Image Map

Intelligence: General Intelligence – Part II

Hopefully, having read Part I of this series, the reader has come to regard intelligence as a sort of ill-defined bulge in the pants pocket of psychology – an entity that surely exists, but whose exact nature is questionable. The last installment left off examining the WAIS-III, a widely used IQ test that employs a multitude of disparate subtests to determine intelligence level.

At first glance, the fact that the WAIS-III really measures anything other than how skillfully one can perform on its specific subtests might seem amazing, yet Part I showed that scores on IQ tests correlate with grades in school, level of achievement, and a number of other real-world proxies of intelligence. Another equally amazing occurrence is that a task such as mental arithmetic has a moderate to strong correlation with a superficially different task such as defining words. With those correlations in mind, let us begin to explore the concept of g, the general factor.

General Intelligence

A common stereotype is a “math guy” whose idea of fine literature is the manual for his TI-89, and the “word guy” who might beat you at Scrabble with QOPHS on a triple word score, but couldn’t multiply if he were a rabbit. Without considering data or looking at the problem from a scientific standpoint, the majority of people would expect various mental tests to verify this belief. Unfortunately, the majority of people would be wrong.

A surprising discovery that has been called “one of the most remarkable findings in all of psychology” (Jensen, 1981) is that all mental tests are positively correlated (Jensen, 1998; Mackintosh, 1998). In other words, someone who tends to do well on one type of mental test has a tendency to do well on any other type of mental test. The implications of this so-called positive manifold are huge. To explore and explain these correlations, Charles Spearman created a method called factor analysis.

Specifics aside, factor analysis is a tool for making sense of a complex set of correlations. For example, in untrained individuals it seems logical that the circumference of various body parts would be positively correlated; the slim tend to be small all over, while the big-boned tend to be large everywhere. Taking the data and applying a factor analysis would reveal a single factor that helps explain the correlations. In this case, we could hypothesize that this single variable is weight. That’s not to say that weight is the only factor involved, it just means that weight is the variable that can explain most of the variance in the data.

It is not always the case that a single factor emerges. Data on football ability would probably reveal a number of different factors, perhaps weight, speed, and reaction time, none of which would be strongly correlated with the others. However, when proper factor analysis is applied to the positive manifold of mental tests, a single factor, referred to as g, emerges. The existence of g means there is some variable that is common to all mental tests – an omnipresent influence an anything mental.

Are g and Intelligence the Same?

In the literature, g is sometimes called “general intelligence,” implying that a definition is finally at hand. It is still dangerous to use the term intelligence, regardless its new modifier, because of the pervasive connotations that come along for the ride. The convention for the rest of the article will be that general intelligence just means g, and g is just the general factor that emerges from a factor analysis of the correlation of mental tests.

There are many theories about g’s exact nature, some claiming its synaptic plasticity (the best candidate in my humble opinion), some maintaining it is an aggregate of multiple simple neural functions, others espousing the view that it is nothing but a mathematical construct that lends itself to erroneous reification. No one knows for sure. What is known, is that g affects just about every aspect of our thinking. If scientists are able to find one biological function that causes g we can be sure it’ll be a rather low-level function of the brain.

A Few Technical Notes and Definitions

Presenting g in a one-page summary doesn’t do it justice at all. G is a powerful, controversial phenomenon that is apt to be misinterpreted. Spearman’s original idea about general intelligence was a two-factor theory. He assumed every mental test measured a specific ability, random error, and g. This concept was shown to be a little off, but it will serve us well for an introductory article. The two-factor theory implies that the less specific abilities a test requires, the more g it will require. The amount of g required for a test is referred to as its g-loading.

The term “mental test” encompasses any test that doesn’t rely heavily on any specific motor or sensory pathway. IQ tests are mental tests, and it’s been established that g causes most of the variance in the scores of IQ tests. That means an IQ test is really a test of g, so studies that show a correlation between IQ and behavior also show a correlation between behavior and g.

Implications of a General Factor

Street Smarts vs. Book Smarts

Often a distinction is made between so-called book-smarts and street-smarts – the dichotomy of intellect separating those who excel scholastically but lack the common sense to spit with the wind, and those who do poor in a structured learning environment but can figure out the easiest way to avoid doing a TPS report every week. It’s Bill Gates versus MacGyver; Bobby Fischer versus Indiana Jones.

