Featured Topic 11/12: Arthur R. Jensen 1923-2012

By Jonathan Kane

Until now, I’ve tried to avoid this group having posts related to psychometrics two months in a row, because it’s a secondary topic compared to this group’s main topic of evolution.  But last month, there has been an unusual situation that I think requires an exception to this rule.  Our featured topic last month was an obituary for J. Philippe Rushton, the psychometrician and evolutionary psychologist who died on October 2nd.  And now, less than a month after Rushton’s death, there has been the death of another researcher in this field who was far more important than Rushton: The U. C. Berkeley psychologist Arthur R. Jensen, who I previously mentioned in this post from February.

To me, it is difficult to imagine it could have been a coincidence that Rushton and Jensen died only 20 days apart, especially considering there are many signs their professional relationship was a close one.  They collaborated on many peer-reviewed papers published over the past ten years, and what I believe to be the last paper Jensen ever wrote was a summary of Rushton’s contributions to the study of mental ability, published around six months before both of their deaths.  It seems reminiscent of how when a person loses their husband or wife of many decades, the widow or widower sometimes will follow them not long afterward.

I mentioned in last month’s post that most of the obituaries for Rushton seemingly existed only to disparage him, but Jensen is respected enough among psychologists that he’s mostly avoided this problem.  His obituaries in both The New York Times and The Los Angeles Times are in fact quite balanced, and also manage to get across how highly-regarded his research has been.  While this is fortunate, both of these obituaries contain another problem that I still find somewhat bothersome: both of them focus almost exclusively on Jensen’s research about IQ in relation to race.  This is the aspect of Jensen’s research that’s always received the greatest amount of media attention, but it is only one of many aspects of intelligence he studied, and probably not the most important one.  Just as my post about Rushton attempted to fill the void left by most of his obituaries, my goal here will be to summarize all of the major topics of research that Jensen is known for, and how his research about race and IQ fit into the broader context of his work.

An educational psychologist by training, Jensen devoted much of his early research to the nature of learning, and the most effective way for schools to educate their students.  One of his most important early discoveries in this area, described in this paper, is that not all teaching methods work equally well for everyone.  Some students are perfectly capable of mastering social interaction, and have no difficulty memorizing information such as new vocabulary words, yet struggle with tasks that require them to form new inferences and conclusions.  Jensen coined the term “level 1” for the ability to perform the first type of task, and “level 2” for the second.  He suggested that because the students who excel at the first type of learning do not always excel at the second, education might be more effective if its methods were tailored to each student’s type of abilities, rather using the same methods for everyone.

Like many other researchers of the 1950s and 1960s, Jensen initially assumed that differences in ability were due entirely to cultural factors, and had researched what sorts of environmental interventions might be able to erase them.  Due to his existing research in this field, in 1968 the journal Harvard Educational Review asked him to write a paper offering his opinion about the Coleman Report, a report about the compensatory education program Operation Head Start.  Among many other conclusions, the Coleman Report stated that this program had not succeeded at producing the gains in IQ and scholastic achievement that had been hoped for.  HER particularly wanted Jensen’s opinion about why Operation Head Start had not succeeded at this, as well as the reason why African-American students tended to over-represented among those in need of compensatory education.

Although Jensen’s article for HER covered more or less exactly what they asked him to cover, in retrospect it is apparent the editors of the journal got more than they bargained for.  Over the course of researching how to best help those students regarded as culturally disadvantaged, Jensen had recently come to the conclusion that most researchers writing about this topic were giving short shrift to the possible role of genetic factors.  Jensen’s article, published in February 1969 under the title “How Much Can We Boost IQ and Achievement?” concluded that the reason why Operation Head Start had not produced the hoped-for gains in IQ was because genetics placed a limit on how much IQ could be improved by cultural factors.  Even more controversially, his paper suggested that because individual differences in IQ were known to be highly heritable, and because no environmental intervention had yet been successful at significantly reducing the gap in average IQ between black and white students, it would be parsimonious to conclude that these group IQ differences involved genetic as well as environmental factors.

While Jensen’s paper was mostly about the effects and heritability of IQ in general, his conclusions about IQ in relation to race were by far what received the most attention, especially from the media and from student protestors.  It is widely regarded as the most controversial paper in the history of American psychology, and was largely responsible for initiating the modern debate over race and intelligence.  As a result of this paper, the term “Jensenism” entered the public vocabulary as a term for Jensen’s theory that genetics play a significant role in both individual and group differences in IQ.  This was initially a term of disparagement, but just as the term “queer” has eventually become a neutral or positive term among members of the LGBT community, the term “Jensenism” now is sometimes used by researchers as a way to show they are not afraid to support Jensen’s conclusions.

