THE SATURDAY EVENING POST

The Her it ability of IntelligenceARTHUR R. JENSEN

The question of why people differ in intelli-gence bas been asked for centuries, but a scientifi-cally acceptable answer did not become whollypossible until psychologists devised techniques formeasuring intelligence quantitatively and objec-tively. The first really useful intelligence test wasdevised in 1905 by the French psychologist AlfredBinet. Binet's early test was later revised and im-proved by Lewis Terman at Stanford University;the now-famous test that resulted from these ef-forts is known as the Stanford-Binet IntelligenceScale. It is stiU the most widely used test of generalintelligence.

There are other intelligence tests, and althoughmany of them appear to be quite different fromone another, all actually measure much the samegeneral ability. That is to say, if we administerseveral seemingly quite different intelligence teststo a large number of persons, their scores on all thetests will be in pretty much the same rank order.Those who score high on one test will tend to scorehigh on the others, and those who score low on onetest will usually score low on all the others. Thisfact of correlation among all tests of intelligenceled Charles Spearman, the famous English psy-chologist, to conclude that there is a general fac-tor, "g," which is common to all tests of intelli-gence. We know that it is practically impossible tomake up a mental test having any degree of com-plexity which does not involve "g." We can per-haps most clearly characterize "g" as an ability forabstract reasoning and problem solving, for seeingrelationships, and for grasping concepts.

A person's score on an intelligence test is usu-ally expressed as an LQ. (for Intelligence Quo-tient). The test is standardized in the general popu-lation in such a way that the average LQ. at any ageis set at IOO, and the middle 50 percent of thepopulation falls within the so-called average range

of l.Q.'s going from 90 to 110.Significance of the LQ.

Can the LQ. tell us anything of practical im-portance? Is it related to our commonsensenotions about mental ability as we ordinarily thinkof it in connection with educational and occupa-tional performance? Yes, indeed, and there is nodoubt about it. The massive evidence from psy-chological, educational and industrial research,and research in the armed forces, is unequivocal.We know, for example, that no other single factthat we are now able to ascertain about a childgives us a better prediction ofhis future scholasficperformance than his LQ. obtained after age 5 or6. (Below this age LQ. tests become less accurateindicators of the child's later mental development,and below 2 or 3 years of age test scores have prac-tically no predictive value.)

The LQ. obtained after 9 or 10 years of age alsopredicts final adult occupational status to almostas high a degree as it predicts scholastic per-formance. When various occupations are rankedfor average income and for the general public'saverage judgment of the occupation's prestige anddesirability, this rank order is found to be highlyrelated to the average LQ. level of the persons inthese occupations. There is of course a wide spreadof l.Q.'s in nearly every occupation, but the aver-age LQ. of persons within a particular occupationis closely related to that occupation's standing interms of its average income and the amount ofprestige accorded to it by the general public.

One of the most convincing demonstrationsthat LQ. is related to "real life" indicators of abil-ity was provided in a classic study by Terman andhis associates at Stanford University. In the 192O'sthey selected a total of 1,528 children with Stan-ford-Binet l.Q.'s above 140. The average LQ. ofthe group was 152. These children were investi-gated periodically over the years up into theiradulthood. (Most of them are now in their fifties.)Terman found that for the most part these high-

l.Q. children in later adulthood markedly excelledthe general population on every indicator ofachievement that was examined: a higher level ofeducation completed; more scholastic honors andawards; higher occupational status; higher income;production of more articles, books, patents, andother signs of creativity; more entries in Who'sWho: ii lower mortality rate; better physical andmental health; and a lower divorce rate. Also, theyhave much brighter children than the average; theiraverage LQ. is 133, a level which is exceeded byonly 2 percent of all the children in the generalpopulation.

Findings such as these establish beyond a doubtthat LQ. tests measure characteristics that areobviously of considerable importance in our pres-ent technological society. To say that the kind ofability measured by intelligence tests is irrelevantor unimportant would be tantamount to repudi-ating civilization as we know it.The Causes of LQ. Differences

The layman usually asks: "Is intelligence due toheredity or environment?" The scientist promptlyanswers: "Both." WitlKJut heredity and environ-ment there simply is no intelligence. Obviouslyevery person must have had a biological inheri-tance of genes from his parents and must havegrown in an environment, or he wouldn't even behere to take an LQ. test. So, of course, both hered-ity and environment are essential for the exis-tence of the individual or any of his physical andmental characteristics.

But when scientists actually study this prob-lem, we find that they do not even ask the lay-man's question. The question to which scientistshave sought an answer can be stated as follows:How much of the variation among persons in agiven population is attributable to differences intheir environments and how much to differencesin their genetic endowments?

Numerous studies conducted by psychologistsand geneticists over the last 40 or 50 years provide

Arthur R. Jensen is Professor of EducationalPsychology and Research Psychologist at theInstitute of Human Learning at the Univer-sity of California (Berkeley). Since 1958 hisresearch has included individual differencesin human learning, and cultural, develop-mental, and genetic determinants of intel-ligence in learning ability. He has publishedover 100 articles and chapters in psychologicaland educational Journals and books. His re-searches have suggested a strong genetic fac-tor in human intelligence. His article in theHarvard Educational Review (Winter 1969) hascaused considerable controversy.

