They were concerned withimproving human breeding stock by cut-ting off the defective germ plasm of thefeeble-minded, etc. The editor of theirJournal, The Eugenical News, was a biol-ogist named Dr. Harry Laughlin. As earlyas 1917 he wrote the following under theheading "The New Immigration Law":"Recently the science of psychology hasdeveloped to . high stage of precisionthat branch of its general subject devotedto the testing of individuals for naturalexcellence in mental and temrramentalqualities. When the knowledge of theexistence of this science becomes generallyknown in Congress, that body will thenhe expected to apply the direct and logicaltest for the qualities which we seek tomeasure in immigrants."

World War I and IQ TestingThe small cloud on the horizon is no

larger than a man's fist, but it's goingto grow and grow. Later in 1917, theU. S. entered World War I and for thefirst time in history, literally millions ofindividuals were subjected to intelligencetesting. The head of the intelligence test-ing pogram in the Army was now Col.Robert M. Yerkes.

In 1918 a small distinguished groupof psychologists and biologists from theeastern seaboard founded somethingcalled the Galton Society which met inthe American Museum of Natural Historyin New York City once a month. It con-sisted at no one time of more than 25distinguished scientists. They provided,as a free service, to the governnient andto private organizations all kinds of help-ful hints about eugenical practices.

In 1920 there was an enormous in-flux of experimental psychologists intothe Eugenics Research Association. Allat the same time in a given year, atremendous number of distinguished ex-perimental psychologists and biologistsin the country joined the Eugenics Re-

.

fined ideas about society and about race.Whether or not these ideas influencedtheir interpretations of intelligence testscores I will leave for you to decide. I'dlike to read a few quotations from theirwritings.

Terman, in his hook published in 1916which presented the famous Stanford-Binet test, described the poor test scoresof a couple of Portuguese children andthen wrote the following: "The dullnessseems to he racial, or at least inherentin the family stocks from which theycome. The fact that one meets this typewith such extraordinary frequency amongIndians, Mexicans and Negroes suggestsquite forcibly the whole question of racialdifferences in mental traits will have tohe taken up anew and by experimentalmethods."

The writer predicts that when this is

done "there will be discovered enormouslysignificant racial differences in generalintelligence, differences which cannot bewiped out by any scheme of mental cul-ture. Children of this group should besegregated in special classes. They cannotmvster abstractions, but they can oftenbe made efficient workers. There is nopossibility at present of convincing societythat they should not be allowed to re-produce. They constitute a grave menacebecause of their unusually prolific breed-ing." This, from the founder of the Amer-ican intelligence testing movement.

"The Menace ofFeeble-Mindedness"

The next year writing under the head-ing "The Menace of Feeble-Mindness,"Prof. Terman wrote, "Only recently havewe begun to recognize how serious amenace it is to the social, economic andmoral welfare of the state. It is responsible

Illly 11111 WI .1,1,1.1 Ill, tlt, (MU

unjust."For example, here is a man who says,

I am wearing $12 shoes, there is a laborerwho is wearing $3 shoes. Why should Ispend $12 while he could only afford $3?I live in a home that is artisticallydecorated carpets, htith priced furniture,expensive p,ctures, etc., and the laborerlives in a hovel. As we have said theargument is fallacious. The fact is, thatworkman may have a 10 year intelligencewhile you have a 20. To demand forhim such a home as you enjoy is as

absurd as it would be to insist that everylaborer receive a graduate fellowship.How can there be such a thing as socialequality, with this wide range of mentalcapacity?

"As for an equal distribution of thewealth of the world, that is equally absurd.The man with intelligence has spent hismoney wisely and saved until he hasenough to provide for his needs in caseof sickness, while a man of low intelli-gence, no matter how much money hewould have earned, would have sontmuch of it foolishly and would neverhave anything ahead. It is said that dur-ing the past year the coal miners incertain parts of the country have earnedmore money than the coal operators. Yettoday when the mines shut down for atime, those people are the first to suffer.They did not save anything, although theirwhole life has taught them that mining isan irregular thing, and that when theywere having plenty of work, they shouldsave against the days when they do nothave work."

