Classics in the History of Psychology

An internet resource developed by
Christopher D. Green
York University, Toronto, Ontario

(Return to index)

Psychology and Industrial Efficiency

Hugo Münsterberg (1913)


OUR plan was to illustrate the possibility of applying psychological experiments to the selection of fit applicants also in cases in which not one characteristic mental function stands out, but in which a large number of relatively independent mental activities are in play. I choose as an illustration of such cases the work of the employees at the telephone switchboard. A study of the psychological factors in this work is strongly suggested by the practical interests of the telephone companies, and may be looked on here exclusively from this point of view. The user of the telephone is little inclined to consider how many actions have to be carried out in the central office before the connection is made and finally broken again. From the moment when the speaker takes off the receiver to the cutting off of the connection, fourteen separate psychophysical processes are necessary in the typical case, and even then it is presupposed that the telephone girl understood the exchange and number correctly. It is a common experience of the companies that these [p. 98] demands cannot be satisfactorily fulfilled when a telephone girl has to handle more than 225 calls in an hour. The official statistics show that this figure is exceeded in not infrequent cases,[13] in extreme cases the number may even rise beyond 300. Moreover, in short periods of reinforced demands it may happen that for a few minutes even the rapidity of 10 calls in a minute is reached. Normally the burden is divided among the employees in such a way that about 150 calls fall to each one in an hour, and that this figure is passed considerably only in one morning and one evening hour. A skillful distribution of pauses and ample arrangements for rest, usually together with very excellent hygienic conditions, make it possible for the fit persons to be able to carry on this work without over-fatigue from 8 to 9 hours a day. On the other hand, it is only natural that such rapid and yet subtle activity under such high tension, where especially the quick localization of the correct hole is a difficult and yet indispensable part, can be carried out only by a relatively small number of human nervous systems. The inability to keep attention at such a high point for a long while, or to perform such rapid movements, or to retain the numbers correctly, does not lead to fatal accidents like those in the case of the unfit motormen, but it does lead to fatigue and finally to [p. 99] a nervous breakdown of the employees and to confusion in the service. The result is that the company is continually obliged to dismiss a considerable proportion of those who have entered the service and who have spent some months in going through the training school of the company. As one single company, the Bell Telephone Company, employs 16,000 operators, the problem is an expansive one, and it has bearing on the health of the employees as well as on the patience of the :subscribers. But above all it refers to the economic interests of the company, inasmuch as every girl who satisfies the entrance conditions of hearing and sight, of school education and general personal appearance, receives some salary throughout the months of training in the telephone school. Since during the first half-year, in which the employee still works entirely under supervision, more than a third of those who had originally entered leave, partly on account of unfitness and inability, partly on account of over-fatigue or similar reasons, the economic disadvantage to the company is evidently a very great one. The candidates are paid for months of mere training, and they themselves waste their energy and time with practice in a kind of labor which cannot be serviceable to them in any other economic activity. Under these circumstances it is not surprising [p. 100] that one city system approached me with the question whether it would not interest me from a scientific point of view to examine how far the mental fitness of the employees could be determined beforehand through experimental means.

After carefully observing the service in the central office for a while, I came to the conviction that it would not be appropriate here to reproduce the activity at the switchboard in the experiment, but that it would be more desirable to resolve that whole function into its elements and to undertake the experimental test of a whole series of elementary mental dispositions. Every one of these mental acts can then be examined according to well-known laboratory methods without giving to the experiments any direct relation to the characteristic telephone operation as such. I carried on the first series of experiments with about thirty young women who a short time before had entered into the telephone training school, where they are admitted only at the age between seventeen and twenty-three years. I examined them with reference to eight different psychophysical functions. In saying this, I abstract from all those measurements and tests which had somewhat anthropometric character, such as the measurement of the length of the fingers, the rapidity of breathing, the rapidity of pulse, [p. 101] the acuity of vision and of hearing, the distinctness of the pronunciation, and so on. A part of i the psychological tests were carried on in individual examinations, but the greater part with the whole class together.

