Testing versus understanding human intelligence

Listen Pause Loading

American Accent British Accent Australian Accent Slow Reading Speed Medium Reading Speed Fast Reading Speed Download MP3 Help

Acknowledgement: Elizabeth J. Austin is now at the Department of Psychology, University of Edinburgh.

Correspondence concerning this article should be addressed to: Ian J. Deary, Department of Psychology, University of Edinburgh, 7 George Square, Edinburgh, EH8 9JZScotlandUnited Kingdom Electronic Mail may be sent to: i.deary@ed.ac.uk.

Psychometric tests of intelligence have high predictive validity and are therefore valued tools in educational and occupational psychology. Nonetheless, they are frequently subject to the criticisms of not explaining intelligence differences and of tautology (cf. Edwin Boring's dictum that intelligence is what the tests test). Current intelligence assessment tools may also be criticized because the lack of understanding of the biology of intelligence means that the units in which intelligence is measured are not well defined. The essence of all of these difficulties is that the assessment and the understanding of human intelligence have developed at different rates, a situation not unknown in the history of other sciences.

The ultimate resolution of these problems must be the establishing of links between test scores and relevant biological indices. Such measures would in addition have the advantage of not being subject to the criticisms sometimes leveled against IQ tests of being biased against certain groups.

A number of putative indicators of the biological bases of intelligence differences has been found. Measures such as inspection time are said to be related to an individual's speed of basic information processing and are found also to be associated with psychometric intelligence. Other physical measures such as the complexity of electroencephalogram (EEG) waveforms and brain metabolic rate give a less consistent picture. Given the provisional level of understanding of these phenomena and the cost and difficulty of performing the relevant measurements, physical tests of intelligence are unlikely to replace IQ tests in the foreseeable future.

Measuring Versus Understanding Physiological Systems

The heart pumps blood around the body. The lungs allow oxygenation of the blood and the escape of carbon dioxide and water from the body to the outside world. The pancreas produces insulin to aid glucose uptake into cells and secretes enzymes to digest nutrients in the gastrointestinal tract. The voluntary muscular system responds involuntarily to keep trunk and limbs in a chosen posture despite external disturbances and voluntarily to produce complex programmed intentional changes in posture, for example, a golf swing. For these and other organs and systems, biological science has progressed on two fronts. (a) Applied science offers diagnostic tests that indicate the functional integrity of the organ in question. That is, some aspect of performance may be measured and used to predict the quality of functioning of the organ (or system) with regard to some aspects of real life. For example, an electrocardiogram indicates whether the heart can cope with exercise without dysfunction. (b) Basic science offers a reductionist account of the functioning of a whole organ. For example, we understand a lot about how the kidney performs its functions, from the level of filtration units (nephrons) down to that of the molecular transporter systems on the surface of specialized kidney cells.

According to a once-popular classification, the brain thinks, feels, and wills. In a similar fashion to the investigation of other organs, we find within differential psychology a twin-track development of (a) the methodology for testing the brain's putative principal functions and (b) an account of the brain's functional integrity at different levels of analysis. The main topic of this piece is that broad function of the brain that we call human intelligence. On the one hand, there is a huge industry surrounding intelligence testing, and on the other, there is an overlapping effort aimed at understanding the nature of human intelligence differences at various levels of description. Figure 1 shows that the structure of individual differences in human abilities has been well described using psychometric tests. Individual differences in cognitive functions are to a degree general and separable, and a three-level hierarchy of mental abilities appears at present to be the favored descriptive framework (Carroll, 1993). At the peak of the hierarchy, it makes sense to talk about a person's general intelligence, or differences between persons in general intelligence, as was discovered by Spearman (1904). At the bottom of the hierarchy, it also makes sense to inquire about differences between people in very specific mental abilities. There is no contradiction in recognizing both general and specific mental abilities as long as one bears in mind the fact that mental abilities below the level of general intelligence are positively correlated. The measurement of cognitive ability differences is validated by the results of psychometric tests being predictive of real-life outcomes, such as in education, work, and in life more generally (Herrnstein & Murray, 1994; Neisser et al., 1996). In this regard mental testing follows the tradition of Binet in using tests of higher-level mental functioning to predict wider aspects of the person's mental performance. This psychometric account of intelligence differences is successful, therefore, but scientifically unsatisfying because it is not rooted in a theory of the causes of intelligence differences. Therefore, psychologists since Galton (1883) and Spearman (1904) have investigated whether there are simpler psychological (or lower biological level), cognitive, or physiological brain functions in which individual differences can be found that account for mental ability differences found at the psychometric level.