How does this bifurcation hold up in light of g? Basically it doesn’t, at least not as a theory of intelligence. If we think of intelligence as a mental ability, and we know all mental abilities are correlated, then it doesn’t make sense that there are two kinds of “smarts.” The theory can be somewhat rectified, however, if it’s realized that the group of attributes that might make someone street-smart or book-smart aren’t all part of intelligence per se. To do well in school certainly requires intelligence, but other things like motivation and punctuality are also necessary. Likewise, salesmanship requires the ability to conceive successful selling strategies (a mental ability), but it also helps to be attractive and have an agreeable personality.

Even with this fix, the theory is misleading because it downplays the importance of g in both types of “smarts.” Since the book-smart student and the street-smart salesperson both rely heavily on some sort of mental abilities, a large part of their success is dependant on g. This suggests that both “smarts” are related. In other words, someone who is street-smart will tend to be book-smart and vice a versa. If that’s the case why have a distinction in the first place?

The Real World: Your Brain

A distinction is needed, some might say, because the real world doesn’t care about intelligence our how intelligence interacts with other factors like personality, motivation, and drive. Some might go as far as to toss intelligence and IQ tests in the garbage can of inapplicable uselessness next to Afghan finger ink. That would be going too far.

In the world of psychometrics, theory is the alcohol-clouded, late-night mate selection, and real world application is the ray of first light that reveals if you made Darwin proud, or if you just pissed in the gene pool. Despite the fact that general intelligence largely leaves out important factors like motivation and personality, it still is a great predictor of real world success.

IQ has been shown in numerous studies to correlate very well with academic achievement (Jenson, 1980; Jenson, 1991; Mackintosh, 1998; Matarazzo, 1972; Snow and Yalow, 1982). This is probably because IQ has also been shown to correlate highly with learning rate (Gettinger, 1984; Christal, 1991) and both oral and reading comprehension (Sticht et al., 1981). Similarly, IQ shows a high correlation with job performance (Walters et al., 1993).

A very interesting study was done on the predictive validity of g with the Armed Services Vocational Aptitude Battery (ASVAB). The ASVAB contains numerous subtests whose g-loadings range from very small to high. When the low g-­loaded subtests were reviewed independently, they had an extremely poor ability to predict job performance. On the other hand, looking at only the subtests with a high g-loading yielded a much higher predictive validity. This indicates that just about all the predictive validity of the ASVAB is due to its ability to measure g (Ree et al., 1994).

IQ score has also been shown to be negatively correlated with crimes, especially those that show violence and impulsiveness (Herrstein & Murray, 1994; Gordon, 1987). This correlation even exists in siblings who have been raised in the same household, which suggests IQ is more predictive of illegal behavior than upbringing. There are a number of theories about the correlation between delinquency and IQ, but there is no real consensus. One theory is that, generally, criminals don’t have the ability to delay gratification or the foresight to realize the consequences of their action. Another hypothesis is that those people with a low IQ have less opportunity in general and turn to crime as an alternate route of gratification (Jensen, 1998).

Even something as commonplace and seemingly unrelated to intelligence such as motor vehicle accidents have been show to be predicted by IQ. A 1990 study by Brian O’Toole showed that the death rate resulting from motor vehicle accidents is nearly three times as high in those with low IQs (80-85) than with high IQs (above 115).

The correlations and predictions go on and on, quickly getting too numerous to list here. The most essential fact to distill from the ocean of studies is that the g factor permeates countless aspects of society and life. Clearly it is powerful and impressive in its predictive power.

Questions and Challenges

IQ and success

Some people have trouble reconciling the unsuccessful intellectual when presented with the ubiquitous nature of g. Joey Adams once said, “A genius is one who can do anything except make a living,” and in some cases that seems all too true. There is a minority of people with an extremely high level of g who are poor, out of work, or who fail out of school. Rather than discrediting g, this is more often a case of misunderstood personal resource allocation.

Society in general roughly equates wealth with success, but someone with a high level of g is more likely to have the reasoning ability to determine for himself what he considers success, and then pursue that avenue. In my own experience with people who are extremely intelligent, I’ve found a large number tend to value free time as much as income, and ability to think freely as much as prestige. That line of reasoning can be used as a cop out for almost anyone, which doesn’t make it any less valid, but does cause some eyes to roll when it’s mentioned.

I know of no empirical data to confirm this, but my own experience and personal correspondences lead me to believe that an IQ of about 130 is most conducive to the accepted version of success. At that level, two standard deviations above the mean, one has abilities that far outstrip the average population, but is more likely to be a somewhat conventional thinker.

