The Organization Man Page 23

  13) Prostitution should be state supervised.

  14) Modern art should not be allowed in churches.

  15) It is worse for a woman to have extramarital relations than for a man.

  16) Foreigners are dirtier than Americans.

  17) “The Star-Spangled Banner” is difficult to sing properly.

  WORD ASSOCIATION QUESTIONS

  Underline the word you think goes best with the word in capitals:

  18) UMBRELLA (rain, prepared, cumbersome, appeasement)

  19) RED (hot, color, stain, blood)

  20) GRASS (green, mow, lawn, court)

  How Good an Organization Man Are You?

  21) NIGHT (dark, sleep, moon, morbid)

  22) NAKED (nude, body, art, evil)

  23) AUTUMN (fall, leaves, season, sad)

  HYPOTHETICAL SITUATIONS—JUDGMENT TYPE

  24) What would you do if you saw a woman holding a baby at the window of a burning house:

  a) Call the fire department

  b) Rush into the house

  c) Fetch a ladder

  d) Try and catch the baby

  25) Which do you think is the best answer for the executive to make in the following situation:

  Worker: “Why did Jones get the promotion and I didn’t?” Executive:

  a) “You deserved it but Jones has seniority.”

  b) “You’ve got to work harder.”

  c) “Jones’s uncle owns the plant.”

  d) “Let’s figure out how you can improve.”

  OPINION QUESTIONS: POLICY TYPE

  26) A worker’s home life is not the concern of the company.

  Agree........ Disagree........

  27) Good supervisors are born, not made.

  Agree........ Disagree........

  28) It should be company policy to encourage off-hours participation by employees in company-sponsored social gatherings, clubs, and teams.

  Agree........ Disagree........

  OPINION QUESTIONS: VALUE TYPE

  29) When you look at a great skyscraper, do you think of:

  a) our tremendous industrial growth

  b) the simplicity and beauty of the structural design

  30) Who helped mankind most?

  a) Shakespeare

  b) Sir James Newton

  CHAPTER 15 The Tests of Conformity

  If personality tests are the voice of the organization it is not that testers mean them to be. A few are lackeys, but the great majority cherish the professional’s neutrality; they try hard to be objective, and “Value judgments” they shun. And this is the trouble. Not in failing to make the tests scientific enough is the error; it is, rather, in the central idea that the tests can be scientific.

  They cannot be, and I am going into quite some detail in this chapter to document the assertion. I do so because the tests are the best illustration of the underlying fallacies of scientism—and the underlying bias. As in all applications of scientism, it is society’s values that are enshrined. The tests, essentially, are loyalty tests, or rather, tests of potential loyalty. Neither in the questions nor in the evaluation of them are the tests neutral; they are loaded with values, organization values, and the result is a set of yardsticks that reward the conformist, the pedestrian, the unimaginative—at the expense of the exceptional individual without whom no society, organization or otherwise, can flourish.

  What I am examining is not the use of tests as guides in clinical work with disturbed people, or their use in counseling when the individual himself seeks the counseling. Neither is it the problems of professional conduct raised by the work of some practitioners in the field, interesting as this bypath is. What I am addressing myself to is the standard use of the tests by organizations as a gauge of the “normal” individual, and the major assumptions upon which the whole mathematical edifice rests.

  The first assumption is the idea that if we can measure minor variations in most human beings, we can to a large degree predict what they will do in the future. If the principle of this were true, tests of aptitude—such as vocabulary, or finger dexterity tests—should have long since proved the point. Unlike personality tests, they measure what is relatively measurable, and they have been used so long and on so many people that a vast amount of before-and-after documentation is available.

  Aptitude tests have proved useful in distinguishing capabilities. Once past establishing whether a person probably can or cannot meet minimum requirements, however, their predictive success is not impressive. The Army Air Force mass testing experience furnishes perhaps the largest body of evidence available. During the war the Air Force tested hundreds of thousands of men with the standardized battery of tests known as the “Stanines.” Now, by going back and comparing the initial prediction with how the men actually performed, we gain a rather clear idea of their usefulness. Here are the conclusions psychiatrist Lawrence Kubie has drawn from the comparison:

  The tests of aptitudes were remarkably accurate as far as they went; they selected accurately a small group at one extreme, most of whom would succeed in training, and another small group at the opposite pole, most of whom would fail. (There were exceptions to the results even at both extremes.)

  As was to be expected, however, the vast majority of the men tested fell into the central zone of the normal curve of distribution, while only a relatively small percentage of the tested population was placed at the two extremes. With rare exceptions, the individuals who fell into the extremes knew their own aptitudes and ineptitudes before going through any tests. From their experiences at play, in sport, in school, and on various jobs, they knew already that they were specially adept or specially maladroit with respect to certain types of activity. Indeed the representatives of the two extreme ends of the scale were usually able to describe their strong and weak points almost as precisely as these could be measured.

