The Bell Curve: Intelligence and Class Structure in American Life provides a provocative examination of the influence of intelligence in life outcomes and offers provocative suggestions for national policies, a number of which impact education, psychology, and the practices of school psychologists. This article responds to some of the issues discussed in the book: the nature of intelligence, its measurement, the importance of personal decisions in determining life outcomes, and the modifiability of intelligence in infants and young children. American psychologists, perhaps even the majority of Americans, have had a longstanding interest in the concept of intelligence and methods to measure it. Most people hold strong opinions about intelligence. Some might even characterize these opinions, when translated into behaviors, as symptomatic of a national bipolar disorder: persons often either respect the concept and hold it in high esteem or abjectly disregard and dislike it. Furthermore, many Americans, including psychologists and other professionals within the behavioral sciences, often hold strong opinions about the effects biological and environmental factors exert on growth and development. Those with more conservative views often believe a person's hereditary and biological qualities largely control a person's outcomes. Try as we might, the trajectories of one's life are pretty self-established by these qualities. In contrast, those with more liberal views often believe in the power of the environment; with effort and resources, persons can pull themselves up by their bootstraps. Opinions about intelligence and its origins, when combined with opinions on race and social class, often foster rancorous and heated debates that may lead to loss of friendships, allegations of racism or naivete, contested public policies, public riots, and even death threats. The issues can be contentious. Our nation has prided itself on enabling those who are tired, poor, and yearning to be free to have opportunities to better themselves through education, hard work, perseverance, and conservation. Nevertheless, we also believe many needy who immigrate with these qualities produce offspring who remain academically and financially marginal generation after generation. While the threads of social change and individual progress have been a mainstay of our social fabric, many believe these threads are frayed and our social fabric in need of weaving. Moreover, contrary to fact, many believe the majority of families who display low academic attainment and continuous poverty are African-American.
he Bell Curve (Herrnstein & Murray, 1994) revisits many of these issues along with others about which few people are neutral. As expected, its examination of differences in intellectual capacity among people and groups and what these differences mean for our country's future strikes at the heart of sensitive issues. The book's contents have been the subject of countless articles in newspapers and magazines, talk shows on radio and television, and have enlisted responses from political leaders including President Clinton.
Many ideas discussed in the book focus on issues central to school psychology theory and practice, including the importance of intelligence on educational attainment and other important life outcomes, allegations of test bias, accurate measurement of intelligence and other cognitive abilities, as well as racial differences in intelligence and their modifiability. Thus, reflections regarding possible implications The Bell Curve may have for the theory and practice of school psychology are important and timely. The information discussed in the book also is of keen interest to me professionally, given my prior research and writing on the nature of intelligence, the validity of using measures of intelligence with children of different racial-ethnic groups, and on issues important to the assessment of minority children.
I do not strongly disagree with the following five conclusions presented in the book: (a) all standardized tests of academic aptitude measure a general factor of intelligence to some degree; (b) measures of intelligence are fairly stable and reliable; (c) the measures are comparably valid for persons from different social, economic, and racial groups; (d) racial differences exist in intelligence; and (e) heredity accounts for between 40-80% of the variance associated with these measures. I believe these conclusions generally are supported by the prevailing evidence currently available from quality and respected scientific sources.
The nature of intelligence. My views about the nature of intelligence differ somewhat from those of the authors. They seemingly believe intelligence can be characterized exclusively as a general factor, a trait with such unity that one number accurately represents it. intelligence is more complex than this. Although I do not dispute the presence of a general factor of intelligence, this general factor accounts for less than one-half the variance associated with intelligence.
However, my views about the nature of intelligence are not restricted only to qualities measurable by intelligence tests. The concept of intelligence can be conceptualized in three forms: genotype, phenotype, and tested. Think of these forms as concentric circles, going from the largest, represented by the genotypic form, to the smallest and most narrow, represented by the tested form.