But low IQ coal miners, according toProf. Goddard, don't have enough intel-ligence to salt away all that extra moneythey're making, etc. You may detect asimilarity between Prof. Goddard's ideol-

b. Via.v. v s .,States. And I would like to run throughbriefly some of the history of that period.

As early as 1912 Prof. Goddard, thePrinceton coal miner speaker, was invitedto Ellis Island by the United States PublicHealth Service and invited to apply the,iew science of intelligence testing toimmigrants from Europe. Goddard re-ported in a scientific paper that he hadgiven the intelligence test to a representa-tive sample of European immigrants. Hediscovered that 83 percent of the Jewstrying to enter the U. S. were feeble-minded, that 80 percent of the Hungar-ians were feeble-minded, that 79 percentof the Italians were feeble- minded, andthat 87 percent of the Russians werefeeble-minded. Science and psychologyscore again. And he wrote in 1917, "Thenumber of aliens deported because offeeble- mindedness increased approximate-ly 350 percent in 1913 57 percentin 1914 over what it had been. This wasdue chiefly to the efforts of the physicianswho were inspired by the belief thatme 'tat tests could be used for the detec-tion of feeble-minded aliens."

Until 1875 there were no federal im-migration laws, whatsoever. Anybodywho wanted to came. In 1875, the firstfederal law was passed. It did not limitthe number of immigrants, but it did ex-chide certain classes of individuals. Andthe first exclusion in 1875 consisted ofthree types of peoplecoolies, convictsand foreign prostitutes. In 1882, lunaticsand idiots were added, and as one goesthrough the years more classes. So by1903 persons who had had two or moreattacks of insanity at any time previouslywere added. The point is that up until theoutbreak of World War I, there was inprinciple unlimited immigration. Certaindefective classes of individuals were ex-

ofcastcalltheinsiltdistas

toful

flu,theattreperinseaass

HebuticsImU.

Sciofwhallon

dethawh.althethistiotThedidatdrahadcou

ace.,cedtestI'dci r

916ord-aresandnessrenthey

vPeonggists

cialto

ntal

s isuslyoral

becul-

benot

ftenno

setyre-

laceeed-ner-

cad-:ss,"laves aandable

1.1.1..111.tt ttttt .J1. tit, 0.4unjust.

"For example, here is a man who says,I am wearing $12 shoes, there is a laborerwho is wearing $3 shoes. Why should Ispend $12 while he could only afford $3?I live in a home that is artisticallydecorated--carpets, high priced furniture,expensive pictures, etc., and the laborerlives in a hovel. As we have said theargument is fallacious. The fact is, thatworkman may have a 10 year intelligencewhile you have a 20. To demand forhim such a home as you enjoy is asabsurd as it would he to insist that everylaborer receive a graduate fellowship.How can there be such a thing as socialequality, with this wide range of mentalcapacity?

"As for an equal distribution of thewealth of the world, that is equally absurd.The man with intelligence has spent hismoney wisely and saved until he hasenough to provide for his needs in caseof sickness, while a man of low intelli-gence, no matter how much money hewould have earned, would have spentmuch of it foolishly and would neverhave anything ahead. It is said that dur-ing the past year the coal miners incertain parts of the country have earnedmore money than the coal operators. Yettoday when the mines shut down for atime, those people are the first to suffer.They did not save anything, although theirwhole life has taught them that mining isan irregular thing, and that when theywere having plenty of work, they shouldsave against the days when they do nothave work."

But low IQ coal miners, according toProf. Goddard, don't have enough intel-ligence to salt away all that extra moneythey're making, etc. You may detect asimilarity between Prof. Goddard's ideol-

11.411, VA. ,,,States. And I would like to run throughbriefly some of the history of that period.