These common tests referred to memory, attention, intelligence, exactitude, and rapidity. I may characterize the experiments in a few words. The memory examination consisted of reading to the whole class at first two numbers of 4 digits, then two of 5 digits, then two of 6 digits, and so on up to figures of 12 digits, and demanding that they be written down as soon as a signal was given. The experiments on attention, which in this case of the telephone operators seemed to me especially significant, made use of a method the principle of which has frequently been applied in the experimental psychology of individual differences j and which I adjusted to our special needs. The requirement is to cross out a particular letter in a connected text. Every one of the thirty women in the classroom received the same first page of a newspaper of that morning. I emphasize that it was a new paper, as the newness of the content was :to secure the desired distraction of the attention. As soon as the signal was given, each one of the girls had to cross out with a pencil every "a" in the text for six minutes. After a certain time, a [p. 102] bell signal was given and each then had to begin a new column. In this way we could find out, first, how many letters were correctly crossed out in those six minutes, secondly, how many letters were overlooked, and, thirdly, how the recognition and the oversight were distributed in the various parts of the text. In every one of these three directions strong individual differences were indeed noticeable. Some persons crossed out many, but also overlooked many, others overlooked hardly any of the "a's," but proceeded very slowly so that the total number of the crossed-out letters was small. Moreover, it was found that some at first do poor work, but soon reach a point at which their attention remains on a high level; others begin with a relatively high achievement, but after a short time their attention flags, and the number of crossed-out letters becomes smaller or the number of unnoticed, overlooked letters increases. Fluctuations of attention, deficiencies, and strong points can be discovered in much detail.

The third test which was tried with the whole class referred to the intelligence of the individuals. Discussion of the question how to test intelligence in general would quickly lead us into as yet unsettled controversies. It is a chapter of the psychology of tests which, especially in the service of [p. 103] pedagogy but to a certain degree also in the service of medicine, has been more carefully elaborated than any other. Often it has been contested whether we have any right to speak of one general central intelligence factor, and whether this apparently unified activity ought not to be resolved into a series of mere elementary processes. The newer pedagogical investigations, however, speak in favor of the view that besides all special processes, or rather, above all of them, an ability must be recognized which cannot be divided any further, and by which the individual adjusts his knowledge, his experiences, and his dispositions to the changing purposes of life. The grading of the pupils in a class usually expresses this differentiation of the intelligence; and while the differences of industry or of mere memory and similar secondary features may sometimes interfere, it remains after all not difficult for an observant teacher to grade the pupils of his class, whom he knows well, according to their general intelligence. The psychological experiments carried on in the schoolroom have demonstrated that this ability can be tested by the measurement of some very simple mental activities. The best method would be the one which would allow the experimenter, on the basis of a single experiment, to grade the individuals in the same order in which they [p. 104] appear in the record of the teacher. Among the various proposed schemes for this purpose the figures suggest that the most reliable one is the following method, the results of which show the highest agreement between the rank order based on the experiments and the rank order of the teachers.[14] The experiment consists in reading to the pupils a long series of pairs of words of which the two members of the pair always logically belong together. Later, one word of each pair will be read to them and they have to write down the word which belonged with it in the pair. This is not a simple experiment on memory. The tests have shown that if instead of logically connected words simply disconnected chance words are offered and reproduced, no one can keep such a long series of pairs in mind, while with the words which have related meaning, the most intelligent pupils can master the whole series. The very favorable results which this method had yielded in the classroom made me decide to try it in this case too. I chose for an experiment 24 pairs of words from the sphere of experience of the girls to be tested. Two further class experiments belonged rather to the periphery of psychology. The exactitude of space-perception was measured by demanding that each divide first the long and then the short edge of a folio sheet into two equal [p. 105] halves by a pencil mark. And finally, to measure the rapidity of movement, it was demanded that every one make with a pencil on the paper zigzag movements of a particular size during the ten seconds from one signal to another.

After these class experiments I turned to individual tests. First, every girl had to sort a pack of 48 cards into 4 piles as quickly as possible. The time was measured in fifths of a second. The following experiment which referred to the accuracy of movement impulses demanded that every one try to reach with the point of a pencil 3 different points on the table in the rhythm of metronome beats. On each of these three places a sheet of paper was fixed with a fine cross in the middle. The pencil should hit the crossing point, and the marks on the paper indicated how far the movement had fallen short of the goal. .One of these movements demanded the full extension of the arm and the other two had to be made with half-bent arm. I introduced this last test because the hitting of the right holes in the switchboard of the telephone office is of great importance. The last individual experiment was an association test. I called six words like "book," "house," "rain," and had them speak the first word which came to their minds. The time was measured in fifths of a second only, as subtler experiments, for which [p. 106] hundredths of a second would have to be considered, were not needed.