In its genotypic form, intelligence generally refers to an estimate of one's intellectual aptitudes which arise due to the person's biologically based qualities together with environmental opportunities and restrictions. For example, the statement that behavior is a function of genetic times environmental qualities is commonly used to succinctly express this viewpoint. In its phenotypic form, intelligence generally refers to the actual behaviors a person displays that reflect intelligence. These include applications of such qualities as memory, judgment, reasoning, abstract thinking, language, and ability to solve novel questions and problems. Intelligence, in its tested form, refers to one's performance on test items that assess the aforementioned qualities when administered under standardized conditions.
Returning to the analogy that these three forms of intelligence comprise three concentric and narrowing circles, one finds that passage from the first to the third results in conceptualizing intelligence in increasingly narrow and restricted ways. Simply put, intelligence as measured by tests is not synonymous with intelligence displayed phenotypically. Intelligence as measured by tests comprises an important yet small sample of qualifies that represent intelligence, mainly qualities that can be readily assessed through tests. For example, using Wechsler's definition that intelligence refers to the ability to act purposefully, think rationally, and deal effectively with one's environment, we find that measures of intelligence tell us little about a person's ability or willingness to act in a manner that shows purpose or a person's ability to effectively use their resources.
I do not dispute our ability to measure forms of intelligence in a reliable fashion and that the data from these measures predict various other qualities moderately well. However, intelligence in its phenotypic and tested forms is not isomorphic. Thus, we should not characterize a person's intelligence only on the basis of an IQ score nor assume such a score accurately predicts a narrow range of qualities within which a person must fall. I will return to this point shortly.
Measures of intelligence often are characterized as measures of aptitudes useful in making predictions about individuals or groups. For example, the academic and vocational futures of persons who score one standard deviation above the mean on an intelligence test are likely to be characterized differently from those who score one standard deviation below the mean. The first is described as having a higher aptitude than the second.
This interpretation rests on the assumption that measures of general intelligence accurately capture the aptitude-related and other personal qualities as well as those of the environment in sufficient number and detail that we can rely on one number, an IQ score, to make accurate predictions about individuals. I do not agree with this assumption.
Expanding our concept of aptitudes. Efforts are underway to discard this narrow notion of aptitudes (i.e., one that considers only one's general intelligence) with one that proposes a broader view (Snow, 1992). Emerging concepts of aptitudes consider temperament (Oakland, Glutting, & Horton, 1995) and personality, academic attainment, as well as physical abilities within the concept of aptitudes. Measures of general intellectual abilities provide some but not all information needed to chart future behaviors.
The importance of personal decisions. As previously noted, behavior generally is characterized as the interaction of a person's genetic (or more broadly conceived, biological) qualities and environmental opportunities and limitations. I believe these two sets of qualities are critical. However, this characterization is too narrow because it overlooks the importance of personal decisions.
Two children born of the same parents, raised in the same home, and who attend the same schools from kindergarten through 12th grade are likely to be very comparable in biological qualities (barring severe illness, accident, and other medical complications) as well as in environmental opportunities. Yet, not uncommonly, the two may take divergent paths through life, differing considerably in their marital status, levels of education, jobs, money earned during their lives, and other indices referred to by Herrnstein and Murray (1994).
People make various and different personal decisions that impact their futures. One decides to spend more time with friends while the other spends more time reading. One prefers to take each day as it comes while the other prepares for college or one's life work. One is concerned with exercising her or his right to be free and un-encumbered by externally imposed obligations while the other leads an organized life in which daily goals are met; the first person puts play before work while the second puts work before play.
The nature of the personal decisions people make about how to best utilize their biologically and environmentally based resources enable them not to be confined only by their biological qualities and environmental opportunities and resources. Persons have no opportunity to influence their genetic qualities, have a little opportunity to influence the nature of their biological qualities, have some opportunity to influence the nature of their environmental qualities, and have considerable opportunity to determine how their biological and environmental qualities are utilized.
For example, cognitive abilities, when properly developed and utilized, can have a decided impact on one's medical, social, psychological, and educational qualities. The use of cognitive abilities to help heal or ameliorate pain, cancer, and arthritis and other medical maladies generally is accepted (Rossi, 1986). Less well-known is research that indicates the process of learning strengthens connections between neurons and helps promote important individual differences (Kandel & Hawkins, 1992). The proper employment of learning strategies can assist in promoting independent learning and academic performance in students (Lenz, 1992). Successful strategies also exist to promote reading achievement among students with learning disabilities (e.g., Black, Oakland, Stanford, & Nussbaum, 1995; Swanson, 1991).