13 THE SATURDAY EVENING POST Summer 1972

an answer to this question. The answer is un-ambiguous and is generally agreed upon by allscientists who have considered all the evidence.This evidence strongly supports the conclusionthat genetic factors are much more important thanenvironmental influences in accounting fot indi-vidual differences in I.Q. How much more impor-tant? The evidence indicates that genetic factorsaccount for at least twice as much of the variationin l.Q.'s as environmental factors. This conclusionhas one main limitation. Since all of the majorstudies in this field were conducted with samplesof Caucasian European and North American popu-lations, we cannot confidently generalize theirconclusions to other populations, especially thosewith very dissimilar environments.

What are the kinds of evidence that lead to theconclusion that genetic differences outweighenvironmental differences in accounting for indi-vidual differences in I.Q.? Most of this evidence, asit is found in the scientific hterature, dependsupon quite technical methods of analysis de-veloped in a specialty known as quantitative genet-ics or population genetics. Some of these methodswere devised originally to analyze the roles ofheredity and environment in agriculture and ani-mal breeding.Experiments in Animal Breeding

Experiments in which we explicitly try tobreedfor some specific trait give us the most certain evi-dence that variation in the trait has a genetic com-ponent. Psychologists have bred rats for speed oflearning mazes, which is a good indicator of ratintelligence. By always mating the fast-learningmales with t~ast-learning females, and mating slow-learning males with slow-learning females, it ispossible, within 6 to 10 generations, to producetwo quite distinct strains ot" rats in respect to maze-learning ability. The slowest-learning rat of the"bright" strain will learn mazes faster than thefastest rat of the "dull" strain. The two strains willdiffer markedly in the number of tries they need tolearn how to run through a maze efficiently, avoid-ing the blind alleys. These experiments definitelyprove that not only physical characteristics butsome behavioral traits as well are largely inheritedthrough the parental genes. Thus we should not besurprised to find in humans that differences insome behavioral characteristics, including inteUi-gence, are a product of genetic inheritance.Identical Twins Reared Apart

One of the most important lines of evidence forthe inheritance of intelligence in humans comesfrom studies of identical twins who were separatedshortly after birth and reared in different homes.Identical twins originate from a single fertilizedovum which splits in the course of early develop-ment to form two individuals. Each member of thepair of twins therefore has exactly the same com-plement of genes. Consequently, any differencebetween the twins must be due entirely to non-genetic or environmental differences.

Twins separated shortly after birth are oftenreared in families that differ markedly in socialclass, and the range of environmental differencesobserved in their foster homes is fairly typical ofthe environmental variations seen in the generalpopulation.

Four major studies of identical twins rearedapart, conducted in England, Denmark, and theUnited States, and totaling 122 pairs of twins, arein remarkably close agreement in showing thattwins reared in different homes are still much more

Continued on page 149

Do I.Q. Tests MeasureIntelligence?

DAVID c. MCCLELLANDPsychology has one great practical success to its

credit in the twentieth century-namely the intelli-gence testing movement. Many tests have beendevised which predict success in school with re-markable regularity. Literally tens of thousands ofvalidity coefficients have been calculated, demon-strating that those who score higher on aptitude orintelligence tests usually do better in their schoolwork. Selecting, at random, a finding which isquite typical for the United States, I recently ob-served in a longitudinal study to be reported byCosta (1972) that Kuhlman-Anderson I.Q. scoresobtained in the sixth grade correlated 0.59 withtwelfth-grade rank in class. In other words know-ing how a child scores on an intelligence test whenhe is eleven or twelve years old enables you to pre-dict fairly accurately how well he will be doing inschool some six or seven years later. Rank in classat graduation from secondary school in turn pre-dicts whether he can go on to the university andhow good a university he will get into. As a con-

David C. McClelland is Professor of Psychol-ogy, Harvard University. Appointed to Har-vard in 1956, he has also served as chairman ofthe Fulbright Advisory Panel on Psychologyand as director of research projects for theFord Foundation, the Carnegie Foundation,the State Department, the National Instituteof Mental Health and the Peace Corps. Pro-fessor McClelland is author of many books,including Personality; The Achieving Society;The Roots of Consciousness; Motivating EconomicAchievement, and Alcohol and Human Motiva-tion.

sequence, knowing a person's intelligence-testscore or schoiastic-aptitude-test score has becomea matter of great importance in the United States,not only to admissions officers who use it to pickpeople for college but also to businesses and civilservice commissions who use it to decide who is"bright enough" to be a policeman, a social workeror a fireman.

Testing has therefore become big business. TheEducational Testing Service which gives theScholastic Aptitude Test used by most of the bet-ter-known colleges and universities in the UnitedStates employs around two thousand people andhas a large plant spread over hundreds of acres inPrinceton, New Jersey. Thousands of young peo-ple pay to take its tests annually to see if they arequalified to get into the college of their choice.The testing technoiogy has been so sold to theAmerican public that only in a few of the more"backward" parts of the society is it not used inthe schools or businesses or civil service. And ofcourse it is spreading fast to the rest of the world,which is beginning to discover the utility of testsfor picking those who will do well in school.

