The Organization Man Read online

  But what is that something? A test’s reliability tells us little about its validity. A test may give eminently consistent results, but the results are worthless unless it can be determined that the test is actually measuring the trait it is supposed to measure. Do the tests measure sociability or introversion or neurotic tendencies? Or do they merely measure the number of times an accumulation of questions about putting out fires or reading certain books will be answered certain ways?

  To show a test is valid, scores must be related to subsequent behavior of the people tested. Examine the “validation” evidence for many tests, however, and you will find they consist chiefly of showing how closely the average scores for the particular test come to the average scores of somebody else’s test. That there should be a correlation between test scores is hardly surprising. Test authors are forever borrowing questions from one another (some questions have been reincarnated in as many as ten or twelve different tests) and what the correlations largely prove is how incestuous tests can be.

  But how much have scores been related to individual behavior? Among themselves psychologists raise the same question, and for muted savagery there is nothing to match the critiques they make of one another’s tests.* The Bernreuter Personality Inventory is a particular case in point. This is by far the most widely used test in business (1953 sales by Stanford University Press, one of several distributors: 1,000,000 copies). Yet a reading of the professional journals shows many reports on it to have been adverse. Some psychologists checked Bernreuter scores against other, more objective, evidence of what the people tested were like and found no significant relationships, and sometimes reverse correlations. “It must be concluded,” writes Cecil Patterson in the Journal of Social Psychology (24, 3-50), “that the results of studies using the Bernreuter Personality Inventory are almost unanimously negative as far as the finding of significant relationships with other variables is concerned. … It is no doubt largely due to the nature of the questionnaire approach, which appears to be a fruitless technique for the study of personality.”

  As top psychologists point out, a really rigorous validation would demand that a firm hire all comers for a period of time, test them, seal away the tests so the scores would not prejudice superiors, and then, several years later, unseal the scores and match them against the actual performance of the individuals involved.* This has rarely been so much as attempted. Dr. Robert L. Thorndike of Columbia Teachers College, certainly no fortress of anti-science, points out that most follow-up studies of personality tests known to the field are “contaminated.” To make a genuine validation, Thorndike says, it is necessary “to apply the procedure and make the evaluation … to keep the results out of the hands of the operating people who can control each man’s career and the appraisal of it … to get appraisals of job success that are entirely independent of the [original] appraisal, and then to bring together the two independent sets of data.”

  Comparisons have been made between groups that have been tested—for example, a group considered productive may be found to have had an average score on a particular test higher than that of another group less productive. The average of a group, however, tells us very little about the individuals involved. Invariably, some of the people in the “best” group will have lower test scores than some of those in the “poor” ones.

  Testers evade this abyss by relying on a whole battery of tests rather than on just one or two. But no matter how many variables you add you cannot make a constant of them. If a man has a high “contentment index” and at the same time a very high “irritability index,” does the one good cancel out the other bad? Frequently the tester finds himself right back where he started from. If he is a perceptive man he may pay little heed to the scores and make a very accurate prognosis; if the prognosis later turns out to be correct, however, this will be adduced as one more bit of evidence of the amazing accuracy of the tests.

  As an example of how values get entwined in testing, let’s watch as a “Worthington Personal History” is constructed. This particular technique is a “projective” one; it is so projective, as a matter of fact, that the tester doesn’t even have to bother seeing the man at all. The client company has the applicant fill out what seems to be an innocuous vital-statistics blank and then mails it to Worthington Associates at Chicago. There an analyst studies it for such clues as whether the man used check marks or underlinings, how he designated his relatives, what part of his name he gives by initials, what by the full word. He is then ready to reconstruct the man.

  Let the reader match his diagnostic skill against the testers. In a hypothetical example given in the Personnel Psychology journal, we are told that the applicant, one Jonathan Jasper Jones, Jr., writes down that he is twenty-six, that the date of his marriage was November 1951, that he has only one dependent, his wife, that her given name is Bernadine Butterfield, her age twenty-eight, and that she works as a nurse-receptionist in a doctor’s office.

  Enigma? Carefully, the analyst writes that these clues indicate that Jones “may not take his general obligations very seriously, may have a tendency to self-importance, wishful thinking, may be inclined to be passive-dependent—i.e., reliant on others for direction and guidance—in his general work relationships, may be inclined to take pleasure in unearned status or reflected glory.”

  How in the world did the analyst deduce all this? Let us peek into the laboratory. The name was the first clue. It told the analyst that Jones was narcissistic. Whenever a man writes his name out in full, the tester notes in his chart that he is “narcissistic.” Two initials, last name: “hypomanic.” First initial, middle and last name: “narcissistic, histrionic.” Any erasure or retracing: “anxious, tense.” Characteristically, the applicant gets the short end of every assumption. Even if he adopts the customary first name, middle initial, last name, he is put down as “mildly compulsive.”

  Deeper and deeper the analyst goes:

  Spouse’s age: twenty-eight

  Fact: married a girl two years older than himself.