Figure 1. A three-level structure of human psychometric intelligence. Specific abilities are represented at Level 1, group factors at Level 2, and general intelligence (Spearman's g) at Level 3 (Gustafsson, 1984 ). Ms = memory span, Cs = speed of closure, CFR = configuration of figural relations, I = induction, V = verbal comprehension, NumAch = mathematical achievement, VeAch = verbal achievement, Vz = visualization, S = spatial orientation, Cf = flexibility of closure, Gf = general fluid ability, Gc = general crystallized ability, and Gv = general visualization ability

The attractions of such an approach are twofold. First, it might free the construct of intelligence and intelligence testing from accusations of tautology. Second, it might root the intelligence testing movement in a theory of the brain's limitations in performing certain basic operations, thereby offering a partial explanation of what it is to be more or less intelligent.

Psychometric Intelligence Tests Cannot Explain What It Means to Be Intelligent

We examine first the limitations of a purely psychometric approach to mental ability testing. Given the complex and poorly understood chain of events between cause (events in the brain) and effect (response to a psychometric item), it is necessary to have evidence that the measurement process involved in mental testing is meaningful. Seashore (1902) emphasized the need to found mental tests in terms of a model of mental functions and at the same time defended psychometric measurements as valid:

An adequate set of tests of normal functional efficiency … is a great desideratum for the present-day needs, and an end by no means beyond the goal of properly directed endeavour. Its starting point is a correct analysis of the most distinctive modes of exercise of the several elementary components of our mental functions; the second step is the devising of tests that shall most simply, naturally, and definitively measure the functional efficiency of a selected factor or process … for a host of comparative purposes the determination of norms or standards of functional mental efficiency is indispensable. That such determination involves conventions and artificialities is true and proper and inevitable. But neither is a foot, nor a meter, nor a candle-power, nor a volt, nor an ohm a natural and predestined ding-an-sich. Yet the arbitrary and conventional character of these units does not interfere with their utility. (p. 1)

Michell (1997) used similar metaphors to undermine, rather than support, the validity of psychometric tests. His argument was that invalidity arises with respect to psychometric tests because the measures used lack the interval scale property invariably found in scales for physical measurements such as length or time. In fact it is questionable that a measurement scale must have interval properties in order to be valid or useful. An interval scale is one where a unit of measurement can be defined so that differences between measured quantities have meaning. For example, it is meaningful to state that two objects differ in length by 2 inches regardless of what their sizes or other attributes are; it is problematic to make an equally strong statement regarding differences between two people's scores on an ability scale. It is well known in social science that many measurement scales may have only the weaker ordinal scale property, for which an ordering of participants on a construct dimension can be established, but absolute differences in scores cannot be interpreted (see, e.g., Blalock, 1972). Ordinal scales can nonetheless be used as a valid basis for inference. Michell is, however, correct to point out that social scientists tend to use statistical techniques valid only for interval scales or ordinal data. Although this statistical misuse is a potentially serious issue, the essence of his criticism of psychometrics—that any scale which cannot be proved to have interval properties fails merely because it lacks the power and refinement of the measuring instruments of physical science—is not well founded.

To judge the validity of current psychometric measures, and to place criticisms of these in context, consider why measurement and prediction is more problematic in psychology than in, for example, the classical physics of stable nonturbulent motion. One reason is the difference in magnitude and meaning of errors in the social and biological sciences. The large errors that are unavoidable because of the intrinsic nonstationarity and variability of living systems contrast with the tiny experimental errors of physics. Statistics is used extensively by biological and social scientists but rarely by physicists. Compared with well-defined manifest biological qualities (e.g., blood pressure), estimation in psychometrics presents additional difficulties because it involves latent constructs such as intelligence, personality traits, and attitudes. The validity of latent constructs can, however, be demonstrated: Intelligence has a well-understood structure and impressive predictive and construct validity (Neisser et al., 1996), as we discussed above; also, broad personality traits are highly robust between studies, methods, and cultures (Matthews & Deary, 1998). Nonetheless, development of units of measurement for these constructs cannot be regarded as finalized. It is therefore encouraging to recall that prediction of the weather or turbulent fluid flow still challenges physicists (Gleick, 1987) and that historically even physicists have found measurement problematic. The history of physics does not reveal an orderly process in which measuring scales were first devised, then validated, then used. Instead, the development of measurement in physical science demonstrates a synergy between experiment and theory (Kuhn, 1961), not an orderly linear progression. Scales of length (Figure 2) developed from (a) the use of a biological reference standard with large errors, through (b) use as reference of the length of a standard bar or a fixed fraction of the Earth's circumference, to (c) a standard based on the distance travelled by light in a given time. The discovery and characterization of fundamental constants such as the velocity of light was of key importance in the refinement of physical measurement scales (Cook, 1994; Petley, 1985). Figure 2 suggests some analogous points in the development of the measurement of intelligence, with assessments by teachers (Spearman, 1904) roughly comparable in accuracy to early length measures and current instruments at the second stage, with population IQ standard deviation playing a role analogous to that of the standard bar. The third stage awaits further understanding of the biological basis of constructs such as intelligence and personality. In the case of intelligence, it appears likely that measures based on information-processing capacity will be developed.