To be sure, the testers themselves loudly insist

that there are other important human qualities be-sides the ability to take scholastic aptitude tests,but as Wing and Wallach (1971) have shown,admissions officers may believe they take theseother qualities into account but in fact their selec-tion decisions can be almost perfectly predicted byaptitude-test scores alone. The desite to selectmore "intelligent" people for schooling or foralmost any occupation proves overpowering. Itquickly reduces other qualifications to insignifi-cance.

While the intelligence-testing movement in theUnited States has been moving on from one tri-umph to the next, some questions have been raisedabout its theoretical underpinnings, both byscholars and by policy makers who wonder if itsgrowing power over people's lives is justified. Onedifficulty with tests has long been known but littlecommented on perhaps hecause its seriousness hasnot been fully appreciated. It is very simply that ifacademic achievement tests are taken seriously asmeasures of real competence, then the quality ofeducation does not seem to contribute to improv-ing competence. Back in the 193O's in the United

States, a number of private schools tried to im-prove the quality of their education as part of whatwas then known as the "progressive educationmovement." Standardized scholastic achievementtests were used to evaluate the effects of this sup-posedly improved education as compared withmore traditional teaching.

By and large no effects of the supposedly high-er-quality education could be discovered in thetest scores. The educators felt they were doing abetter job but the test scores did not indicate thatthey were. The same finding has turned up againand again since that time. Certain colleges in theUnited States are widely acknowledged to be bet-ter than other colleges—in the sense that they havebetter faculties, more books in the library, higherendowments, better laboratory facihties, and soforth. Yet repeated studies as summarized byJacob (1957) have failed to show any test-scoredifferences attributable to the better educationsupposedly obtained in the elite colleges. If thegraduates of those colleges perform better onachievement tests, it is because they scored higheron them at entrance to college, not because theyreceived a better education subsequently.

THE SATURDAY EVENING POST 149

nonsense and order the machineryhe needs to carry out the improve-ments on his property. But he stillinsisted on waiting until he heardfrom the Professor.

While waiting, we all checked intothe GovernorChittenden Motel. Therest of our group has been seeing thesights of this beautiful city: theparks, the University, the ShelburneMuseum, the Lake. And I have beenwriting this letter, which I wil! mailtomorrow morning.

I am sorry I cannot as yet reportany definite success. But sometime,somehow, something may turn up.

Yours hopefully,Alexander Bolts

P.S. The next morning. Tuesday,January 25, 1972. Somethinghasin-deed turned up. When I entered thePost Office this morning to mail thisletter, my eagle eye spotted, on abulletin board, two photographs-one full face and the other profile—of none other than Professor IsaacNewton Norton, Ph. D., alias thisand that. He was wanted for grandlarceny, fraud, and other crimes. Iconsulted the local police. They saidthey were certainly after the man.And they surmised that the mys-terious remark, "The fuzz is startingto buzz," was a warning that thecops were after him. 1 took the newsto Monkeyface. We both decided theProfessor was on the lam and wouldnot return. Monkeyface then signeda magnificent order for Earthwormequipment -which is enclosed.

All this makes me feel so goodthat I am repeating the statement Imade in an earlier letter: I am notcharging anything for my servicesthis time. I am also turning down anoffer from Monkeyface, who inci-dentally is making a generous con-tribution to Miss Gloriana Smith andthe other New York ladies. They arenow very good friendswith Monkey-face, and they are working on anewslogan for their organization—some-thing like "Convert polluters of theenvironment by kindness, not mere-ly by kicking them in the pants."

This wraps everything up, exceptfor one nagging doubt: IsthisSuper-mohole idea really as crazy as itsounds?

Yours,A.B. n

A tree that never had to fightFor sun and sky and air and lightBut stood out in the open plainAnd always had its share of rainNever became a forest king.But lived and died a scrubby thing.

The man who never had to fightWho never had to win his shareOf sun and sky and light and airNever became a manly man.But lived and died as he began.

Good timber does not grow inease

The stronger the wind, thetougher the trees.

J. Willard Marriott

alike in I.Q. than are fraternal twinsreared together. Fraternal twins aremerely siblings who happen to beconceived and born at the same time,and therefore half of them are of theopposite sex. In I.Q. and other traitsthey resemble one another no morethan do ordinary siblings born at dif-ferent times.

Identical twins reared apart dif-fer, on the average, by only 6 to 7I.Q. points. But even if we test thevery same person on two occasions aweek apart, we find that his testscore will vary, on the average, by 2or 3 I.Q. points. This is the test's"measurement error." When weeliminate this error from the twindata, we find that the twins differonly 4 or 5 points in I.Q. Identicaltwins reared together differ by only2 or 3 points, not including measure-ment error. The largest I.Q. dif-ference ever found in a pair of identi-cal twins reared apart is 24 points.More than 17 percent of siblingsreared together differ by more than

continued front page 12

tween unrelated children brought upin different homes but in the samesocioeconomic class. We see that un-related children brought up togetherin the same home differ from oneanother in I.Q. at least 3 or 4 timesmore than genetically identical twinsreared in different homes. And theunrelated children reared togetherdiffer almost as much in I.Q. as un-related children simply picked atrandom from different homes.