IQ and Genius

When the subject is IQ, the topic of genius is quick to come up. While there are few, if any, people who would argue that a genius is just someone with a high IQ, there are countless individuals ready to knock down that straw man. The widely held belief is that achieving genius or eminence in a field is a function of a high level of general ability, personality factors like obsession and motivation, and other external factors such as luck. Psychologist Howard Gardner has estimated that an IQ of at least 120 is required to become a true “genius” (Jensen, 1998). Others, myself included, believe that number is a bit on the low side.

IQ: The second most exaggerated number online

If this article has piqued your interest in IQ and you decide to try your hand at an online test, let me suggest taking about 2 grams of r-ala to prevent hyperglycemia from the sugar coating. Well, that’s not fair for all online IQ tests, just the vast majority.

To put some actual IQ scores in perspective, look at this table from David Wechsler’s 1944 book The Measurement of Adult Intelligence:

Classification I.Q. Limits Percent Included
Defective 65 and below 2.2
Borderline 66-79 6.7
Dull Normal 80-90 16.1
Average 91-110 50
Bright Normal 111-119 16.1
Superior 120-127 6.7
Very Superior 128 and over 2.2

Notice that only 2.2 percent of the population has an IQ of 128 or over. Wechsler doesn’t even go into IQs like 140 or 160 that are so routinely claimed online. Those scores are too rare to be of clinical interest to most psychologists. Think of that next time someone claims a stratospheric IQ but can’t figure out how to take Caps Lock off.

If “know thyself” is a Maxim you subscribe to, an IQ test should be included right next to the centerfold. No other psychological test has such widespread predictive validity. Of course, IQ tests aren’t perfect. An important thing to remember is that an IQ score should be looked at as a guideline rather than a definitive limitation. A vast array of human faculties is influenced by general intelligence, but that is not their only influence. A motivated individual can accomplish a great many things more than expected from general intelligence alone. In addition, it is human nature that most people automatically are drawn towards goals that they can accomplish. You don’t see too many mentally impaired individuals who like to spend their days lounging around the house proving theorems.

IQ testing itself is still in its infancy. I suspect the next 50 years will bring a revolutionary new era of accurate ability measurement and possibly enhancement. In the end, IQ tests and general intelligence should be seen as quantifiable bits of the fabric that make us living, thinking, conscious entities.

Works Cited

Christal, R.E. (1991). Comparative validities of ASVAB and LAMP tests for logic gates learning. Brook, AFB, TX: Manpower and Personnel Division, Air Force Human Resources Laboratory.

Eysench, H. J. (1995). Genius: the natural history of creativity. Cambridge University Press, Cambridge.

Gettinger, M. (1984). Individual differences in time needed for learning: A review of literature. Education Psychologist, 19, 15-29.

Gordon, R.A., (1987) SES versus IQ in the race-IQ-delinquency model. International Journal of Sociology and Social Policy, 7, 203-320.

Herrstein, R. J., & Murray, C. (1994) The bell curve: Intelligence and class structure in American life. New York: Free Press.

Jensen, A.R. (1981). Straight talk about mental tests. Methuen, London.

Jensen, A.R. (1998). The g Factor: The Science of Mental Ability. Westport: Praeger Publishers.

Jensen, A.R. (1980). Bias in mental testing. New York: Free Press.

Jensen, A.R. (1991). Spearman’s g and the problem of educational equality. Oxford Review of Education, 17, 169-187.

Mackintoch, N.J. (1998). IQ and Human Intelligence. New York: Oxford University Press.

Matarazzo, J. D. (1972). Wechsler’s measurement and appraisal of adult intelligence (5th ed.). Baltimore: Williams & Wilkins.

Ree, M. J., Earles, J. A. & Teachout, M.S. (1994). Predicting job performance: Not much more than g. Journal of Applied Psychology, 79, 518-524.

Snow, R. E. & Yalow, E. (1982). Education and intelligence. In R.J. Sternberg, Handbook of human intelligence (pp.493 – 585). Cambridge: Cambridge University Press.

Sternberg, R.J., Forsythe, G. B., Hedlund, J., Horvath, J. A., Wagner, R. K., Williams, W. M., Snook, S. A., Grigorenko, E. L. (2000). Practical intelligence in everyday life. New York: Cambridge University Press.

Sticht, T. G., Hooke, L.R. & Carylor, J.S. (1981) Literacy, oracy, and vocational aptitude as predictors of attrition and promotion in the Armed Services. Alexandria, VA: Human Resources Research Organization.

Walters, L.C., Miller, M.R. & Ree, M.J. (1993) Structured interviews for pilot selection: No incremental validity. International Journal of Aviation Psychology, 31, 25-38.

×