Mental Measurements

In psychiatry, as in other sciences, precise measurement and objective statement present great advantages. It is desirable to express whenever possible in quantitative terms the conduct and mental status of the patient. In this way errors of personal interpretation may be avoided and reliable comparisons made of conditions, individuals, and recorded observations of clinicians.

Thus, it is quite as possible to measure memory as it is to determine the pulse. To describe the former as "rather poor" is as inexcusable as to report the latter as "somewhat slow." Even such complex symptoms as incoherence, dis-tractibility, retardation, dilapidation of school knowledge, lend themselves, with certain limitations, to measurement.

In the endeavor to express in precise language the deviations of conduct, capacity and experience, the psychiatrist finds frequent need for the employment of technique and materials elaborated by the psychologist.

Quantitative methods involve a considerable amount of time and a degree of professional skill which cannot be hastily acquired. In incompetent hands their results may be given a significance never vouched for by their elaborators, or they may fail to reveal the significant data potential in them. In slovenly and careless hands they may yield a false impression of accurate report. In the hands of the unprepared the results of their use may often be set forth as conclusive without due regard to other significant factors. For such reasons it is desirable in practice for the expert psychologist to be consulted in his own field, just as are the chemist, toxicologist, and X-ray specialist.

Psychological measurement proceeds by providing uniform experimental situations or stimuli, establishing by preliminary research the normal or standard responses to these stimuli, and so scoring the subject's reactions that they may be graded in terms of achievement or of value.

Normal Curves Of Distribution

Measurements of mental traits have shown that individuals are distributed, with respect to them, according to the familiar curve of the probability integral. Human beings do not fall into sharply separated types or species, such as the slow and the fast, the elated and the depressed, the normal and the abnormal. Instead, in any mental trait that can be measured, the human family would be found to constitute but a single species, to fall within the limits of a normal curve of distribution. Such a curve of frequency means that all degrees of a given trait will be found to occur. Certain degrees of it, the median, modal or average degree, occur most frequently. Those individuals possessing this median degree of the trait, or deviating from it only by a stated amount, will constitute the typical. As one goes above or below this degree the individuals become gradually fewer and fewer.

In Fig. 1 are shown typical curves of normal distribution. Points on the base line or abscissa indicate in progressive order amount or degree of a given trait. Points on the ordinates indicate frequency of occurrence. OF on both curves represents the median or average degree of the trait as well as the fact of its most frequent occurrence. The lines ab and cd are equally distant from OY, and the area abYcd is one-half of the total area under the curve. The measure aO or Od represents average deviation and is technically known as Probable Error.

For example, the average educated adult performance in the well-known Tapping test is 376 reciprocal innervations in one minute. Half of the individuals of this class would fall within 37 taps above or below this average, i.e., between 339 and 413 taps. The number 37 is, then, the Probable Error of the distribution. Unusual or atypical performance would fall on the base line at a distance from the median greater than that represented by the Probable Error; the more unusual or atypical it is, the farther from the median it will fall.

The Practice Of Psychiatry Applications Of Psychol 1The Practice Of Psychiatry Applications Of Psychol 2

Fig. 1.

In a normal surface of frequency a distance of one Probable Error to one or the other side of the median includes 25% of the total number of cases. A distance of two Probable Errors includes an additional 17%; three, an additional 6%, and four, the remaining 2%, approximately, in each case.

Turning again to the Tapping test as an example, an individual who should be able to tap 450 times in a minute exceeds the average by 74 taps, which is double the Probable Error. He is thus two Probable Errors removed from the average in the direction of superiority. He would be excelled by only 8% of the group and would himself be number nine from the top if one hundred representative individuals were arranged in an order of tapping ability. It has thus been possible not only to measure his actual tapping rate, but also to show what degree of deviation from the average the record indicates and the frequency of this capacity.