As early as 1912 Prof. Goddard, thePrinceton coal miner speaker. was invitedto Ellis Island by the United States PublicHealth Service and invited to apply thenew science of intelligence testing toimmigrants from Europe. Goddard re-ported in a scientific paper that he hadgiven the intelligence test to a representa-tive sample of European immigrants. Hediscovered that 83 percent of the Jewstrying to enter the U. S. were feeble-minded, that 80 percent of the Hungar-ians were feeble-minded, that 79 percentof the Italians were feeble-minded, andthat 87 percent of the Russians werefeeble-minded. Science and psychologyscore again. And he wrote in 1917, "Thenumber of aliens deported because offeeble-mindedness increased approximate-ly 350 percent in 1913 and 57 percentin 1914 over what it had been. This wasdue chiefly to the efforts of the physicianswho were inspired by the belief thatmental tests could be used for the detec-tion of feeble-minded aliens ?'

Until 1875 there were no federal im-migration laws, whatsoever. Anybodywho wanted to came. In 1875, the firstfederal law was passed. It did not limitthe number of immigrants, but it did ex-clude certain classes of individuals. Andthe first exclusion in 1875 consisted ofthree types of peoplecoolies, convictsand foreign prostitutes. In 1882, lunaticsand idiots were added, and as one goesthrough the years more classes. So by1903 persons who had had two or moreattacks of insanity at any time previouslywere added. The point is that up until theoutbreak of World War I, there was inprinciple unlimited immigration. Certaindefective classes of individuals were ex-

of psychologists and biologists from theeastern seaboard founded somethingcalled the Galton Society which met inthe American Museum of Natural Historyin New York City once a month. It con-sisted at no one time of more than 25distinguished scientists. They provided,as a free service, to the government andto private organizations all kinds of help-ful hints about eugenical practices.

In 1920 there was an enormous in-flux of experimental psychologists intothe Eugenics Research Association. Allat the same time in a given year, atremendous number of distinguished ex-perimental psychologists and biologistsin the country joined the Eugenics Re-search Association. The secretary of theassociation, Dr. Laughlin, got a new job.He remained secretary of the Associationbut he was also appointed "Expert Eugen-ics Agent" for the House Committee onImmigration and Naturalization of theU. S. Congress.

In 1921, the National Academy ofScience published, under the editorshipof Col. Yerkes, an enormous volumewhich contained within it the summary ofall the intelligence testing data gatheredon draftees in World War I.

Most people know that that volumedemonstrated massively for the first timethat blacks scored much lower thanwhites in intelligence tests. But whatalmost nobody seems to know was thatthe immediate practical application ofthis data was not to a black-white ques-tion but to the question of immigration.There was a special chapter in this volumeedited by Yerkes, which presented thedata on the intelligence test scores ofdraftees who were immigrants and whohad been born in different Europeancountries, To page 8

SOUTH TODAY 7

Profound Economic Conflicts, PrRacism Behind Misuse of IQ

From page 7

The U. S. Army test scores could beclassified in terms of A, B, C, D, E. Ifyou are really stupid you are E. Whatthat shows is that the average intelligenceof draftees who are immigrants fromdifferent European countries differ. Itturns out that immigrants coming fromEngland, Holland, Denmark, Scotland,Germany, Switzerland, etc, seem quitebright. The Russians, the Italians andthe Poles are all just plain stupid. ThePoles did not score higher than the blacks.Now in describing this set of interestingdata, the writer used the value-free state-ment, "The Slavic and Latin countriesstand low."

"The Passing ofthe Great Race"

Madison Grant, president of the GaltonSociety had published a very influentialracist book called The Passing of theGreat Race, in which he divided all ofEurope into Mediterranean, Alpine andNordic stocks. Nordics were later re-christened Aryan by other eugenic au-thorities. In any event, Grant had madethe observation that all that was good,beautiful and true in culture and civiliza-tion had come to us from the Nordics andnothing much from the Latin and Slaviccountries. Here we see proof that theintelligence of the Latins and Slays isindeed visibly dull.