Figure 2. A schematic representation of some parallels between the development of measures for length and intelligence. The later dates for the standard bar refer to the use of controlled conditions of temperature

Biological Bases of Intelligence: The Saviour of Psychometrics?

Will an appropriate unit of information processing provide a fundamental constant of psychology? The vision is simple and attractive: A person's potential brain efficiency will be construed in the context of a biological theory of cognitive performance parameters, and basic limitations on cognitive functioning will be assessed using physiological techniques. Mental functions will thus be assessed using the units of natural science, not by problematic IQ-type scales. The limitations of human cognitive performance thus measured will predict a substantial part of the variance in human psychometric intelligence differences. With such a vision in mind, Vickers and Smith (1986) argued as follows:

One major strategy guiding attempts to measure speed of mental functioning has been to isolate some process sufficiently elementary to be relatively immune from influence by higher cognitive activities or by motivational and social factors. In its focus on a simple, component process, likely to play a limiting role in most (if not all) more complex processes, this strategy resembles the employment of standard algorithms as benchmark tests of processing speed of a digital computer. (p. 609)

How far this strategy—on the basis of the brain's fundamental processing limitations—would take us from the allegedly rootless, atheoretical approach of the journeyman psychometrician! From the beginnings of formal mental testing, there has been tension between the need for practically valid tests and the demand for an explanatory account of human intelligence differences. Whereas Seashore (1902)—see above—and Spearman (1923) assumed that the latter would precede the former, the history of progress in the physical sciences suggests that the two efforts of assessing versus understanding human intelligence are interdependent. One may therefore debate whether it is surprising that the testing of mental functions has been achieved long before their understanding in biological–information processing terms. One of the first batteries of psychometric tests of largely mental functions was proposed by McKeen Cattell (1890), who introduced 10 tests and 50 more detailed tests with the minimum of theoretical background:

Psychology cannot attain the certainty and exactness of the physical sciences, unless it rests on a foundation of experiment and measurement. A step in this direction could be made by applying a series of mental tests and measurements to a large number of individuals. The results would be of considerable scientific value in discovering the constancy of mental processes, their independence, and their variation under different circumstances. Individuals, besides, would find their tests interesting, and perhaps, useful in regard to training, mode of life or indication of disease. (p. 373)

This seminal article contained little more than a list of some suggested tests. It was not asked why the tests were to be used and on what model of mental functioning they were to be based. Of course, the tests must be based on some implicit theory of mental functions, but it was not expounded. Galton (1890) made a series of remarks in reply to McKeen Cattell (1890). The bulk of these are details of test apparatus, but Galton opened thus,

One of the most important objects of measurement is hardly if at all alluded to here and should be emphasised. It is to obtain a general knowledge of the capacities of a man by sinking shafts, as it were, at a few critical points. In order to ascertain the best points for the purpose, the sets of measures should be compared with an independent estimate of the man's powers. (1890, p. 380)

It became something of a London School tradition to be unsatisfied with a merely psychometric frame of reference for human intelligence differences and to try to root test results in more basic psychological functioning. Spearman, in his various writings (1904, 1923, 1927), came up with three different—though not exclusive—sources of differences in human intelligence. These were (a) the three fundamental psychological processes of apprehension of experience, eduction of relations, and eduction of correlates; (b) sensory discrimination (also favored by Galton [Deary, 1994]); and (c) a general physiological energy. Spearman's levels of description sweep across a broad reductionistic range. Spearman's dissatisfaction at the construct of intelligence's not being rooted in theory was picked up by his graduate students, one of whom wrote of Binet's successful tests, “They do not know what these tests measure or signify … . The tests are isolated from the main body of scientific psychology. They neither derive much light from it, nor do they impart much to it” (Abelson, 1911, p. 269).