  The next important lesson of the entire experiment with the “Stanines” was that for the majority, who fall in the great middle zone of the normal curve of distribution, their minor variations in aptitudes do not determine either success or failure, happiness or unhappiness in a career. For most of us (that is, for the Average Man), a subtle balance of conscious and unconscious forces determines how effectively we use our native aptitudes, whether intellectual, emotional, sensory, neuromuscular, or any combination of these aptitudes. For most of us it is not the minor quantitative differences in the machine itself, but the influence of these conscious and unconscious emotional forces on our use of the human machine which determines our effectiveness. For me this was the ultimate lesson from the experience of the Air Force with the “Stanines.”

  The personality tester can reply that this merely proves his point —i.e., if a combination of aptitude tests cannot predict the future because of subtle emotional forces, why, then, all we have to do is add some tests that will measure these forces. Once again, the fatal step from science to scientism. In aptitude testing the responses are of a character that can be rated objectively—such as the correctness of the answer to 2 plus 2 or the number of triangles in a bisected rectangle. The conclusions drawn from these aptitude and intelligence scores are, furthermore, limited to the relatively modest prediction of a man’s capability of doing the same sort of thing he is asked to do on the tests. If the tests indicate that a man has only 5,000 words in his vocabulary, it is a reasonable assumption that he won’t do particularly well in a job requiring 50,000 words. If he is all thumbs when he puts wiggly blocks together, he won’t be very good at a job requiring enough manual dexterity to put things like wiggly blocks together.

  To jump from aptitude testing to personality testing, however, is to jump from the measurable to the immeasurable. What the personality testers are trying to do is to convert abstract traits into a concrete measure that can be placed on a linear scale, and it is on the assumption that this is a correct application of the scientific method that all else follows. But merely defining a trait is immensely difficult, let alone deter
mining whether it can be measured as the opposite of another. Is “emotionalism,” for example, the precise statistical opposite of “steadiness”? People are daily being fitted onto linear scales for such qualities, and if their dimensions don’t fit they are punished, like those on Procrustes’ bed, for their deviance.

  But what is “personality”? The surface facets of a man—the way he smiles, the way he talks? Obviously not, the psychologists admit. We must go much deeper. But how much deeper? Few testers would dream of claiming that one can isolate personality from the whole man, yet logic tells us that to be able statistically to predict behavior we would have to do just this. The mathematics is impeccable—and thus entrapping. Because “percentiles” and “coefficients” and “standard deviations” are of themselves neutral, the sheer methodology of using them can convince people that they are translating uncertainty into certainty, the subjective into the objective, and eliminating utterly the bugbear of value judgments. But the mathematics does not eliminate values; it only obscures them.

  Let’s take the interpretation of test scores. Testers argue that the human element has largely been eliminated from this process; except with “projective” tests the scoring is standardized—if you choose answer (d.) you get so many points, and what the tester may think has nothing to do with it. But you don’t take just one test; you usually take several different tests, and the crux of the process comes when the tester attempts to build a composite picture out of all these different sub-scores. The more different scores there are to put together, the more a job of interpretation, not less, the tester has.

  With even the most disciplined mind, it is impossible to erase the influence of one’s own environment and outlook. Testers likewise, and when any have some stray neuroses themselves, their interpretations can be downright dangerous. Several years ago, a midwest executive sent a job applicant he had investigated and thought very well of to a psychological consultant for testing. The report that came back was surprisingly ominous: according to the analyst, the man was abnormally anti-authority and would have “insufficient feelings of loyalty to the organization.” The executive hired the man anyway—he didn’t care whether the man loved the company or not; he just wanted a job done. The man worked out fine. A year later, when a similarly ominous report came in on a similarly able man, the executive became curious. He decided to go over and chat with the analyst. “The poor guy was pathetically jealous,” the executive recalls. “He was eating his heart out because men his own age that I was sending over had gone way past him. I asked him about the first man he had warned me against. Obviously unstable, he told me; the man had two kids, yet he had bought a convertible. Also, he was building an ‘ultra-modern’ house.”

  In the case of projective tests, interpretation is even more critical. Originally, they were meant to be used only as part of an exhaustive clinical diagnosis, and the few men who are expert in them have warned against their application to selection. As they point out, the tests can sometimes be more a projection of the man who is doing the evaluation than of the man who is being evaluated.

  What do symbols symbolize? David Riesman tells of a Thematic Apperception Test taken by a graduate student in history. In these tests you are shown a picture—of a man going out a doorway, for example—and asked to tell a story about it. The history student, not too surprisingly, told a story about a famous historical figure who had had a difficult choice to make. Ah-ha! said the man who interpreted the tests, maladjustment. The student had talked about people who were dead. This was the first thought that a historical figure had called up in the mind of the tester.