In 1922, Col. Yerkes had a civilianjob in the National Research Council inWashington. The National ResearchCouncil under his leadership set up aspecial commission of scientists calledthe Commission for the Scientific Studyof Problems of Human Migration. TheNational Research Council began to sup-port research relevant to problems ofhuman migration. The first research itsupported was that of Prof. Carl Brigham,at that time an assistant professor atPrinceton University in the psychologydepartment. In 1923, the Princeton Uni-versity Press published Brigham's workentitled, A Study of American Intelli-gence. What Brigham did was take theArmy immigrant data and demonstratethat the longer an immigrant had beenin the United States, the higher was hisIQ.

We must assume that he assumed thetest measured native intelligence. Thepsychologists had built it to measurenative intelli once. How can it be that

of Europe is comparatively free fromthis taint."

The idea was that the lowerl classesbreed too much and the upper classesdon't breed enough. So, since intelligenceis inherited, over the years the averageintelligence is going to sink and sinkuntil eventually it gets below zero. Herein the U. S. it's going to be worse, be-cause not only do we have stupid immi-grants, but we have stupid blacks as well.

Brigham wrote, "But the decline ofAmerican intelligence will be more rapidthan the decline of intelligence of Eur-opean national groups, owing to thepresence here of a Negro. The steps thatshould be taken must of course be dic-tated by science and not by politicalexpedience. Immigration should not onlybe restricted but highly selective. Andthe revision of the immigration and na-turalization laws will only afford a slightrelief from our present difficulties. Thereally important steps are those lookingtoward the prevention of the continuedpropagation of the defective strains inthe present population."

For those of you who missed the point,Prof. Brighlm advocated the compulsorysterilization of the "defective." For thosewho want to know what happened toProf. Brigham, he went on to a job assecretary to the College Entrance Exam-ination Board. Also the Scholastic Apti-tude Test was devised by Prof. Brigham.His book was not without influence.

Congressional HearingsIn 1923, the Eugenics Research As-

sociation elected as its chairman a non-scientist. That layman was the Hon.Albert Johnson, chairman of the HouseCommittee on Immigration and Natural-ization. The committee began to havehearings. I have read through the Con-gressional Records and hearing testimonyin some detail. Ill just read a couple ofexcerpts. Dr. Arthur Sweeney testifyingbefore the committee in 1923, said, "Wehave been overrun by a horde of theunfit. We've had no adequate means ofdetermining who are the unfit. While wecan measure objectively the physicalqualifications of the immigrants, we havehad no yardstick with which to performan accurate estimate of the intellectualside. Psychological tests in the Army havefurnished us with the necessary yard-stick. Immigration from eastern andsouthern Europe is most undesirable. Wecan gauge the desirability of immigrants

northern European group. From thissource the stream of intelligent citizenshipis polluted. Who are the unfit? The groupsat the lower end of the psychologicalscale."

Francis Kinnicutt, testifying before aSenate Committee on Immigration andNaturalization in 1923, said, "The largeproportion of this immigration fromsouthern and eastern Europe comes fromPoland or from Russia. The immigrationfrom the last two countries consists largelyof the Hebrew elements. Some of theirlabor unions are among the most radicalin the whole country. The recent Armytests show that all three of theseclasses ranked far below the average inintelligence." Here, he was referring toBrigham's book, A Study of AmericanIntelligence. Col. Yerkes also vouches forBrigham's book and speaks in the highestterms of the author, who, by this time, isassistant professor of psychology atPrinceton University.

Also testifying at these hearings wasMadison Grant, who said, "The countryat large has been greatly impressed bythe results of the Army intelligence tests,carefully analyzed by Yerkes andBrigham. The experts who have analyzedthe statistics and who have tested the testsbelieve that the tests give as accurate ameasurement of intelligence as is possibleat the present time. The questions wereselected with a view to measuring innateability. Experts have told us that hadmental testing been in operation, over .6million Americans now living in this coun-try, free to become the fathers andmothers of future Americans, wouldnever have been admitted."