In these and countless other examples, persons engaged in learning activities designed to improve their command and use of their cognitive abilities demonstrate improvement in important life outcomes. We need to abandon the premise that cognitive abilities are unmodifiable. We should not think that human behaviors are controlled only by biological and environmental conditions over which we may have little self-control. The nature of personal decisions one makes to utilize their abilities and talents in an active and constructive fashion also critically impacts their life outcomes.
Two students with similar aptitudes may display very different qualities. These differences are likely to be attributable to three qualities, governed in part by personal decisions: knowing their strengths and weaknesses, utilizing this information well, and displaying persistence and motivation. More successful students are likely to know their strengths and weaknesses, to use their strengths and minimize their weaknesses, and to be highly motivated and persistent in their quest to achieve. In contrast, less successful students are unlikely to know their strengths and weaknesses, to disregard this information when making personal decisions, or to decide not to display the persistent drive and motivation needed to succeed. These ideas are somewhat consistent with those of Steinberg (Steinberg & Detterman, 1986) who proposes broadening the general concept of intelligence. He views intelligence as cognitive activities involved in goal directed adaptations which involve shaping and selecting important environments to help ensure success.
Attempts to develop measures of intelligence that fully assess Sternberg's model reportedly have not been successful. Some may criticize the viability of his model for this reason. However, the degree to which theories on intelligence are directly transferable to tests of intelligence should have little or no direct bearing on the theory's vi-ability. Its viability rests, in part, on the accumulation of data that suggest the theory to be robust yet succinct, descriptive, explanatory, and that enable us to improve decision making. We must resist efforts that impose a need for a test to be developed based on the theory as a litmus test of a theory's validity.
Herrnstein and Murray (1994) indirectly acknowledge the importance of personal decisions when they suggest that putting a library in a community without one could increase the intellectual level within the community. They go on to add, "but it may also spread out the range of scores by adding points to the IQs of the library users, who are likely to have been at the upper end of the [IQ] distribution to begin with" (p. 394). Although I do not agree with the outcome, their important point is the following: we may increase intelligence by providing beneficial resources and facilitating ways people make good decisions that result in their taking advantage of their resources.
Herrnstein and Murray make use of correlations to show relationships between two variables. For example, they report correlations between IQs and job performance ratings (.53) and years of education (.64). Although these correlations are high and statistically significant, they account for only 28 and 41% of common variance.
Intelligence and academic attainment. I previously examined relationships between the WISC-R IQs and reading comprehension concurrently and over a 3-year period (Oakland, 1983). The concurrent correlations were .71 for the total group, .73 for whites, .63 for African-Americans, and .64 for Hispanics. Although these correlations also are highly significant, they account for less than 54% of the variance associated with reading. Although important, intelligence, at best, tells only about one-half of the story. Correlations enable us to know something about the relationships between the variables and, in using the above correlations, typically lead us to conclude intelligence critically affects the attainment of some important qualities. The uncritical reader may assume correlations of .50 to .70 between intelligence and achievement allow us to determine levels of achievement with a high degree of accuracy once we know one's intelligence. However, while vital, considerably more information is needed to make accurate decisions about individuals.
Being unable to locate reliable information that estimates the range of academic achievement typically attainable given a particular narrow range of intelligence, I examined data on 476 children and youth in order to better understand this issue (Oakland, 1983). Children were grouped into one of four groups: those with WISC-R IQs between 83-87, 93-97, 103-107, and 113-117. Means and standard deviations were determined for these four groups on reading comprehension as measured by the California Achievement Test (Table 1).
As expected, in contrast to those students with higher levels of intelligence, students with lower levels of intelligence generally display lower levels of reading comprehension. However, the ranges of reading achievement within each of the four IQ groups are fairly broad and overlap with those in adjoining IQ groups. The range of reading comprehension percentile scores for two-thirds of the students in the lowest IQ group is between 7 and 49; the percentile ranges for those in the three higher groups are 19 through 67, 43 through 83, and 52 through 98, respectively.