  Empirical observation: the majority of men marry women younger than themselves.

  Primary deduction: may have been influenced in marrying her by unconsciously considering her as a mother-surrogate.

  Tentative inference: may like to have an older woman, or people, take care of him.

  Provisional extension: may be inclined to be passive-dependent—i.e., reliant on others for direction and guidance—in his general work relationships.

  And so on.

  Analysts, of course, have just as much right to read between the lines as the next man. What makes their posture interesting is the claim that theirs is the scientific method. There are all the inevitable tables (“the biserial correlation for tenure on Worthington cutoff is 0.34”) in which the accuracy of the internal mathematics is confused with the accuracy of the premises. There is the usual spurious humility: the authors caution that no single deduction is necessarily correct; it is only when all the deductions are fitted into a “formal, quantitative scoring procedure based on an organized system of psychodynamic factors” that they become correct. Add up enough wrongs, in short, and you get a right.

  The analysis of J. J. Jones is a rather extreme example but it leads us to a defect characteristic of all tests. Against what specifications is the man being measured? Assuming for the moment that we are able to diagram Jonathan Jones exactly, how are we to know the kind of work the diagram indicates he is suited for? There is not much point in testing people to find out if they will make good salesmen or executives unless we know what it is that makes the good ones good.

  There are, of course, some fairly simple common denominators such as energy and intelligence, but beyond these we get into very muddy waters indeed. Is the executive, as is so often assumed, one who was more attached to his father than his mother? And what, for example, of the liking for other people that is supposed to be a key executive trait? Many successful executives are very gregarious, but a great man
y equally successful ones are not, and some are very cold fish. Conversely, are writers, research people, and the like necessarily introverts? One writer I know rather well was advised by a testing laboratory to go into clerical work and steer clear of writing. While he scored just slightly “introverted,” the laboratory still felt that his score indicated enough enjoyment of the company of other people to prove him unsuited for writing. Another laboratory tells people whether they are good at communication by determining their “ability to write freely without blocking or searching for the right word or phrase.” You are given a topic and so many minutes in which to write on it. No matter how much gibberish it may be, the more words you write on this glibness test the more “creative” you are rated to be.

  Testers often admit their knowledge of personality requirements is still too sketchy, but they feel this can be cleared up by more testing. And thus we come to the “profile.” Testers have been busy collating in chart form the personality scores of people in different occupations to reveal how they differ with other adult groups on particular personality traits. These comparisons are generally expressed as a “percentile” rating—if the total of thirty salesclerks’ scores on sociability averages somewhere around the 8oth percentile, for example, this indicates that the average salesclerk is more sociable than 79 out of 100 adults. With such data a man being considered for a particular kind of job can be matched against the master profile of the group. The closer he fits the better for him.

  Profiles are also worked up for work in individual companies. At Sears, Roebuck there are charts that diagram the optimum balance of qualities required. Here is the one an executive values:

  A man does not have to match this profile exactly, but it won’t help him at all if his line zigs where the chart zags. Take a man who scores considerably higher than the 10th percentile on aesthetic values, for example; such people, Sears notes, “accept artistic beauty and taste as a fundamental standard of life. This is not a factor which makes for executive success. … Generally, cultural considerations are not important to Sears executives, and there is evidence that such interests are detrimental to success.”

  Sears has every right to de-emphasize certain qualities and emphasize others; and in hewing to this type, it should be noted, Sears has built up one of the most alert management groups in the country. But the process should not be confused with science. When tests are used as selection devices, they are not a neutral tool; they become a large factor in the very equation they purport to measure. For one thing, the tests tend to screen out—or repel—those who would upset the correlation. If a man can’t get into the company in the first place because he isn’t the company type, he can’t very well get to be an executive in it and be tested in a study to find out what kind of profile subsequent executives should match. Long before personality tests were invented, of course, plenty of companies had proved that if you hire only people of a certain type, then all your successful men will be people of that type. But no one confused this with the immutable laws of science.

  Bias doesn’t have to be personal. Now it’s institutionalized. For the profile is self-confirming; when it doesn’t screen out those who fail to match it, it will mask the amount of deviance in the people who do pass. Few test takers can believe the flagrantly silly statement in the preamble to many tests that there are “no right or wrong answers.” There wouldn’t be much point in the company’s giving the test if some answers weren’t regarded as better than others. Telling the truth about yourself is difficult in any event. When someone is likely to reward you if you give answers favorable to yourself the problem of whether to tell the truth becomes more than insuperable; it becomes irrelevant. In this respect, let us give the tests their due; they may tell little about personality, but they do tell something about intelligence. “Do you daydream frequently?” In many companies a man either so honest or so stupid as to answer “yes” would be well advised to look elsewhere for employment.