New Restrictive ImmigrationLaws

In 1921, as the result of an intense na-tional debate, temporary legislation limit-ing the number of immigrants was passedby Congress for the first time. Accordingto the law, the number of immigrants al-lowed into the United States from anyEuropean country depended upon thenumber of people from that Europeancountry already residing in the U. S. Thistemporary law assigned to each Europeancountry an immigration quota based onthree percent of its population residingin the U. S. Even though the temporarylaw passed in 1921, the population statis-tics were based on the 1910 Census.

Just so that you don't miss the point,the permanent law which was passed in

the intelliEuropeanbeen borschools ireally genRussian astupid as

Needleed that i

Polish,stupid. Istudy. Rpsychololaw passhopeful shas toshould bno quotCanadatesting udescentligence othe averand NeMexicanIn ourgetting tpeople.FrenchNew Efilled wiof the Frapproacintelligeforeigning foreface thethick, msugar loImmigrain extrabrunette

I thinmarizedFrank BAssociatbattle wquotasbiologistsociatio"Eugenito meThis isbut itof reseyou rselof immthe aidwouldnently.'

d Economic Conflicts, Profoundm Behind Misuse of IQ Tests

iuld beE. If

. Whatligence

fromfer. Itg fromotland,

quiteis and. The

blacks.restingst ate-

untries

Galionuent ialof theall of

le ander re-ic au-madegood,

iviliza-ics andSlavicat thelays is

civilianncil in'search

up acalledStudy

n. The.0 sup-ms ofrch itgham,sor athologyn Uni-

workIntelli-lie thenstrate

beenas his

ed theThe

casurethat

of Europe is comparatively free fromthis taint."

The idea was that the lower' classesbreed too much and the upper classesdon't breed enough. So, since intelligenceis inherited, over the years the averageintelligence is going to sink and sinkuntil eventually it gets below zero. Herein the U. S. it's going to be worse, be-cause not only do we have stupid immi-grants, but we have stupid blacks as well.

Brigham wrote, "But the decline ofAmerican intelligence will be more rapidthan the decline of intelligence of Eur-opean national groups, owing to thepresence here of a Negro. The steps thatshould be taken must of course be dic-tated by science and not by politicalexpedience. Immigration should not onlybe restricted but highly selective. Andthe revision of the immigration and na-turalization laws will only afford a slightrelief from our present difficulties. Thereally important steps are those lookingtoward the prevention of the continuedpropqation of the defective strains inthe present population."

For those of you who missed the point,Prof. Brighlm advocated the compulsorysterilization of the "defective." For thosewho want to know what happened toProf. Brigham, he went on to a job assecretary ttI the College Entrance Exam-ination Board. Also the Scholastic Apti-tude Test was devised by Prof. Brigham.His hook was not without influence.

Congressional HearingsIn 1923, the Eugenics Research As-

sociation elected as its chairman a non-scientist. That layman was the Hon.Albert Johnson, chairman of the HouseCommittee on Immigration and Natural-ization. The committee began to havehearings. I have read through the Con-gressional Recordsand hearing testimonyin some detail. I'll just read a couple ofexcerpts. Dr. Arthur Sweeney testifyingbefore the committee in 1923, said, "Wehave been overrun by a horde of theunfit. We've had no adequate means ofdetermining who are the unfit. While wecan measure objectively the physicalqualifications of the immigrants, we havehad no yardstick with which to performan accurate estimate of the intellectualside. Psychological tests in the Army havefurnished us with the necessary yard-stick. Immigration from eastern andsouthern Europe is most undesirable. Wecan gauge the desirability of immigrants

northern European group. From thissource the stream of intelligent citizenshipis polluted. Who are the unfit? The groupsat the lower end of the psychologicalscale."