Before the publication of his 1969 HER paper, the relationship between IQ and race had not been a major part of Jensen’s research.  But in the face of the controversy arising from this aspect of his paper, he decided to formally present his theory about race and IQ as a scientific hypothesis in his 1973 book Educability and Group Differences.  The most important point of Jensen’s argument rests on the mathematical relationship between the heritability of a trait between individuals, and the degree to which heredity accounts for differences between group averages.  It is incorrect to assume, for either IQ or any other trait, that high heritability among individuals requires between-group heritability to be equally high.  However, within-group heritability places certain limitations on what types of environmental factors are capable of causing between-group differences, and Jensen’s argument was that for group differences in average IQ, there were no available environmental explanations which met the necessary criteria.

One of the most popular early criticisms of Jensen’s research in this area was the assertion that any apparent difference in average IQ between ethnic groups was only an artifact of the testing methods used, because IQ tests were systematically underestimating the ability of minority groups.  Jensen began to focus on this idea in particular in the mid-1970s, eventually leading to the publication of his 1980 book Bias in Mental Testing.  This book did not attempt to address the nature/nurture question of the cause of group differences, but made the point that evidence did not support the claim that the tests themselves were unfair.  One important fact about IQ tests was that they are very good predictors of real-life outcomes such as scholastic achievement and job performance.  Jensen argued that if IQ tests were underestimating the ability of minority students, they would also consistently underestimate these students' performance in school or on the job, while in fact IQ tests predict these real-life outcomes equally well for all groups of native English speakers in the United States.  Although Jensen’s conclusion that group IQ differences contain a genetic component has yet to enter the scientific mainstream, the conclusion that test bias is not adequate to explain them is now widely-accepted in the psychology community, and has been endorsed by a 1982 report from the National Academy of Sciences as well as a 1996 report from the American Psychological Association.

It was easy enough to measure what things IQ did or did not correlate with, but by the 1980s Jensen has become interested in a more fundamental question.  The question was, what is the actual nature of the trait that IQ tests are measuring?  In the 1960s, Jensen had not quite known the answer to this—in his 1969 HER paper he had remarked “Intelligence, like electricity, is easier to measure than to define.”  In search of an answer to this question, Jensen turned to the work of a much earlier researcher: the British psychologist Charles Spearman.

One of Spearman’s most important discoveries was that when a person’s performance is compared across a wide variety of different types of intelligence tests, there is a substantial positive correlation in how well the person performs on all of them.  The type of thinking required to answer a reading comprehension question is different from the type required to solve a math problem, but more often than not a person who does well at the first type of task will also have above-average ability at the second.  Spearman was a pioneer of a mathematical technique known as factor analysis, which can be used to find the factors responsible for variation among a number of data points.  Applying this method to scores on various types of intelligence tests, Spearman found that a person's performance on each type of test could be explained by the interaction of two factors.  One was an ability specific to each type of task, which Spearman termed s, and the other was a general factor which affected performance on all mental tasks, which Spearman termed g.  As Jensen’s research progressed, Spearman’s g factor became steadily more important in it.

A common criticism of g is that that because factor analysis does not specify how many factors should result from such an analysis, there is no way to prove that a central factor of g is the correct model to explain variance in intelligence test scores.  Jensen’s argument for the importance of the g model centers around one of the most important characteristics of any scientific theory: predictive validity.  He suggested that Spearman’s model has a greater predictive validity than any other theory of intelligence, in two separate ways.  First, g predicts life outcomes such as scholastic achievement and job performance even more accurately than IQ does.  In this respect, g is sometimes considered the “active ingredient” of IQ tests—that is, the degree to which IQ tests predict these outcomes is based on the degree to which they are measuring g.  A second type of predictive validity is that g correlates with certain biological variables, such as overall brain size and the brain’s efficiency of glucose metabolism, as well as that the intelligence tests which are most strongly measuring g tend to be those on which performance is the most heritable.  Jensen presented his argument for the importance of g in his 1998 book The g Factor.

Jensen’s research on g and its correlates provided him with a new insight into his earlier research on the difference between level-1 and level-2 tasks.  The reason why a person who is good at level-1 tasks will not necessarily be good at level-2 tasks is because performing level-1 tasks depends on an ability specific to that type of test, while ability to perform level-2 tasks is more heavily influenced by g.  Jensen coined the term “g-loading” to describe how strongly g affects a person’s ability to perform a particular task.  An example of the difference between a level-1 task and a level-2 task is the difference between reciting a sequence of digits that’s been read aloud to you, and reciting the same sequence of digits in reverse.  The first task is primarily a test of short-term memory, but the second ability requires one to think harder and thus is more g-loaded.

This line of research enabled Jensen to answer a question about race and IQ he had long been puzzled by, which is why group differences in test performances were larger on some types of test than on others.  Jensen found that the most strongly g-loaded tests tended to be those on which group IQ differences were the largest.  In this paper, he proposed that the reason for this correlation was because group differences in average IQ in fact consisted mostly of differences in g.  He termed this theory “Spearman’s Hypothesis”, as Charles Spearman had suggested the idea decades earlier, although Jensen was the first person to empirically test it.  Jensen’s colleague J. Philippe Rushton has suggested that when a strong correlation exists between a test’s g loading and some other variable, this should be called a “Jensen effect”, although this term hasn’t quite caught on outside of the psychology community the way “Jensenism” has.