Contemporary Work on Information Processing and Intelligence

The search for basic information-processing/physiological measures, which would predict a substantial part of variance in intelligence, has motivated recent research (Brody, 1992). One of the simplest hopes was that intelligence could be understood as mental speed and linked to reaction times for elementary or complex decisions, peripheral nerve conduction velocities, and the latencies of event-related components of brain electrical activity. Measures of speed of simple perceptual judgments in inspection time and rapid visual-processing tasks (Deary, McCrimmon, & Bradshaw, 1997; Deary & Stough, 1996) may explain up to 25% of variance in psychometric intelligence test scores. Reaction time for elementary decisions, and variability of reaction time, are also related to intelligence (Jensen, 1993) but somewhat less strongly. Variation in strategies used in reaction-time tasks, such as trading accuracy for speed, may partly account for the weaker relationship, but it has been noted that the important parameters in psychological models of information processing are unimportant in predicting intelligence, whereas parameters that predict it have no special function in these models (Lohman, 1994).

Brain event-related potentials (ERPs) are waveforms recorded using EEG technology that yield indirect information about the detailed time course of stimulus processing and decision making. Early attempts to link ERPs to intelligence were flawed technically, but recent work has repeatedly linked lower intelligence with longer latency to particular landmarks on the waveform (Deary & Caryl, 1997); smart peoples' brains do work faster. However, differences in speed of perceptual intake, also linked to intelligence, have been related in several laboratories (Deary & Caryl, 1997) to differences in the shape of the ERP waveform (reflecting differences in the orchestration of early stimulus analysis rather than in its speed); smart people also operate their neural machinery differently. Links between intelligence and latency of one particular ERP landmark (P3 or P300) are well established. Whether or not a strong relationship is observed in these studies depends on the exact task and on what group is considered, for example, Alzheimer's disease patients or their control group (Deary & Caryl, 1997). There is as yet no theory to explain which task or task parameters to choose for a particular group. Studies of complexity of brain electrical waveforms have linked both higher and lower complexity, under difference conditions, with high intelligence. Brain imaging studies, linking brain metabolism to intelligence, yield a similarly ambiguous outcome: Higher intelligence may be associated with higher or lower metabolic rate depending on whether the image is scanned during relaxation or while performing a mental task (Deary & Caryl, 1997).

The straightforward mental ability changes associated with ageing—slowing of responses and a drop in fluid abilities (e.g., mental calculation) while crystallized abilities (e.g., vocabulary) are maintained—have been linked to increases in reaction time and in ERP latencies with age. The hidden complexities of this example hint that a simple psychobiological explanation of intelligence differences is still a distant prospect. As task complexity increases, reaction time and ERP latencies in young and elderly adults increase differentially (Bashore, Osman, & Heffley, 1989) because age slows some brain processes but also affects the strategies that participants adopt. The results support each of two competing models of age-related slowing (Cerella, 1985)—hinting that better understanding of biological mechanisms will not necessarily yield a unitary biological measure of the processes involved!

Replacing Psychometric Tests With Tests of Basic Psychological Processes?

With each discovery that a putative basic function correlates with psychometric intelligence, there has been a flurry of speculation as to whether old-style IQ-type tests will be replaced with a more scientifically respectable measure. Typically, the benefits are thought to be culture fairness, repeated testing (fewer practice effects related to content of the answers), and objectivity. Thus, psychophysical tests, ERPs, functional brain scanning tests, and other advances have been proclaimed as the new face of psychological testing (e.g., Brody, 1992; Matarazzo, 1992). These claims are naive. Such tests usually involve highly specialized equipment, are available in only a few laboratories, and require expertise to administer. The procedures used in the tests, and the variables derived, change regularly as psychological theories evolve and new equipment and techniques become available. The repeatable associations between experimental measures and psychometric intelligence scores are modest at best, and there is no evidence to indicate that any experimental test would predict real-life variance better than a psychometric test or battery. In many if not most cases of associations between information-processing measures and intelligence, there has been insufficient research to establish the direction of causation between the supposedly lower level and the psychometric measure (for an exception, see Deary, 1995).