In studying the results so far as the memory experiments were concerned, we found that it would be useless to consider the figures with more than 10 digits. We took the results only of those with 8, 9, and 10 digits. There were 54 possibilities of mistakes. The smallest number of actual mistakes was 2, the largest, 29. In the experiment on attention made with the crossing-out of letters, we found that the smallest number of correctly marked letters was 107, the largest number in the six minutes, 272; the smallest number of overlooked letters was 2, the largest, 135; but this last case of abnormal carelessness stood quite isolated. On the whole, the number of overlooked letters fluctuated between 5 and 60. If both results, those of the crossed-out and those of the overlooked letters, are brought into relation, we find that the best results were a case of 236 letters marked, with only 2 overlooked, and one of 257 marked, with 4 overlooked. The very interesting details as to the various types of attention which we see in the distribution of mistakes over the six minutes were not taken into our final table. The word experiments by which we tested the intelligence showed that no one was able to reproduce more than 22 of the 24 words. The smallest number [p. 107]of words remembered was 7. The mistakes in the perception of distances fluctuated between 1 and 14 millimeters; the time for the sorting of the 48 cards, between 35 and 58 seconds; the association-time for the 6 associated words taken together was between 9 and 2l seconds. The pointing experiments could not be made use of in this first series, as it was found that quite a number of participants were unable to perform the act with the rapidity demanded.

Several ways were open to make mathematical use of these results. I preferred the simplest way. I calculated the grade of the girls for each of these achievements. The same candidate who stood in the 7th place in the memory experiment was in the 15th place with reference to the number of letters marked, in the 3rd place with reference to the letters overlooked, in the 2lst place with reference to the number of word pairs which she had grasped, in the 11th place with reference to the exactitude of space-perception, in the 16th place with reference to the association-time, and in the 6th place with reference to the time of sorting, As soon as we had all these independent grades, we calculated the average and in this way ultimately gained a common order of grading. It is evident that this kind of calculation contains accidental factors, especially as a consequence of the fact that we [p. 108] give equal value to every one of these results. It might be better, for instance, to attribute a higher value to the attention experiment or to the intelligence experiment. This could be done by multiplying the results of some of these grades by 2 or by 3, which would bring the high or low grade of a girl for a particular function to stronger influence in the final result. But in this first trial I contented myself with the simplest uniform scheme in order to exclude all arbitrariness, and therefore considered the mere average of all the grades as the expression of the experimental result.

With this average rank list, we compared the practical results of the telephone company after three months had passed. These three months had been sufficient to secure at least a certain discrimination between the best, the average, and the unfit. The result of this comparison was on the whole satisfactory. First, the skeptical telephone company had mixed with the class a number of women who had been in the service for a long while and had even been selected as teachers in the telephone school. I did not know, in figuring out the results, which of the participants in the experiments these particularly gifted outsiders were. If the psychological experiments had brought the result that these individuals who stood so high [p. 109] in the estimation of the telephone company ranked low in the laboratory experiment, it would have reflected strongly on the reliability of the laboratory method. The results showed, on the contrary, that these women who had proved most able in practical service stood at the top of our list. Correspondingly, those who stood the lowest in our psychological rank list had in the mean time been found unfit in practical service and had either left the company of their own accord or else had been eliminated. The agreement, to be sure, was not a perfect one. One of the list of women stood rather low in the psychological list, while the office reported that so far she had done fair work in the service, and two others to whom the psychological laboratory gave a good testimonial were considered by the telephone office as only fair.

But it is evident that certain disagreements would have occurred even with a more ideal method, as on the one side no final achievement in practical service can be given after only three months, and because on the other side a large number of secondary factors may enter which entirely overshadow the mere question of psychophysical fitness. Poor health, for instance, may hinder even the most fit individual from doing satisfactory work, and extreme industry and energetic [p. 110] will may for a while lead even the unfit to fair achievement, which, to be sure, is likely to be coupled with a dangerous exhaustion. The slight disagreements between the psychological results and the practical valuation, therefore, do not in the least speak against the significance of such a method. On the other hand, I emphasize that this first series meant only the beginning of the investigation, and it can hardly be expected that at such a first approach the best and most suitable methods would at once be hit upon. A continuation of the work will surely lead to much better combinations of test experiments and to better adjusted schemes. But it would be most desirable that such studies be undertaken at various places according to various schemes in order to come nearer to the solution of a problem which is economically important to the whole public and to many thousands of employees. As soon as methods are really perfected it would seem not at all impossible that by a short experiment of a few minutes thousands of applicants might be saved long months of study and training which are completely wasted. For us here the detailed analysis of this particular case did not mean a suggestion to use to-day in the telephone offices of the country the special scheme which we applied, but it stood only as a clear, [p. 111] simple illustration of a method by which not the specific work itself is tested, but by which the industrial work of the individual is resolved into a long series of parallel functions each one of which is tested independently. The experimental aid which the laboratory has to supply in such cases is not a newly invented device, such as we needed in the case of the motormen, but simply the methods well known as so-called mental tests.