As Jensen researched the g-loadings of various types of tests, he discovered something else unusual.  The g factor of mental ability not only predicts one’s ability to accomplish various mental tasks, it also predicts the speed at which one can accomplish them.  And in fact, this correlation is strongest for very simple tasks that anyone can accomplish in a few seconds, such as adding together a pair of one-digit numbers or pressing a button when one sees it light up.  The reason why this correlation exists is not known for certain, but it seems that measuring the brain’s speed of information processing taps into some of the same biological variables that influence intelligence as measured on traditional IQ tests, such as processing efficiency and nerve conduction velocity.

Like g itself, the idea of reaction speed as correlating with intelligence did not originate with Jensen, but he is largely responsible for the importance placed on it in modern psychometrics.  Jensen has proposed that tests of reaction speed have the potential to be even more useful than traditional IQ tests, because while IQ tests can only rank people’s abilities against one another based on who is or isn’t able to answer certain test items, reaction speed can provide an absolute ranking of a person’s ability in milliseconds.  Jensen’s research in this field has resulted in a third thing being named after him, after “Jensenism” and “Jensen effect”: he is the inventor of what’s now known as a Jensen box, a device which measures reaction speed based on how quickly the test subject presses any of several buttons when lights go on next to them.  This field, which is known as mental chronometry, is the subject of Jensen’s 2006 book Clocking the Mind.  Clocking the Mind was Jensen's last major book, although he continued periodically writing peer-reviewed papers until earlier this year.

The importance of Jensen’s research in modern psychology is easy to underestimate if one pays attention to only his research about IQ in relation to race, but other psychologists generally don’t make this mistake.  An article in Review of General Psychology lists him among the 100 most eminent psychologists of the 20th century, in 47th place.  He received the 2003 Kistler Prize for adding to the understanding of the relationship between genetics and society, and in 2006 received the lifetime achievement award from the International Society for Intelligence Research.  At least a half-dozen books have been devoted to discussing him and his views, and in 1998 a special issue of the peer-reviewed journal Intelligence was devoted to honoring him, with the title “A King Among Men: Arthur Jensen”.

Yet while Jensen was among one of the most admired psychologists of the 20th century, he was also among the most hated.  Due to threats against his life, at several points it was necessary for his mail to be opened by a bomb squad, and for him to be accompanied by a bodyguard while traveling to and from the classes he taught at Berkeley; on other occasions he had to literally flee from angry mobs for his physical safety.  The reactions and threats against him from protestors are described in more detail on pages 343 to 347 in this paper.  Equally significant are the reactions of other academics when he first presented his ideas about race and IQ, such as a 1970 statement by the American Anthropological Association describing him as a “chauvinist, biased racist”, and a 1973 statement in the New York Times signed by 1,000 academics saying that his work did not deserve any protection in the name of academic freedom.  Contrary to Harvard Educational Review’s usual policies, Jensen was initially not allowed to purchase reprints of his famous 1969 article, or to reply to critiques of his paper published in that journal.

Ultimately, the question of whether or not a person is a racist boils down to their motives and morals, and Jensen articulated his own fairly clearly.  Although his research was sometimes abused by pro-segregation groups in the 1970s, Jensen himself was an opponent of racial segregation, and tended to react with hostility when these groups sought his support.  As he explained in this paper, he believed that education methods should be based on students’ individual abilities regardless of what group the belonged to, and his goal was to help those students who did not receive a good education because of their schools failing to do this.  On the question of whether or not Jensen is a racist, one of his most vigorous defenders is in fact the person who was his prominent opponent in debates over race and IQ: James Flynn, after whom the Flynn Effect is named.

More than anything else, what one thinks of Jensen’s research depends on where one’s priorities lie.  There’s no question that he contributed a great deal to our understanding of the nature of human mental ability, and his obituaries in both the New York Times and the Los Angeles times point that out, although they don’t go into as much detail about those contributions as I’d like.   But on the other hand, his discoveries include a lot of things that many people probably wish weren’t the case.  Among its conclusions are that many important life outcomes are heavily influenced by a single variable of mental ability; that this variable is highly heritable; and that it is not distributed uniformly between ethnic groups, even if it remains an unanswered question whether or not these group differences involve a genetic component.  When someone’s message includes ideas like this, for many people it’s hard to avoid wanting to shoot the messenger.

To a science-minded individual, though, one would hope that knowledge is valuable regardless of how unpleasant it might be, especially when understanding the world’s unpleasant realities can assist us in helping people who are less fortunate than us.  I also have a special admiration for people who help us to see when our comforting misconceptions about the world are without foundation, whether about natural history or about human abilities.  That is what Arthur Jensen was: the indefatigable g-man, reminding everyone to wake up and smell the ashes.