  No matter what tests are used, the interview experience itself is surcharged with values. The dehumanized literature of the field gives little hint of the highly personal overtones that mark the meeting of interviewer and interviewee. Consider the situation in which a man who has reached middle age sees himself weighed in the balance by a man whom he has never seen before. Even were both impossibly “normal” the relationship would be difficult, for all the amenities of civilized discourse cannot suppress in either the consciousness that there is a conflict of interests between the two. There is no interviewee who does not have something within himself that he fears to reveal, and there are few interviewers who do not wish to find it. For professional reasons alone the interviewer wishes to probe for the bared nerve.

  And sometimes for personal reasons too. I remember vividly a conversation a colleague and I had with a well-known consultant. Quite voluntarily, he bared his own nerve. In the course of explaining his technique of interviewing he referred to the OSS testing program in World War II and how candidates were submitted to a grueling series of experiences to test their reaction under shock. He spoke warmly of the “final” interview: in this the candidate was taken into a room filled with men dressed as superior officers. They told him that he had done fine and now had only one more test to pass. This was a test of binocular vision in which he was to peer into a box and by turning handles bring two objects together. Unknown to him, one of the onlookers would manipulate a knob which made it impossible to accomplish the task. While the poor devil of a candidate fumbled away, the officers would begin making disparaging remarks. On the very threshold of acceptance, he was told that he had failed. The onlookers then observed his reaction.

  The consultant explained that for obvious reasons this sort of thing could not be done in an industrial situation. But the principle could be used. He explained his interview technique. “I sit down with a man with a record of the tests and his vital data in front of me. I am very friendly with him. It is a tense situation, and by increasing the stress I can get him to reveal much more of himself. For example, I will be running down the list and I will read aloud ‘married: seventeen years.’ Then I will read ‘children: none.’ I will let my eyebrows go up just a little and then pause thoughtfully. He is probably very sensitive on this point and in a few minutes he will begin to blurt out something about his wife or himself being sterile, and how maybe they have seen doctors about it. Just at this point I may ask him how his sexual relations with his wife are now. After several more minutes of stress I build him up again. Toward the end of the interview I usually smile and say, ‘Well, why don’t we stop while we’re ahead.’ That makes him relax and makes him think everything is going to work out all right. Then I shoot a really tough one at him. He is caught off guard.”

  I am not trying to suggest that testers are abnormal, though there would be, I am tempted to add, a certain poetic injustice in such a suggestion. When they are confronted with recalcitrance or criticism, many testers, as is so characteristic of followers of scientism, do not address themselves to the ideas in dispute but speculate, instead, on the hidden maladjustments that drove one to take a position contrary to theirs. They use sympathy like a weapon.

  But turnabout would not be fair play. Most testers are fair and as normal as the next fellow, and as far as underlying hostilities of their own are concerned, they would be abnormal if they didn’t have some. They would also be abnormal if they suppressed them completely. The interviewer is sorely tempted to play God—and if the difference in age or salary or background or temperament between himself and the person he is evaluating is wide, the temptation to play somebody else is strong too. This can be resisted by the man with great insight into himself as well as others, the man with wisdom, forbearance, humility. To all such men in the testing field my remarks do not apply.

  We have been talking of how testers interpret answers; now let’s turn to the questions themselves. Are they free of values? In designing the questions, testers are inevitably influenced by the customs and values of their particular world. Questions designed to find a man’s degree of sociability are an example. Do you read books? The reading of a book in some groups is an unsocial act, and the person who confesses he has at times preferred books to companions might have to be quite introverted to do such a thing. But the question is relative. Applied to someone brought up in an envi
ronment where reading books is normal—indeed, an excellent subject for much social conversation—the hidden “value judgment” built into the test can give a totally unobjective result. People are not always social in the same terms. A person who would earn himself an unsocial score by saying he would prefer reading to bowling with the gang is not necessarily unsocial and he might even be a strong extrovert. It could be that he just doesn’t like bowling.

  If the layman gags at the phrasing of a question, testers reply, sometimes with a superior chuckle, this is merely a matter of “face validity.” They concede that it is better if the questions seem to make sense, but they claim that the questions are not so important as the way large numbers of people have answered them over a period of time. To put it in another way, if a hundred contented supervisors overwhelmingly answer a particular question in a certain way, this means something, and thus no matter whether the question is nonsensical or not, it has produced a meaningful correlation coefficient.

  Meaning what? This is not the place to go into a lengthy dissertation on statistics, but two points should be made about the impressive test charts and tables that so often paralyze common sense. A large proportion of the mathematics is purely internal—that is, test results are compared with other test results rather than with external evidence. Now, this internal mathematics is valuable in determining a test’s “reliability”—that is, whether it is consistent in its measurements. If a group of people take Form B of a test, for example, and a mathematical correlation shows that their percentile scores rank just about as they did when they took Form A of the same test, we have an indication of the test’s reliability in measuring something.