  Even when the man who should have looked elsewhere slips through, the profile will be self-confirming. For the profile molds as well as chooses; it is, as Sears puts it, a statement of “the kind of behavior we have found to be desirable.” If he is going to get ahead, let alone survive, the man is going to have to adjust. Several years of give and take, and the organization will have smoothed him out, and thus when the psychologists do their “validating,” or rechecking, later he will score near enough to the median to show them how right they were all along. At Sears psychologists retested employees about two and a half years after their first test. Of the thirteen personality factors they were investigating, nine showed a definite change. “All the factors,” concluded Sears’ V. C. Benz, “indicate a ‘better’ or more adequate psychological adjustment. … The individuals are better people.”

  Up to a point the company “type” has some virtue; any first-rate organization must have an esprit de corps, and this requires a certain degree of homogeneity. But the pitfalls are many, for while a self-confirming profile makes for a comfortable organization, it eventually can make for a static one. Even the largest corporations must respond to changes in the environment; a settled company may have its very existence threatened by technological advances unless it makes a bold shift to a new type of market. What, then, of the pruning and molding that adapted it so beautifully to its original environment? The dinosaur was a formidable animal.

  The profile can be self-confirming for the individual too. for tests intensify a mutual deception we practice on one another. Who is “normal”? All of us to some degree have a built-in urge to adjust to what we conceive as the norm, and in our search we can come to feel that in the vast ocean of normality that surrounds us only we are different. We are the victims of one another’s façades.

  And now, with the norm formally enshrined in figures, we are more vulnerable than ever to this tyrant. “Science” seems its ally, and thus, faulty or not, the diagnosis can provoke a sense of guilt or inadequacy; for we can forget that the norm is often the result of the instinctive striving of previous test takers to answer as they think everyone else would answer.

  If the organization man escapes the danger of self-tyranny he faces another. At first superiors may scoff at the diagnosis, but if they have been putting reliance on testing they have a stake themselves in the correctness. Suspicion, unfortunately, demands proof, and sometimes it so counterbalances judgment that unconsciously a management will punish the man so that faith in the tests may be confirmed. One large midwestern company was about to promote a man when it decided to have him take a test. The report that the consultant firm mailed back to the company was freighted with warnings about the man’s stability. The company was puzzled. The man had consistently done a fine job. Still. … The more the company mused, the more worried it became; at last it decided to tell the man the promotion he had expected so long was going to someone else. Six months later, the company reports, the man had a nervous breakdown. As in all such stories, the company says what this proves is how accurate the test was.

  Are the people who don’t score well necessarily the misfits? Almost by definition the dynamic person is an exception—and where aptitude tests reward, personality tests often punish him. Look at a cross section of profiles and you will see three denominators shining through: extroversion, disinterest in the arts, and a cheerful acceptance of the status quo. Test scoring keys reveal the same bias. As I note in the Appendix, if you want to get a high score you will do well to observe these two rules:

  (1) When asked for word associations or comments about the world, give the most conventional, run-of-the-mill, pedestrian answer possible.

  (2) When in doubt about the most beneficial answer to any question, repeat to yourself:

  I loved my father and my mother, but my father a little bit more.

  I like things pretty much the way they are.

  I never worry much about anything.

  I don’t care for books or music much.

  I love my wife a
nd children.

  I don’t let them get in the way of company work.

  If you were this kind of person you wouldn’t get very far, but, unfortunately, you won’t get very far unless you can seem to be this kind. Check the norms and you will find that the advice is not flippant.* Norms are based on group scores, and as often as not the group turns out to have been 1,000 college freshmen, 400 high-school students, or some similarly less-than-outstanding aggregation. There are norms for selected groups like executives, chemists, and the like, and there will in time be more and more. But these can be illusory too. Usually, the norms are based on the responses of people who are being tested by The Organization; in such a situation self-preservation demands a certain circumspection in answering, and the norms are more a playback of what the people taking the tests think The Organization wants to hear rather than the reality of their own selves. Daydream? Of course not.

  The sheer mechanics of the tests punish the exceptional man. A test with prefabricated answers is precisely the kind of test that people with superior intelligence find hardest to answer. Writing of the using of “objective,” forced-choice college examinations, Jacques Barzun says in his Teacher in America, “I have kept track for some ten years of the effects of such tests on the upper half of each class. The best men go down one grade and the next best go up. It is not hard to see why. The second-rate do well in school and in life because of their ability to grasp what is accepted and conventional. … But first-rate men are rarer and equally indispensable. … To them, a ready-made question is an obstacle. It paralyzes thought by cutting off all connections but one. Or else it sets them to thinking and doubting whether in that form any of the possible answers really fits. Their minds have finer adjustments, more imagination, which the test deliberately penalizes as encumbrances.”

  If the reader has taken the test in the previous chapter he has probably come to the same conclusion about personality-test questions. How big was that fire in the house with the mother and child? What about the worker who wanted a promotion? Was he any good? Maybe the office did want the boss’s nephew to get the job. These are not quibbles; they are the kind of questions that occur to the intelligent mind, and the ability to see shadings, to posit alternatives, is virtually indispensable to judgment, practical or otherwise.