The l.Q.'s of adopted children al-so show little or no relationship tothe l.Q.'s of their adopting parents,but they are almost as closely relatedto the l.Q.'s of their natural parentsas we find in the case of children whoare reared by their natural parents.

Children reared in the commonenvironment of an orphanage differfrom one another in I.Q. to approxi-mately the same degree as childrenpicked at random from the totalpopulation. The l.Q.'s of orphanagechildren who have never known theirown parents show almost the same

wm

" / guess this just wasn't our day."U>^'^

24 I.Q. points. The same is true offraternal twins. But siblings (andfraternal twins) have only half oftheir genes in common, and they dif-fer on the average by 12 I.Q. points(excluding measurement error), evenwhen reared together.

The studies of identical twinsshow clearly that individuals who aregenetically identical are almost asmuch alike in mental ability as theyare alike in physical traits, and this istrue even when they have grown upin different environments.Unrelated Children Reared Together

The opposite situation to identi-cal twins reared apart is that ofgenetically unrelated childrenadopted at birth by foster parentsand reared together. Such childrendiffer from one another, on theaverage, by 1 5 to 16 I.Q. points (ex-cluding measurement error). Com-pare this with the 17 to 18 I.Q.points difference between unrelatedchildren reared in different homes,or the 15 to 16 points difference be-

degree of correlation with theirparents' level of ability as we find inthe case of children reared by theirown parents.Resemblance Between Parentsand Children

Now and then we notice that verybright parents can have an intellec-tually mediocre child, or that ratherdull parents can have an exception-ally bright child. These observationsare often pointed to mistakenly asevidence that intelligence is not in-herited. But the fact is that genetictheory predicts precisely that weshould find such discrepancies be-tween parents and their offspring.For example, parent-offspring dif-ferences in height are of about thesame relative magnitude as their dif-ferences in I.Q. Children resembletheir parents physically and inmental ability to about the same de-gree that they resemble their ownsiblings. The average LQ. differencebetween a parent and his (or her)child is the same as the difference

between siblings-that is, about 12LQ. points. The difference betweena child and the average of both of hisparents'I.Q.'sisabout lOpoints.

A parent with a high LQ. willusually, but by no means always,have children whose l.Q.'s are some-what lower than his own but are stillabove the average for the generalpopulation. A parent wit ha lowLQ.,on the other hand, will usually, butnot always, have children whosel.Q.'s are somewhat higher than hisown but are still below the average ofthe population. This phenomenon,discovered by Sir Francis Galton, iscalled "regression toward themean," and it holds true for heightand other inherited physical traits aswell as for LQ.l.Q.'s of Husbands and Wives

It is interesting that in our societyhusbands and wives are at least asmuch alike in LQ. as brothers andsisters. If men and women pickedtheir mates strictly at random, as bya lottery, spouses would differ by anaverage of 18LQ. points. But in factmen and women choose one anotherpartly for intelligence, and sospouses differ by only 10 or 11points in LQ.The Effect of Inbreeding on I.Q.

Every person harbors a number ofmutant, recessive genes. Most ofthese are defective genes. They arepassed on from parent to child, butthey usually will not produce anyharmful effects to the child unlessthe other parent also contributesexactly the same defective gene. Thereason this usually does not occur isthat each parent's normal genes aredominant over the other parent'sdefective, recessive genes. Whenmating occurs between a man and awoman who are blood relations,however, the chances are muchgreater that they will both possessmany of the same defective genes.When these defective genes arepaired together in the related cou-ple's children, they subtract unfavor-ably from the traits that are con-trolled by these genes under normalconditions. This depression due toinbreeding is known to occur in in-herited physical traits, such as stat-ure, and the same thing has beenfound for LQ. It is well established,for example, that cousin marriagesproduce children who, on the aver-age, have l.Q.'s several points lowerthan the l.Q.'s of children whoseparents are unrelated but arematched with the married cousins onLQ., age, educational level, andsocioeconomic status. More extremeare the cases of children who haveresulted from incestuous relation-ships, such as father-daughter andbrother-sister matings. These chil-dren show a much higher incidenceof severe mental retardation thanchildren born to the same parentswhen they have mated with un-related persons. These interesting

150 THE SMTVRDAY EVENING POST Summer 1972

findings are entirely predictablefrom basic principles of genetics thatapply to all living beings. Moreover,it is virtually impossible to explainsuch facts without concluding thatLQ. differences are very strongly in-fiuenced by genetic mechanisms.The Relative Effects of Heredityand Environment

How can we summarize brieflywhat is now known about the rela-tive importance of heredity and en-vironment in causing individual dif-ferences in LQ.? In the terminologyof genetics a summary answer con-sists of saying that the "heritability"of LQ. is close to 0.80. This meansthat 80 percent of the "variance" inl.Q.'s in the general population is at-tributable to genetic differences and20 percent is attributable to non-genetic or environmental differ-ences.

"Variance" is essentially a quanti-tative index of the total amount ofdifferences that exist among allmembers of some population. So in-stead of talking about variance, wecan more easily describe our con-clusions in terms of average differ-ences.