The experiments with such tests by which single mental functions are measured approximately in short quick examinations, has been much discussed in psychological circles. For a long while the thorough scholars remained very reluctant to accept such an apparently superficial scheme, when these tests were proposed especially for the pedagogical interests of the schoolroom. It was a time in which the scientific efforts were completely devoted to the general problems of the human mind and in which individual differences were very little considered. Moreover, the questions of applied psychology still seemed so far distant that the true scholar instinctively took his standards from the methods of purely theoretical research. Seen from such a point of view, it could not be denied that the tests were not sufficient to give us a complete scientific analysis of [p. 112] the personality in its subtler structure. The theorists knew too well that if the reactions, or associations, or memories, or tendencies of attention, or emotions of a subject were measured really with that scientific thoroughness which is the ideal of research, long months of experiments would be needed, and little could be hoped for from tests to be performed in half an hour. But this somewhat haughty reserve which was quite justified twenty years ago has become obsolete and would be meaningless to-day. On the one side the methods themselves have been multiplied; for each mental act like memory, attention, and so on, dozens of well-studied tests are at our disposal, which are adjusted to the finest ramifications of the functions.[15] On the other side the interest in individual differences and in applied psychology has steadily grown, and through it an understanding for the real meaning of the tests has been gained. Their value, indeed, lies exclusively in their relation to the practical problems. Where theoretical questions are to be answered and scientific studies concerning the laws and variations of the mind are to be undertaken, the long series of laboratory experiments carried on with patience and devotion are indispensable and can never be replaced by the short-cut methods of the tests. But where practical tasks of [p. 113] pedagogy or jurisprudence or medicine, or especially of commerce and industry, are before us, the method of tests ought to be sovereign. It can be adapted to the special situations and can succeed perfectly, if the task is to discover the outlines of the mental individuality for particular practical work.

The only real difficulty of the method lies in the ease with which it can be used. A device which presupposes complicated instruments deters the layman and will be used only by those who are well trained. Moreover, the amateur would not think of constructing and adapting such apparatus himself. But when nothing is necessary but to use words or numbers or syllables or pictures, or, as in those experiments which we just described, newspapers and so on, any one feels justified in applying the scheme or in replacing it by a new apparently better one according to his caprice. The manifoldness of the proposed tests for special functions is therefore enormous to-day. What is needed now is surely much more that order be brought into this chaos of propositions, and that definite norms and standards be secured for certain chief examinations, than that the number of variations simply be increased.

The chief danger, moreover, lies in the fact that those who are not accustomed to psychological [p. 114] laboratory research are easily misled. They fancy that such an experiment can be carried out in a mere mechanical way without careful study of all the conditions and accompanying circumstances. Thereby a certain crudeness of procedure may enter which is not at all suggested by the test method itself. The psychological layman too seldom recognizes how many other psychical functions may play a rôle in the result of the experiment beside the one which is interesting him at that moment. The well-schooled laboratory worker almost automatically gives consideration to all such secondary circumstances. While his experiments may refer to the process of memory, he will yet at the same time carefully consider the particular situation as to the emotional setting of the subject, as to his attention, as to his preceding experience, as to his intelligence, as to his physiological condition, and many other factors which may have indirect influence even on the simplest memory test. Hence the real performance of the experiments ought to be undertaken only by those who are thoroughly familiar and well trained in psychological research. And they alone, moreover, can decide what particular form such an experiment ought to take in a given practical situation. It must be left to them, for instance, to judge in which cases the mental function [p. 115] of economic importance ought to be tested after being resolved into its components and in which it ought to be examined in its characteristic unity.


[13] Investigation of Telephone Companies: Bureau of Labor. (Washington, Government Printing Office, 1910.)

[14] Ries: Beiträge zur Methodik der Intelligenzprüfung. (Zeitschrift für Psychologie, 1910, vol. 56.)

[15] For a survey of a large number of such tests and bibliography, compare: G. M. Whipple: Manual of Mental and Physical Tests. (Baltimore, 1911.)