Perhaps it is just too tempting, when considering intelligence, to make reductionist analogies with simple physical measures such as length or temperature, or in biology with an organ such as the heart that is highly specialized to perform one function brilliantly. As anyone will appreciate who has driven through traffic while holding a conversation, learning the route through an unfamiliar city, and wondering what next week will bring, the brain performs multiple functions in parallel. A better analogy than with the heart muscle might be with a notional golf-swing muscle system. At present, simple biological measures to diagnose the cause(s) of a faulty golf swing or of poor intellectual performance are equally distant prospects; the golf course or the intelligence test battery yield clearer assessments than the psychology or physiology laboratory. However, of course, to understand mechanisms one does need the laboratory.

The great success of psychometric tests is their predictive validity combined with their technological simplicity and ease of administration. In his first published article, Burt (1909–1910) appreciated that to understand intelligence, it was necessary to use tests of simple cognitive functions, whereas tests of higher-level mental functions offer better prediction. In Burt's (1909–1910) study, Binet-type tests had better correlations with imputed intelligence than Spearman-type tests of sensory discrimination:

The main significance of this hierarchy of experimental performances, is, as it appears to me, that we are led to infer that all of the functions of the human mind, the simplest and the most complicated alike, are probably processes within a single system. A process typical of higher psychophysical ‘levels’ may be connected with a process typical of lower psychophysical ‘levels’ far less intimately than either is with a process of intermediate ‘levels’. Yet this relatively small correlation is not a disproof, but a consequence of, their inclusive organisation within a single integrative system of psychical dispositions or neural arcs. (p. 164)

The most fundamental objection to replacing psychometric tests with tests of basic processes in the foreseeable future is that there is little agreement as to what the basic processes are and the level at which they are best conceptualized. The information-processing enterprise is premised on there being a stable, tractable account of relevant psychological processing. Psychologists are still rewriting these accounts. In the meantime, rapid advances have been made in molecular genetics, which allows psychologists to study the associations between phenotypes and specific gene actions. Two examples make it clear that the likelihood of explaining some of the variance in human intelligence differences in biological terms is increasing. First, differences in a DNA marker in the gene for insulin-like growth factor-2 receptor are related to psychometric intelligence test scores in children (Chorney et al., 1998). Second, possession of a certain allele of the apolipoprotein E gene makes one more susceptible to dementia and to cognitive decrements following environmental brain insults such as head injury and cardiac bypass surgery (Schmechel et al., 1993). From this type of research, alternative mechanistic pathways for understanding human intelligence differences and their changes with age and in the face of insults to the brain may begin to appear at the biological level and perhaps obviate the requirement to express aspects of brain function at any higher psychological level.

Conclusion

Crystal ball gazing yields guesses that are less likely to be right than wrong. Nevertheless, Matarazzo (1992) bravely attempted to predict developments in psychological testing in the 21st century. He opined that there would be few radical changes in the traditional everyday tests of intelligence (often called paper-and-pencil tests) but that a new range of validated tests of intelligence (based on physiological measures of nervous system functioning and on measures of information processing derived from the cognitive psychology laboratory) would be developed and brought into use in specialized assessments, if not in everyday practice. We have argued, using a historical perspective, that such new information-processing and physiological measures will not provide estimates of intelligence that are any better or less contentious than the traditional psychometric methods for assessing mental abilities. More important, the view that these approaches should provide tools for assessing people misrepresents their importance and promise. They offer, not alternative (more complex and more expensive) ways to assess human mental abilities and their interrelationships, but the prospect of a better and deeper understanding of the mechanisms underpinning abilities and ultimately the causes of human intelligence differences (Deary, 2000).

References

Abelson, A. R. (1911). The measurement of mental ability of backward children. British Journal of Psychology, 4, 268–314.

Bashore, T. R., Osman, A., & Heffley, E. F. (1989). Mental slowing in elderly persons: A cognitive psychophysiological analysis. Psychology and Aging, 4, 235–244.

Blalock, H. M. (1972). Social statistics. New York: McGraw Hill.

Brody, N. (1992). Intelligence. San Diego, CA: Academic Press.

Burt, C. (1909–1910). Experimental tests of general intelligence. British Journal of Psychology, 3, 94–177.

Carroll, J. B. (1993). Human cognitive abilities. Cambridge, UK: Cambridge University Press.

Cerella, J. (1985). Information processing rates in the elderly. Psychological Bulletin, 98, 67–83.

Chorney, M. J., Chorney, K., Seese, N., Owen, M. J., Daniels, J., McGuffin, P., Thompson, L. A., Detterman, D. K., Benbow, C., Lubinski, D., Eley, T., & Plomin, R. (1998). A quantitative trait locus associated with cognitive ability in children. Psychological Science, 9, 1–8.