If we should determine the differ-ences in LQ. between every person inthe population and every other per-son, the average of all these differ-ences would turn out to be 18 LQ.points. These differences are dueboth to genetic and to environ-mental factors. Now we can asktheoretically: What would be theaverage LQ. difference among allpersons in the population if every-one had grown up in identical en-vironments from the moment ofconception, while genetic differ-ences remained as they are? Underthis hypothetical condition of com-pletely equal environments foreveryone, the average LQ. differencewould be 16 points. Thus, therewould be a reduction of 2 points inthe average difference that nowexists. Let us now ask the reverse:What would be the average differ-ence if everyone had exactly thesame genetic endowment, but en-vironmental differences remainedunchanged? Under this hypotheticalc o n d i t i on of complete geneticequality the average LQ. differenceamong persons would be only 8points, or just half the differencethat would exist with equal environ-ments.

So the conclusion we come to—which is certainly valid at least in thewhite European and North Ameri-can populations in which the re-search was conducted—is this: Inaccounting for the causes of the dif-ferences among persons in LQ., thegenes outweigh the effects of en-vironment by 2 to 1. As environ-mental conditions are improved andmade more alike for all persons inthe society, the average intelligencelevel of the population will be some-what increased, and the LQ. dif-ferences among persons will beslightly reduced. But of course thedifferences that remain will in-

evitably be due even more to geneticfactors.

* * *Scholastic Achievement and

Intelligence Statement ofArthur R. Jensen to theU.S. Senate Committee

on EducationIntroduction

Mr. Chairman and members ofthe Committee, my name is ArthurR. Jensen. I am Professor of Educa-tional Psychology at the Universityof California at Berkeley, where Ihave been a member of the facultyfor the past 14 years.

First, I want to thank this Com-mittee for inviting me to summarizemy position on the topics mentionedin your invitation: "heredity, en-vironment, intelligence, and scholas-tic achievement." My position onsome aspects of these questions, asyou know, is a controversial and un-popular one, but it is a positionwhich I have come to through myintensive studies over the past

evant evidence contradict the be-lief. The reactions have often beenextreme. In a recent Newsletter ofthe American Anthropological Asso-ciation, for example, it was pro-posed, apparently in all seriousness,that members of the AAA shouldburn—literally burn-all copies ofthe Harvard Educational Reviewcontaining my article! 1 am thereforegrateful for this opportunity tomake clear my position on these im-portant matters to the Committee,and also to see that a number of dis-tinguished scientists are concernedwith some of the issues I have raisedand are here today to express theirviews. Though I do not expect alltheir opinions to coincide com-pletely with mine—indeed, theywere selected expressly to insurethat a variety of views might be ex-pressed before this Committee—I amespecially gratified to see that thereare prominent scientists who, likemyself, are trying to come to gripswith these difficult questions nowunder discussion.

"Hi. My name is Percy . . . P as in prison, E aa in evasion,R as in receipt, C as in catch, Y as in you . . ."

decade of virtually all the relevantevidence in this field and through myown experimental and statistical in-vestigations. The reporting of theseconclusions in scholarly journals, asin my well-known article in theWinter 1969 issue of the HarvardEducational Review, has not beenwithout some penalty. At least Ihave heard of no one taking the con-trary position who as a result has hadto endure personal threats, pro-longed harassment, invasion of hiscollege classes by disruptive demon-strations, or has had his universityadministration take such pre-cautions as putting two plainclothespolice bodyguards on him-in lec-tures, on the campus, in going to andfrom classes and to the parking lot.Much of this emotional reaction tomy Harvard Educational Reviewarticle, I believe, is a result of the factthat a generation of social scientistsand educators has been assiduouslyindoctrinated to believe that geneticfactors are of little or no importancein human behavior and human dif-ferences. My summaries of the rel-

Scholastic Achievementand Intelligence

Scholastic achievement is whatchildren learn in school-repertoireof knowledge and skills, includingthe 3 R's. Scholastic achievementcan be most reliably measured bymeans of standardized tests at everystage of schooling from kindergartento college. When entire school popu-lations are tested on scholasticachievement, there is revealed a widerange of individual and group dif-ferences at any given grade level. Inone fifth-grade classroom in SanFrancisco, for example, achievementlevels ranged from first to eleventhgrade. What most educators, govern-ment officials, and writers in thepopular press who talk about thepresent problems of education are infact referring to is not primarily dis-satisfaction with so me a &.SO/U re levelof achievement, but rather with thelarge group differences in educa-tional attainments that show up soconspicuously in our educationalsystem-the achievement gaps be-tween the affluent and the poor, the

lower class and the middle class, themajority and the minority, the urbanand the suburban, and so on. Educa-tional differences, not absolute levelof performance, are the main causeof concern. Whether we like to admitit or not, the problem of achieve-ment differences today is where theaction is, where the billions of dol-lars of educational funds are beingpoured in, where the heat is on, andwhere the schools are being tornapart. We are trying to understandmore about the causes of these dif-ferences. Massive surveys and statis-tical analyses such as James Cole-man's well-known report on Equal-ity of Educational Opportunity(1966) have shown that only a smallfraction (about one-tenth to one-fifth) of the total variation in scho-lastic achievement is attributable tofactors in the schools themselves. Inother words, differences amongschools and school systems nation-wide are not sufficiently large toaccount for more than a small frac-tion of the total variation in scholas-tic achievement at any grade level.