Francis Kinnicutt, testifying before aSenate Committee on Immigration andNaturalization in 1923, said, "The largeproportion of this immigration fromsouthern and eastern Europe comes fromPoland or from Russia. The immigrationfrom the last two countries consists largelyof the Hebrew elements. Some of theirlabor unions are among the most radicalin the whole country. The recent Armytests show that all three of theseclasses ranked far below the average inintelligence." Here, he was referring toBrigham's book, A Study of AmericanIntelligence. Col. Yerkes also vouches forBrigham's book and speaks in the highestterms of the author, who, by this time, isassistant professor of psychology atPrinceton University.

Also testifying at these hearings wasMadison Grant, who said, "The countryat large has been greatly impressed bythe results of the Army intelligence tests,carefully analyzed by Yerkes andBrigham. The experts who have analyzedthe statistics and who have tested the testsbelieve that the tests give as accurate ameasurement of intelligence as is possibleat the present time. The questions wereselected with a view to measuring innateability. Experts have told us that hadmental testing been in operation, over '6million Americans now living in this coun-try, free to become the fathers andmothers of future Americans, wouldnever have been admitted."

New Restrictive ImmigrationLaws

In 1921, as the result of an intense na-tional debate, temporary legislation limit-ing the number of immigrants was passedby Congress for the first time. Accordingto the law, the number of immigrants al-lowed into the United States from anyEuropean country depended upon thenumber of people from that Europeancountry already residing in the U. S. Thistemporary law assigned to each Europeancountry an immigration quota based onthree percent of its population residingin the U. S. Even though the temporarylaw passed in 1921, the population statis-tics were based on the 1910 Census.

Just so that you don't miss the point,the permanent law which was passed in

the intelligence of children of southeasternEuropean immigrants whose children hadbeen born in this country had gone toschools in this country. If the stupidity isreally genetic, then the children of Polish,Russian and Italian immigrants should bestupid as well.

Needless to say, Prof. Hirsch discover-ed that indeed the children of Russian,Polish, Italian immigrants were alsostupid. Let me read to you from hisstudy. Remember this is in a professional,psychological journal. "The immigrationlaw passed by Congress is one of mosthopeful signs. That part of the law whichhas to do with non-quota immigrantsshould be modified." (You see, there wasno quota for immigrants either fromCanada or from Mexico.) "All mentaltesting upon children of Spanish Mexicandescent has shown that the average intel-ligence of this group is even lower thanthe average intelligence of Portugueseand Negro children in this study. YetMexicans are flowing into the country.In our immigration from Canada, we aregetting the less intelligent, working-classpeople. The increase in the number ofFrench Canadians is alarming. WholeNew England villages and towns arefilled with them. The average intelligenceof the French Canadian group in our dataapproaches the level of the average Negrointelligence. I have seen gatherings of theforeign born in which narrow and slop-ing foreheads were the rule. In everyface there was something wrong. Lipsthick, mouth coarse, chin poorly formed,sugar loaf heads, goose-bill noses, . . .

Immigration officials report vast troublein extracting the truth from certainbrunette nationalities."

I think what happened is fairly sum-marized in the presidential address ofFrank Babbott to the Eugenics ResearchAssociation in 1927. By that time thebattle was over and the national originquotas had been passed. Speaking to thebiologists and psychologists in the As-sociation, Babbott said the following:"Eugenics has made its strongest appealto me by its influence on immigration.This is an indirect result of eugenics,but it comes as a natural developmentof research on the part of people likeyourselves. It is possible that restrictionof immigration would have come withoutthe aid of our society, but I doubt if itwould have come so soon or so perma-nently."

T n " A L

National xesearen uouncii began to sup-port research relevant to problems ofhuman migration. The first research itsupported was that of Prof. Carl Brigham,at that time an assistant professor atPrinceton University in the psychologydepartment. In 1923, the Princeton Uni-versity Press published Brigham's workentitled, A Study of American Intelli-gence. WhLt Brigham did was take theArmy immigrant data and demonstratethat the longer an immigrant had beenin the United States, the higher was hisIQ.