Cook, A. (1994). The observational foundations of physics. Cambridge, UK: Cambridge University Press.

Deary, I. J. (1994). Sensory discrimination and intelligence: Post-mortem or resurrection?American Journal of Psychology, 107, 95–115.

Deary, I. J. (1995). Auditory inspection time and intelligence: What is the direction of causation?Developmental Psychology, 31, 237–250.

Deary, I. J. (1996). Reductionism and intelligence: The case of inspection time. Journal of Biosocial Science, 28, 405–423.

Deary, I. J. (2000). Looking down on human intelligence. Oxford, England: Oxford University Press.

Deary, I. J., & Caryl, P. G. (1997). Neuroscience and human intelligence differences. Trends in Neurosciences, 20, 365–371.

Deary, I. J., McCrimmon, R. J., & Bradshaw, J. (1997). Visual information processing and intelligence. Intelligence, 24, 461–479.

Deary, I. J., & Stough, C. (1996). Intelligence and inspection time: Achievements, prospects and problems. American Psychologist, 51, 599–608.

Galton, F. (1883). Inquiries into human faculty. London: Dent.

Galton, F. (1890). Remarks on ‘Mental tests and measurements’ by J. McK. Cattell. Mind, 15, 380–381.

Gleick, J. (1987). Chaos: Making a new science. London: Macdonald.

Gustafsson, J. E. (1984). A unifying model for the structure of intellectual abilities. Intelligence, 8, 179–203.

Herrnstein, R. J., & Murray, C. (1994). The bell curve. New York: Free Press.

Jensen, A. R. (1993). Spearman's g:. Links between psychometrics and biology. In F. M. Crinella & J. Yu (Eds.), Brain mechanisms: Papers in memory of Robert Thompson (pp. 103–129). New York: Annals of the New York Academy of Sciences.

Kuhn, T. S. (1961). Measurement in modern physical science. In H.Woolf (Ed.), Quantification, a history of the meaning of measurement in the natural and social sciences (pp. 31–63). Indianapolis, IN: Bobbs-Merrill.

Lohman, D. F. (1994). Component scores as residual variation (or why the intercept correlates best). Intelligence, 19, 1–11.

Matarazzo, J. D. (1992). Psychological testing and assessment in the 21st century. American Psychologist, 47, 1007–1018.

Matthews, G., & Deary, I. J. (1998). Personality traits. Cambridge, UK: Cambridge University Press.

McKeen Cattell, J. (1890). Mental tests and measurements. Mind, 15, 373–380.

Michell, J. (1997). Quantitative science and the definition of measurement in psychology. British Journal of Psychology, 88, 355–383.

Neisser, U., Boodoo, G., Bouchard, T. J., Boykin, A. W., Brody, N., Ceci, S. J., Halpern, D. F., Loehlin, J. C., Perloff, R., Sternberg, R. J., & Urbina, S. (1996). Intelligence: knowns and unknowns. American Psychologist, 51, 77–101.

Petley, B. W. (1985). The fundamental physical constants and the frontier of measurement. Bristol, UK: Adam Hilger.

Raven, J. C. (1992). Raven's standard progressive matrices. (1992 ed.). Windsor, UK: NFER-Nelson.

Schmechel, D. E., Saunders, A. M., Strittmatter, W. J., Crain, B. J., Hulette, C. M., Joo, S. H., Pericak-Vance, M. A., Goldgaber, D., & Roses, A. D. (1993). Increased amyloid beta-peptide deposition in cerebral cortex as a consequence of apolipoprotein E genotype in late-onset Alzheimer disease. Proceedings of the National Academy of Sciences of the United States of America, 90, 9649–9653.

Seashore, C. E. (1902). A method of measuring mental work: The psycherograph. University of Iowa Studies in Psychology, 3, 1–17.

Spearman, C. (1904). “General intelligence,” objectively determined and measured. American Journal of Psychology, 15, 201–293.

Spearman, C. (1923). The nature of “intelligence” and the principles of cognition. London: MacMillan.

Spearman, C. (1927). The abilities of man. London: MacMillan.

Vickers, D., & Smith, P. L. (1986). The rationale for the inspection time index. Personality and Individual Differences, 5, 609–623.

Wechsler, D. (1981). Wechsler Adult Intelligence Scale: Revised Manual. New York: The Psychological Corporation.

Submitted: February 15, 1998 Revised: July 14, 1998 Accepted: July 18, 1998