The single most powerful pre-dictor of children's scholastic perfor-mance is inteUigence as measured byany one of a variety of standard in-telligence tests: group or individualtests, verbal or nonverbal. No othervariable has yet been found whichmakes as large an independent con-tribution to variance in scholasticachievement as does intelligence. Ihave found in my own research thata composite of several differentmeasures of intelligence will predictnearly all the true variance in scho-lastic achievement scores, and thepredictive validity of the intelligencetests becomes better with advancinggrade level. Furthermore, the valid-ity of intelligence tests for predictingscholastic achievement is not signif-icantly different for white and blackchildren. In this respect, intelli-gence tests are quite color blind.That is to say, a white child and ablack child with the same I.Q. can beexpected to perform about equallywell in school. In short, if intelli-gence tests can be said to be good foranything, they are good for pre-dicting scholastic achievement.

Intelligence, in the technicalsense that psychologists use thisterm, is not the same thing as scho-lastic achievement. Schools do notteach intelhgence per se. Intelligenceis mental brightness; it is a capacityfor conceptualization, abstractreasoning and problem solving, forprocessing information in the formof words and symbols, for integrat-ing and understanding what islearned, and for making broad trans-fer from past learning to the solutionof novel problems. As Harvardpsychologist Professor LawrenceKohlberg recently noted, scholasticachievement merely rides on theback of intelligence: " . . . brightkids learn the stuff they're taught inschool faster, but learning the stuffthey're taught in school doesn'tmake them brighter" {Education

THE SATURDAY EVENING POST iSi

Summary, August 6, 1971).Intelligence can be quite reliably

measured by appropriate tests, andthese measurements, often calledl.Q.'s, show substantial correlationswith a number of educationally, oc-cupationally, and socially importantcriteria. The correlationsare not appreciably dif-ferent for children of therich than for children of thepoor, for whites or forblacks. Although intelli-gence, as psychologists usethis term, is the single mostpowerful, though by nomeans perfect, predictor ofacademic performance, it issurely not the whole ofmental ability and humancompetence. To equate in-telligence with all virtue is aridiculous mistake. But indiscussing achievement dif-ferences in schools withtheir present curricula andtheir present instructionalmethods, intelligence dif-ferences, we know, are ofcentral importance. Thereare undoubtedly othersocially important mentalabilities besides intelli-gence, and on some of theseabilities we find little ornone of the social class andracial differences that wefind for intelligence. Muchof my own research hasbeen concerned with iden-tifying such abilities andwith trying to determinetheir relevance to instruc-tional methods that mightmake better use of theseother abilities for scholasticlearning.

Both intelligence andscholastic achievementgrow or increase in a quiteregular manner in mostchildren, and their indi-vidual mental growthcurves become increasinglystable over the years frominfancy to maturity.The Heritability ofIntelligence

A number of geneticstudies carried out in Eu-rope and the United Statesover the past forty yearsprovide evidence whichshows quite conclusivelythat, in the populationsstudied, a very substantialproportion of the variabil-ity among persons in intelli-gence is attributable togenetic, i.e., inherited, fac-tors. The vast majority ofstudies have found that theproportion of populationvariance attributable togenetic differences is some-thing between the extremelimits of about .60 to .90, afigure known as the heri-tability (in the broadsense). (Since heritability,i.e., the proportion of the

total variance which is geneticallydetermined, is a population statisticsubject to samphng error and othersources of variation, it has no univer-sal or constant value for all times, alltests, and all populations. But em-pirically determined values are usu-

ally of the order of .70 to .80.) 1 be-lieve it is safe to say that in Europeanand North American Caucasianpopulations at the present time,genetic or hereditary factors areroughly twice as important as en-vironmental variation as a cause of

individual differences in intelligenceas assessed by standard tests.

Educational ImplicationsAt present, neither I nor anyone

else, I'm afraid, has any more thanrather general notions concerning

Prosperity '^ yoursfor 50^ a day!

THE LAND BUY OF A LIFETIME...SAN LUIS VALLEY RANCHESIN BOOMING COLORADO!

'395 PER ACRE

GUARANTEED INVESTMENT PLANNO INTERESTNO DOWN PAYMENTNO SALESMAN'S COMMISSIONS

COLORADO IS BOOMINGEarly buyers reaped fortunes in the Denver area. Landbought for $200.00 an acre was sold for $20,000.00. Theopportunity for profits like this in Denver are gone . . . buta new area of Colorado is opening for public sale for thefirst time . . .

SAN LUIS VALLEY RANCHESThe future here is unlimited. The San Luis Valley Ranchesare located in the heart of the beautiful San Luis Valley,surrounded by ranges of the Rocky Mountains. This is beau-tiful land, level to rolling. The air is clean; the climatehealthy. All Ranches are accessible by graded roads. Thisis a perfect place for raising children, retirement or as avacation paradise.

RECREATION " ^Every outdoor recreation is here. Hunting for duck, elk,deer and bear. Camping and exploring in the Rio GrandeNational Forest. You'll find great fishing, boating and waterskiing within 3 miles of your property. Wolf Creek Pass,only 60 miles away, offers wonderful skiing and wintersports. Yes, the San Luis Valley is truly a scenic and recre-ation wonderland.

