IV. MECHANICAL ABILITY
In 1915 Stenquist, Thorndike, and Trabue, working with dependent children in a county of New York State, used tests of various mental functions, including a test of ability to put simple mechanisms together. These correlations showed that whereas tests of ability to handle language and tests of general intelligence (Binet-Simon) gave positive coefficients as high as .90, the test of mechanical ability yielded a coefficient much lower, when correlated with these. They therefore suggested that mechanical ingenuity might be a relatively specialized form of capacity, not reliably predictable from knowledge of general intelligence.
Subsequently, one of these investigators, Stenquist, made extended tests, and standardized a measuring scale to gauge mechanical ability. Measuring individuals for general intelligence and for mechanical ability, a positive coefficient of correlation amounting to about .40 is ordinarily obtained. This relationship is obviously not close. Ability to put mechanisms together is not reliably predictable from status in general intelligence. The chances are, however, that a pupil who is superior in general intelligence will score higher in mechanical ability, than a generally stupid pupil will score. There is no negative or compensatory relation between the two functions, as is sometimes assumed.
Wider studies, including tests of learning mechanical processes, will give further light upon the extent to which ability to deal with concrete mechanisms coheres with general intelligence, and to what extent comprehension of mechanical principles is so correlated. It may be that the correlation between performances in Stenquist’s tests and in tests like those used to measure general intelligence is reduced through factors like selective attention operating over a period of years. It may be that the relatively unintelligent become relatively more proficient in concrete acts, like assembling a bicycle bell or putting a lock together, because they have not the degree of intelligence that would enable them to prefer reading as an activity. Thus when 40 fifteen-year-old boys, 20 of whom have IQ’s (Stanford-Binet) from 150 to 170, and 20 of whom have IQ’s from 90 to 100, are faced with a series of tasks similar to those mentioned above, those of lower IQ might conceivably produce a record equal to or surpassing that of the first group, because their ability had enabled them to practice only tasks at a comparatively low level of general capacity. With an equal amount of attention to these matters, not previously of much interest to them, the boys of 150 to 170 IQ might surpass their competitors greatly. In a test of cake-baking, a hundred housewives, selected at random on a given date, will surpass the hundred most eminent men of science; but not after both groups have attended to the matter for an equal length of time.
The tests of mechanical ability do not as yet eliminate the influence of mechanical interest upon the outcome of the test. Extremely high intelligence may well be relatively little interested in concrete materials and processes, preferring to manipulate ideas. Thus on a given date lower intelligence, long acting on that level, may surpass. Yet the higher IQ may really be capable under incentive, of surpassing in work with things as well as in work with ideas. Tests of learning mechanical processes would, therefore, be a most valuable supplement to what has already been done in this field.
Great inventors of mechanical devices are probably, as a group, very far above the average in general intelligence. This statement cannot be made with positive certainty, as the general intelligence of a large number of inventors has never been measured. It rests only on deduction from the fact that invention evidently calls for a high degree of selective thinking, and of interest in problem situations. Even “invention by accident” which may occasionally occur, calls for a high degree of ability to “notice” a new element in the familiar situation, in relation to other elements.