We must assume that he assumed thetest measured native intelligence. Thepsychologists had built it to measurenative intelligence. How can it be thatimmigrants that have been in this countrya long time are S.) much brighter than im-migrants who haven't been in this coun-try a long time? Prof. Brigham concludedmat this had nothing to do with knowingthe English language or knowing Ameri-can cultureno. The immigrants, duringthe early 1900s. had arrived from north-western Europe, full of Nordic blood,whereas the immigrants who arrivedfrom 1913 to 1918 were from southeast-ern Europe and consisted of Poles, Ital-ians, Russians, Jews, etc., immigrantswith very little Nordic blood.

From this Prof. Brigham concludedthat there was a clear parallel betweenthe proportion of Nordic blood and theintelligence of the immigrants. That lefta certain policy recommendation. Let meread to you the concluding passages ofProf. Brigham's book. (The foreword inthis book was written by Col. Yerkes.)Brigham says, "We must face a possibilityof racial admixture here that is infinitelyworse than that faced by any Europeancountry today. For we are incorporatingthe Negro into our racial stock, while all

hearings. I have read through the Con-gressional Records and hearing testimonyin some detail. I'll just read a couple ofexcerpts. Dr. Arthur Sweeney testifyingbefore the committee in 1923, said, "Wehave been overrun by a horde of theunfit. We've had no adequate means ofdetermining who are the unfit. While wecan measure objectively the physicalqualifications of the immigrants, we havehad no yardstick with which to performan accurate estimate of the intellectualside. Psychological tests in the Army havefurnished us with the necessary yard-stick. Immigration from eastern andsouthern Europe is most undesirable. Wecan gauge the desirability of immigrantsby the relative proportion of those in theA and B classes. We strenuously objectto the immigration from Italy, Russia,Poland, Greece, and Turkey. The Slavicand Latin countries show a marked con-trast in intelligence with the western and

4,

George L. Walker NI

Dr. Leon Kamin

In 1921, as the result of an intense na-tional debate, temporary legislation limit-ing the number of immigrants was passedby Congress for the first time. Accordingto the law, the number of immigrants al-lowed into the United States from anyEuropean country depended upon thenumber of people from tha' Europeancountry already residing in the U. S. Thistemporary law assigned to each Europeancountry an immigration quota based onthree percent of its population residingin the U. S. Even though the temporarylaw passed in 1921, the population statis-tics were based on the 1910 Census.

Just so that you don't miss the point,the permanent law which was passed in1924 reduced the percentage of the resid-ing European population to two percent,based not on the 1910 census, certainlynot on the 1920 census, not on the 1900census but on the 1890 census, when allthe immigrant blood was Nordic. If onereads the congressional debate, nobodyfooled anybody. The use of the 1890census was explicitly stated to be an ex-clusionary racist policyto keep the in-ferior blood from southeastern Europeaway from our fair shores, while makingit possible for northwestern Europeans toimmigrate. 1-, is was largely rationalizedon the basis that the tests of innateability developed by psychologists hadclearly demonstrated, in study after study,that Italians, Poles, Russians and Jewsscored 25 or 30 IQ points lower than theNordics from northwestern Europe.

You can't appreciate the full flavor ofthe "scientific findings" without readingthe literature. Prof. Hirsch conducted astudy in 1926, supported by the NationalResearch Council, and published in theprofessional journal, Genetic PsychologyMonographs, that raised the question of

Frank BabboAssociationbattle was oquotas hadbiologists ansociation, B"Eugenics hto me by itThis is anbut it comeof researchyourselves.of immigratithe aid of owould havenently."