ALL THIS AND MODERN CONVENIENCES TOO!The cities of Alamosa and Blanca are friendly neighbors,with stores, supermarkets, schools, churches and a hospi- 4tal. Blanca is less than 3 miles and Alamosa 14 miles fromSan Luis Valley Ranches. Alamosa is the home of Adams

state College, one of Colorado's fastest growing Institutionsof higher learning.

YOU CAN BUY WHERE THE GIANTS INVEST!You can feel a lot better about buying land where you knowlarge corporations are investing. In the San Luis Valley,Forbes, Inc., publishers of the world's most respectedfinancial magazine, recently bought a giant 168,000 acreranch nearby San Luis Valley Ranches. Malcolm Forbes,their president, publicly stated that he will turn the ranchinto the largest and best known game preserve in the coun-try, leaving the natural beauty untouched

THE PRICE IS STILL LOW — N O DOWN PAYMENT!*Just $395.00 per acre in 5 acre parcels. Your total monthlypayment is $15.00, LESS THAN 50c A DAY. There is nodown payment, no interest, no salesman commissions, nocarrying charges of any kind. Every cent you pay is appliedto the purchase price of your land.

WATCH YOUR DOLLARS G R O W . . .Remember, this is acreage, not just a little plot of ground.Your 5 acre ranch will be equal in size to 20 city lots.We all know about the fortunes made by original investorsin Denver. So . . .

ACT NOW —RANCHES ARE GOING FASTi As you might imagine, property at this price and these

terms is going fast. Return the Reservation Form and a 5acre parcel will be reserved in your name. No obligation,ACT NOW!

GUARANTEEDINVESTMENT PLAN

BUYER MAY AT ANY TIME, WITHINONE YEAR FROM DATE OF THE CON-TRACT, COMPLETE A COMPANYGUIDED TOUR OF THE SAN LUISVALLEY RANCHES AREA, AND, IF NOTSATISFIED. REOUEST tN WRITING AREFUND OF ALL MONIES PAID ONTHE CONTRACT. THIS REQUESTMUST BE EXECUTED IMMEDIATELYUPON COMPLETION Of SAID TOUR.

SEP-fi-72

SAN LUIS VALLEY RANCHES / BOX 836 / ALAMOSA, COLORADO 81101YES, RESERVE a 5 acre Ranch for me. I understand the price is $395 peracre (total price $1,975) and my total monthly payment for the 5 acres willbe $15 . . . no down payment. . . no interest nor carry charges. Please rushcomplete details, FEDERAL LAND REGISTRATION REPORT, purchase contractand map showing the exact location of my property, n Enclosed is my $1reservation deposit which will be credited to my account or fully refunded atmy request within 30 days. Reservations subject to prior sale.THIS OFFER IS SUBJECT TO ALL LOCAL, STATE AND FEDERAL REGULATIONS.

Name.

-State & Zip_

153 THE SATURDAY EVENING POST Summer 1972

the educational implications of thewide range of apparent differences ineducability in our population.

There is fundamentally, in myopinion, no difference, psychologi-cally and genetically, between indi-vidual differences and group differ-ences. Individual differences oftensimply get tabulated so as to show upas group differences betweenschools in different neighborhoods,between different racial groups, be-tween cities and regions. They thenbecome a political and ideological,not just a psychological, matter. Toreduce the social tensions that arisetherefrom, we see proposals to abol-ish aptitude and achievement test-ing, grading, grade placement, spe-cial classes for the educationally re-tarded and the academically gifted,neighborhood schools, the class-room as the instructional unit, theacademic curriculum, and even ourwhole system of education. Theremay be merit in some of these pro-posals. But I think they are too oftenaimed at covering up problemsrather than coming to grips withthem.

Greater Attention to LearningReadiness. The concept of develop-mental readiness for various kinds ofschool learning has been too ne-glected in recent educational trends,which have been dominated by theunproved notion that the earliersomething can be taught to a child,the better. Forced early learning,prior to some satisfactory level ofreadiness (which will differ marked-ly from one child to another), couldcause learning blocks which later onpractically defy remediation. Themore or less uniform lockstep se-quencing of educational experiencesmay have to be drastically modifiedfor the benefit of many children, butthe recent massive insistence on"earliness" and equality of educa-tional treatment of all children hasmilitated against large-scale researchon the implications of readiness forchildren with below-average edu-cability within the traditional schoolsystem.

Greater Diversity of Curriculaand Goals. Public schools, which aimto serve the entire population, mustmove beyond narrow conceptions ofscholastic achievement to find agreater diversity of ways for childrenover the entire range of abihties tobenefit from their schooling-tobenefit especially in ways that willbe to their advantage when they areout of school. The academic goals ofschooling are so ingrained in ourthinking and our values that it willprobably call for radical efforts tomodify public education in wayssuch that it will maximally benefitlarge numbers of children with verylimited aptitude for academicachievement. I believe that a well-intentioned but misconceived socialegalitarian ideology has preventedpublic education in the UnitedStates from facing up to this chal-lenge.