A closer examination of scores reveals numbers of students in the lowest group display levels of reading comprehension that are at or higher than levels of reading comprehension displayed by students in the two higher IQ groups. For example, within the lowest IQ group, students display reading comprehension at the 47th percentile and beyond while, not uncommonly, students in the two higher IQ groups display reading comprehension at or below the 19th and 43rd percentiles, respectively. As with other important life outcomes, qualities in addition to intelligence strongly impact reading (Anderson, Hiebert, Scott, & Wilkinson, 1985; Barr, Kamil, Mosenthal, & Pearson, 1991; Singer & Ruddell, 1989).
One task facing school psychology is to broaden our understanding of aptitudes, not by discarding general intelligence, but by going beyond it to identify qualities that impact achievement, and then to discover methods that enable students to know and utilize their aptitudes well while either working to improve those that are deficient or finding means to minimize their impact on important life outcomes.
Differences in intelligence are related both to race and social class. An examination of previously reported data (Oakland, 1983) clearly indicate this trend. WISC-R means and standard deviations were determined for middle and lower class white, African-American, and Hispanic students. The following rank orders (with means and standard deviations) reflect the importance of both race and social class: white middle class (111, 13), African-American middle class (101, 15), Hispanic middle class (96, 13), white lower class (93, 14), and Hispanic and African-American lower class (88, 11). Important social class differences exist within each racial ethnic group: mean score differences between middle and lower class children are 18 for whites, 14 for African-Americans, and 9 for Hispanics. Thus, intellectual development is related to both race and social class.
Herrnstein and Murray (1994) note that, "changing cognitive [intellectual] abilities through environmental interventions has proved to be extraordinarily difficult" (p. 314). This passage elicits two responses. First. I do not believe the authors intended to imply that cognitive abilities do not change. Second, the implied conclusion that we cannot change the rate at which intellectual abilities change is somewhat overly stated.
Let me restate what I believe the authors intend to imply: changing cognitive abilities through environmental intervention is not difficult. For example, the impact of schooling on the attainment of knowledge and skills is obvious. However, changing the rate at which cognitive abilities develop has proved to be extraordinarily difficult. We know cognitive abilities improve and thus change with age, often through the mid-twenties for those qualities more directly influenced by cortical and neuropsychological abilities (e.g., fluid abilities) and possibly through the fifties for those qualifies more directly influenced by education and vocation (e.g., crystallized abilities). Thus, cognitive abilities change and can improve through events related to environment (Kaufman, 1990).
However, attempts to improve the rates at which cognitive development changes have been less successful. The terms mental age or learning age often are used to reflect rates of change. Persons whose cognitive abilities change at a rate similar to that of their peers are described as having an average rate of change (i.e., an average learning age). Those with abilities that change at a slower rate (or a faster rate) than their peers are described as having a below average (or an above average) learning rate. Thus, persons with IQs that lie at the 16th percentile (or the 84th percentile) typically are expected to maintain their relative placement and to continue to exhibit below (or above) average rates of change. Nevertheless, their cognitive abilities are advancing albeit at steady but different rates.
Considerable effort has been directed toward trying to alter the rates at which young children develop cognitively. This quest often translates into seeking changes in development quotients as reflected on measures of infant development or intelligence. Interventions are available that positively impact rates of development. However, most efforts have had little permanent impact. Herrnstein and Murray (1994) acknowledge the Abecedarian and Perry Preschool Projects, along with others, as examples of interventions that demonstrate success, only to find that gains children display related to the early interventions later dissipate after they leave the enriched program. Hunt (1961) summarizes much of the earlier work on this topic. In general, we find attempts to improve the rates at which cognitive abilities develop are most successful when interventions focus on infants and young children, in particular those with mild levels of developmental delays, and involve parents and professionals working in a consistent and coordinated fashion while the delays are more specific in nature, prior to their generalizing and adversely affecting other qualities.
