Long experience has likewise demonstrated, fairly clearly, that the best results will be obtained in any industrial organization or educational staff by making one person chiefly responsible for the proper administration of the intellectual and educational measurements. If a personnel director is at hand who can study his tests just as scientifically as he studies his men, progress and improvement in the methods and results are inevitable.
Measurements of intelligence are by no means the only or final criteria by which the successful personnel manager wins success in his work and saves money for his employers. He makes use of every piece of information about his men that it is possible for him to pick up anywhere. The trade tests particularly offer a wide field in which measurements of intelligence may be supplemented and made more useful. Of two men who are to-day working in the same trade, receiving the same wages and making the same score on their trade tests, that one is more promising who has the higher intelligence score. On the other hand, of two equally intelligent men, as measured by the intelligence tests, that one who has attained within a given time the higher proficiency in his trade is superior.
The chief value of the group intelligence tests will probably always be in the classification of large groups of persons into smaller, well-defined groups, the members of which groups may then be studied more carefully and by more exact methods in the hands of a trained psychologist, if necessary. Until the group method of examination was developed, making it possible to test the intellectual ability of every employee without tremendous expense in time and money, it would have been most foolish to talk about maintaining a continuous inventory of the mental strength of an organization, and yet such an inventory is now possible—just as possible as the record of the condition and capacity of each machine owned by the company.
Prospective users of the Mentimeters need to bear in mind that mental powers are far less constant in their amounts than are the dimensions and measurements of a piece of steel or lumber. Even the length of a steel rail varies between winter and summer, but the variation that occurs in the strength of mental connections from day to day or from hour to hour is very much greater than the variations of the steel rail. Except by chance one would not obtain exactly the same score a second time in taking a Mentimeter test, or any other test of mental ability. Being for the most part constructed on the “increasing difficulty” plan, however, the Mentimeters will prove much less influenced by recency of drill and nearness to the lunch hour than will most other tests, especially less than those speed tests which measure how many simple tasks one can do within a given time limit. The Mentimeter ideal is to test power rather than speed.
No single set of tests should be used as final and conclusive in the public schools with regard to the kind of work which a given boy or girl should undertake. The Mentimeter tests may be used as a first “drag-net,” but those caught in this net should then be carefully studied by the most refined methods known to psychologists before being recommended for particular types of special instruction or sent to special schools. One of the most hopeful signs in the entire educational field is the number of cities that are employing psychologists to follow up the results of group examinations in the schools. Many of the state universities have established bureaus to serve the local communities[[1]] in such matters. The very finest measurements are of no avail unless something is done about the results disclosed.
[1]. There has recently been established in Teachers College, Columbia University, New York City, a Bureau of Educational Service, the Director of which would be glad to answer questions or advise with any one interested in measuring intelligence or educational results, regardless of the state or community in which one may live.
For each of the Mentimeter tests, the authors have classified the possible scores into five general groups: Superior, High Average, Average, Low Average, and Inferior. This classification is very rough and should not be wrongly interpreted. An individual who is tested with three or four or more of the Mentimeter tests should not be expected to receive the same classification in each test. In the Handwriting test, for example, a person might well be expected to make a rating of “Superior” in quality of writing while making only “Low Average” in speed of writing. The same person might well make a score on the test of Poetic Discrimination which would classify him as “Inferior.” Although there is a tendency for people who are superior in one line to have high abilities in other lines, it is only a general tendency, which will not hold good in all cases and with regard to all varieties of ability.
For the most accurate scientific work the reader will probably disregard entirely the fivefold classification of scores mentioned above. The finer distinctions made by the numerical scores will be studied, and interpretations will be made for the specific purposes of the examiner. It is probable, for example, that comparatively few children at the age of eight years would be classified as being better than “Inferior,” if these rough general classifications were to be the only record kept of performance on these tests. On the other hand, very few clerical workers of proved ability and success would make a classification as low as “Average,” except possibly in a few specialized-ability tests. The important point to be considered by the teacher of a second-grade class, or by an employer of clerical workers, or by any other person who wishes to make serious use of these tests, is the relation of the scores in the test to the relative abilities of the persons in the special group tested. The tentative classification of scores made at the end of each section of the chapter which follows this is for human beings in general and will not fit well any specialized group of persons.
In order to assist readers who have no statistical training in the evaluation for their special purposes of any particular Mentimeter test, a few pages will be devoted to an elementary statement of how to try out scientifically the relationship between a test, on the one hand, and demonstrated ability in any special line of endeavour, on the other. It may be stated here again that not all traits of mind are important in every task that must be done in life. Some positions require only a little intellectual ability while others require a great deal, and some tasks require very great development of a few traits which may be very little called for in other equally important tasks. The authors have used their best judgment as to which tests will probably select the type of persons needed in a certain type of position, but the judgments of other equally experienced men would be just as good. The final proof of reliability in a test can come only by actual trial of that test upon men of various degrees of demonstrated ability in the trade or profession concerned. What follows is a statement of how to measure this correspondence between demonstrated degree of success and score in a test, or between the scores of the same persons in two or more different tests.
No measure of relationship between success in life and success in a test can be any more accurate than the original measures of success from which the calculation is made. If the measures of success in life are unreliable, then the measure of their relationship to success in a test will be even more unreliable. The more definite and certain one can be of his measures of success, the more reliable will his measure of relationship be.