The belief that equality of edu-

cational opportunity should neces-sarily lead to equality of perfor-mance, I believe, is proving to be afalse hope. It is the responsibility ofscientific research in genetics,psychology, and education to deter-mine the basis for realistic solutionsto the problems of universal publiceducation. Though it may be pre-mature to prescribe at present, Iventure the prediction that futuresolutions will take the form not somuch of attempting to minimize dif-ferences in scholastic aptitudes andmotivation, but of creating a greaterdiversity of curricula, instructionalmethods, and educational goals andvalues that will make it possible forchildren ranging over a wider spec-trum of abilities and proclivitiesgenuinely to benefit from their yearsin school. The current Zeitgeist ofenvironmentalist equalitarianism hasall but completely stifled our think-ing along these lines. And I believethe magnitude and urgency of theproblem are such as to call for quiteradical thinking if the educational

eliminate human differences. Ratherthan making over a large segment ofthe school population so they willnot be doomed to failure in a largelyantiquated, elitist-oriented, educa-tional system which originally evolv-ed to serve only a relatively small seg-ment of society, the educationalsystem will have to be revamped inorder to benefit everyone who is re-quired by the society to attendschool. It seems incredible that asystem can still survive which virtu-ally guarantees frustration and fail-ure for a large proportion of thechildren it should intend to serve.

But we should not fail to recog-nize that to propose radical diversityin accord with individual differencesin abilities and interests, as con-trasted with uniformity of educa-tional treatment, puts society be-tween Scylla and Charybdis in termsof insuring for all individuals equal-ity of opportunity for the diversityof educational paths. The surest wayto maximize the benefits of school-ing to all individuals and at the same

them, if not in the usual academicsense, then in ways that can bettertheir chances for socially useful andself-fulfilling roles as adults. ECalifornia Institute of Technology En-gineering and Science magazine, Pasadena,California, April, 1970.

. You, Peabody, are one crummy Puritan!

system is truly to serve the whole ofsociety. We have invested so muchfor so long in trying to equalizescholastic performance that we havegiven little or no thought to findingways of diversifying schools to makethem rewarding to everyone whilenot attempting to equalize every-one's performance in a common cur-riculum. Recommendations havealmost always taken the form of ask-ing what next we might try to makechildren who in the present schoolsystem do not flourish academicallybecome more like those who do. Theemphasis has been more on changingchildren than on revamping thesystem. A philosophy of equaliza-tion, however laudable its ideals,cannot work if it is based on falsepremises, and no amount of propa-ganda can make it appear to work.Its failures will be forced upon every-one. Educational pluralism of somesort, encompassing a variety of verydifferent educational curricula andgoals, I think, will be the inevitableoutcome of the growing realizationthat the schools are not going to

time to make the most of a society'shuman resources is to insure equalityof educational opportunity for all itsmembers. Monolithic educationalgoals and uniformity of approachesguarantee unnecessary frustrationand defeat for many. On the otherhand, educational pluralism runs therisk that social, economic, ethnicbackground or geographic origin,rather than each child's own charac-teristics, might determine the edu-cational paths available to him. Theindividual characteristics appropri-ate for any one of a variety of educa-tional paths and goals are to befound everywhere, in every socialstratum, ethnic group, and neighbor-hood. Academic aptitudes and spe-cial talents should be cultivatedwherever they are found, and a wisesociety will take all possible mea-sures to insure this to the greatestpossible extent. At the same time,those who are poor in the traditionalacademic aptitudes cannot be left bythe wayside. Suitable means andgoals must be found for making theiryears of schooling rewarding to

contxnuvd from page 14

too far from familiar territory. Yet ifhe says he would go home he isjudged by the testers to be less intel-ligent. It is also easy to see fromthese examples why there is a cor-relation between test performanceand later performance in school be-cause the teacher, as a representativeof standard middle-class culture, willexpect the same language and typesof behavior as the person who madeup the so-called intelligence test. Theteacher will either not understandthe dialect that is used in class or willgive the child a lower mark for using"bad" language and the ghetto childwill before long go through life stig-matized as being less intelligent and apoor student.

Looking at the problem this wayforced psychologists to considerseriously another possible explana-tion for many of the existing correla-tions between intelligence testscores, doing well in school, andholding down higher-status jobs laterin life. Those who control not onlyeconomic and social opportunitiesbut also what language and values arethe standards by which others will bejudged, may in fact be able to usetest scores to maintain their power.All one needs to assume is that morepowerful families are in a much bet-ter position to help their children gethigher-status jobs: they know theright people; they can send their chil-dren to the right schools; they canuse their influence to get them jobsdirectly. So it turns out that peoplein higher-status jobs score higher onso-called intelligence tests.

But where is the direct evidencethat the higher score on the test infact indicated that the person wasbetter able to do the higher-statusjob? As every psychologist knows,correlation does not mean causation.It doesn't follow that because pro-fessionals score higher than laborerson certain tests that it is the abilityto perform those particular testswhich enabled them to be profes-sionals rather than laborers. Thereason why people have assumedthat causation was involved is thatthe test scores were supposed to indi-cate how intelligent the person was,and it seems reasonable to assumethat being a professional requiresmore of something called intelli-gence than being a laborer does.However it is by no means as self-evident as it once was that these testscores measure the kind of intelli-gence implied by the logic of this

Continued on page 154