Today'The majo

country in tof immigraeconomic coThe majorday, our polwelfare memajor econtracism. Thetire biologicprepared toCommitteezation. I'mage of bioloprepared to

The datanow presentis no moreby their sci1920s on thI ige nce amon

That datawas absurd."innate intetoday's data50 years.

it)-ofitm.at

gy['i-rk/U-beitcen:Us

hee

!CC

attryva-

un-tedng

cri-ingth-

,od,vedast-t al-ints

dedeentheleftmeofin

es.)ilityttelycanting

all

hearings. I have read through the Con-gressional Records and hearing testimonyin some detail. just read a couple ofexcerpts. Dr. Arthur Sweeney testifyingbefore the committee in 1923, said, "Wehave been overrun by a horde of theunfit. We've had no adequate means ofdetermining who are the unfit. While wecan measure objectively the physicalqualifications of the immigrants, we havehad no yardstick with which to performan accurate estimate of the intellectualside. Psychological tests in the Army havefurnished us with the necessary yard-stick. Immigration from eastern andsouthern Europe is most undesirable. Wecan gauge the desirability of immigrantsby the relative proportion of those in theA and B classes. We strenuously objectto the immigration from !tidy, Russia,Poland, Greece, and Turkey. The Slavicand Latin countries show a marked con-trast in intelligence with the western and

George L. Walker III

Dr. Leon Kamin

In 1921, as the result of an intense na-tional debate, temporary legislation limit-ing the number of immigrants was passedby Congress for the first time. Accordingto the law, the number of immigrants al-lowed into the United States from anyEuropean country depended upon thenumber of people from that Europeancountry already residinp in the U. S. Thistemporary law assigned to each Europeancountry an immigration quota based onthree percent of its population residingin the U. S. Even though the temporarylaw passed in 1921, the population statis-tics were based on the 1910 Census.

Just so that you don't miss the point,the permanent law which was passed in1924 reduced the percentage of the resid-ing European population to two percent,based not on the 1910 census, certainlynot on the 1920 census, not on the 1900

census but on the 1890 census, when allthe immigrant blood was Nordic. If onereads the congressional debate, nobodyfooled anybody. The use of the 1890census was explicitly stated to be an ex-clusionary racist policyto keep the in-ferior blood from southeastern Europeaway from our fair shores, while makingit possible for northwestern Europeans toimmigrate. This was largely rationalizedon the basis that the tests of innateability developed by psychologists hadclearly demonstrated, in study after study,that Italians, Poles, Russians and Jewsscored 25 or 30 IQ points lower than theNordics from northwestern Europe.

You can't appreciate the full flavor ofthe "scientific findings" without readingthe literature. Prof. Hirsch conducted astudy in 1926, supported by the NationalResearch Council, and published in theprofessional journal, Genetic PsychologyMonographs, that raised the question of

1-rank Babbott to the Eugenics ResearchAssociation in 1927. By that time thebattle was over and the national originquotas had been passed. Speaking to thebiologists and psychologists in the As-sociation, Babbott said the following:"Eugenics has made its strongest appealto me by its influence on immigration.This is an indirect result of eugenics,but it comes as a natural developmentof research on the part of people likeyourselves. It is possible that restrictionof immigration would have come withoutthe aid of our society, but I doubt if itwould have come so soon or so perma-nently."

Today's Data "Absurd"The major domestic issue facing the

country in the early 1920s was the issueof immigration. It involved profoundeconomic conflicts and profound racism.The major issue facing the country to-day. our politicians tell us, is "the greatwelfare mess," which again involvesmajor economic conflict and profoundracism. There was no shortage of scien-tific biologists and psychologists in 1920prepared to act as teachers to the HouseCommittee on Immigration and Naturali-zation. I'm afraid there will be no short-age of biologists, psychologists and othersprepared to act as teachers today.

The data which such teachers arenow presenting to policymakers, I think,is no more valid than the data presentedby their scientific predecessors of the1920s on the differences in innate intel-ligence among the racial stocks of Europe.

That data, as all would agree today,was absurd. It had nothing to do with"innate intelligence." In my opinion,today's data will look just as absurd in50 years.

8 SOUTH TODAY

DOCUMENT RESUME - ERIC · 2014. 1. 2. · DOCUMENT RESUME ED 084 318 UD 013 879 AUTHOR TITLE. SPONS...

Documents

Transcript of DOCUMENT RESUME - ERIC · 2014. 1. 2. · DOCUMENT RESUME ED 084 318 UD 013 879 AUTHOR TITLE. SPONS...