I recently completed a 3-year study of the effectiveness of a home-based intervention program to improve the rate of development of children (from birth through 3) living in the Gaza Strip (Oakland & Abu Ghazaleh, in press). The information utilized the Portage Program which contains a series of more than 600 developmental tasks children are expected to acquire in the following areas: motor, communication, cognition, social, and self-help skills. Its implementation requires paraprofessionals to give mothers weekly in-home instruction about what and how to teach theft youngsters during the following week. Our experimental and control groups were matched on gender, age, pretest data, and social status. After 1 year in the program, the children in the experimental group performed near the 70th percentile while those in the control group performed near the 21st percentile. During the two succeeding years the children in the experimental group maintained their levels of development while those in the control group declined to the 4th percentile. A 1-year follow-up study is being conducted to determine if, after leaving the program, these differences are maintained.
I am not suggesting the use of the Portage Program under similar conditions within our country will lead to a similar level of improvement in the cognitive abilities of those zero to 3. However, this and other methods which show promise for advancing the development of very young children with mild developmental delays deserve further study. I disagree with the authors' statement that next to nothing is to be learned about how to raise IQ by research involving early intervention programs with infants (Herrnstein & Murray, 1994, p. 413).
School psychology's commitment to a scientific-practitioner model requires us to review evidence about the effects various qualities exert on growth and development in objective and dispassionate ways. Moreover, somewhat widely held and pessimistic views about prevailing social trends within our nation require us to conduct various forms of self-evaluations in an attempt to identify ways these Wends may be reversed. The information and positions advanced in The Bell Curve are somewhat useful to these ends by encouraging serious reflection and honest debate. In summary, for me, the book's contents reinforce my belief about the importance of intelligence on important life outcomes and the importance of better understanding the influences of personal decisions on these outcomes.
TABLE 1 Relationships Between IQ and Reading Comprehension for Four Groups of Students Ranges of WISC-R IQs Reading Comprehension (in percentiles) 83-87 93-97 103-107 113-117 M 28 43 64 80 SD 21 24 20 23
Anderson, R., Hiebert, E., Scott, J., & Wilkinson, I. (1984). Becoming a nation of readers. Washington, DC: National Institute of Education.
Barr, R., Kamil, M., Mosenthal, P, & Pearson, P D. (Ed.). (1991). Handbook of reading research: Volume II. New York: Longman.
Black, J., Oakland, T., Stanford, G. & Nussbaum, N. (1995). An evaluation of the Texas Scottish Rite Hospital dyslexic program. Paper presented to the annual meeting of the Orton Dyslexia Society, Houston, Texas.
Herrnstein, R., & Murray, C. (1994). The bell curve: Intelligence and class structure in American life. New York: The Free Press.
Hunt, J. Mcv. (1961). Intelligence and experience. New York: Ronald Press.
Kaufman, A. (1990). Assessing adolescent and adult intelligence. Boston: Allyn and Bacon, Inc.
Kandel, E., & Hawkins, R. (1992). The biological basis of learning and individuality. Scientific American, 267, 78-87.
Lenz, K. (1992). Self-managed learning strategy system for children and youth. School Psychology Review, 21, 211-228.
Oakland, T. & Abu Ghazaleh, H. (manuscript submitted). A primary prevention program for infants in Gaza. American Journal on Mental Retardation.
Oakland, T. (1983). Concurrent and predictive validity estimates for the WISC-R IQs and ELPs by racial-ethnic and SES groups. School Psychology Review, 12(1), 57-61.
Oakland, T., Glutting, J., & Horton, C. (1996). Student styles questionnaire. San Antonio, TX: The Psychological Corporation.
Rossi, E. L (1986). The psychobiology of mind body healing. New York: W. W. Norton & Company.
Singer, H., & Ruddell, R. (Eds.). (1985). Theoretical models and processes of reading. Newark, N J: International Reading Association.
Snow, R. (1992). Aptitude theory: Yesterday, today, and tomorrow. Educational Psychologist, 27, 5-32.
Sternberg, R., & Detterman, D. (Ed.). (1986). What is intelligence? Norwood, NJ: Ablex Publishing Corporation.
Swanson, H. L. (Ed.). (1991). Handbook on the assessment of learning disabilities. Austin, TX: Pro-Ed.