The American Scene; Predicting Success

Good morning, Mr. Jules, and I'm here to see our subjects morning and evening recess. Specifically, I'm concerned with the youth and the government whether or not a person has a chance of becoming successful in any given endeavor. Good morning, Mr. Jules, anger for the American scene. Our subject this morning is predicting success, and specifically we're concerned with the use of tests to determine whether or not a person can has a chance of becoming successful in any given endeavor. Testing since World War I has taken an increasingly large chunk out of the American business endeavor. We're told that 10 years ago over 60 million tests were given. The figure has been estimated has doubled by today. That's 120 million tests. Testing has become very popular. On the other hand, however, testing has elicited a tremendous amount of critical

comment. Criticism directed against the techniques of testing, the philosophy of testing, the men who do the tests, the motives for testing. Discuss some of these problems with us this morning. Some of the uses of testing, the future of testing. We have two guests. Our first guest is Mr. Richard. I'm Smith, who is the Director of Educational Services, the Illinois Belt Telephone Company. Our second guest is the George S. Spear, Director of the Institute for Psychological Services. I wonder if we might start, Mr. Smith, you in your capacity as an administrator for a major employer personally on the state of Illinois. You must use some tests. What do you test for? What do you use? Yes, we do use quite a few tests. We use tests principally for predicting mental ability or for predicting certain basic aptitudes or skills. We give all of the women and all the men that come to work for us

regardless of what the job is. Test of their mental ability. Mr. Standard, what do you mean by an IQ test? It's an IQ type test. Only in a few instances do we give a test that comes up with an actual IQ. We give trade tests in addition to that such as Comptometer Test or Typing Test. The test of particular skill. The ability to do a specific job. That's right. We have several tests that are developed for specific jobs which we have that are not common to industry in general and for which there have not been real good test developed. I see. For example, an information operator has more or less unusual job and we have test developed to enable us to determine whether or not she can do it. Mr. Speer, what else is being tested for today? This handles two of the areas. General mental ability, specific skills. There are a variety of traits that are

tested for these days. You have the mental ability, intellectual capacity of one type or another. There are various skills as Mr. Smith has already referred to. Achievement in particularly in school situations where you're measuring how much has been learned or in some cases you want to know how much the individual knows as a basis for determining whether or not he is ready to take a next subject, the next step. Might this conceivably be a college entrance exam. Mr. Speer, it definitely would be of this type usually. You're measuring his readiness for the next step in the academic work. How much has he learned and profited by his high school program. Then there are diagnostic tests in particular in the educational field to determine what kinds of problems an individual may have with math, arithmetic, reading, things of this sort,

personality, interest, motivation. You name it as a test. That's what I thought. You say personality. In other words, that whole area called personality testing. This particular area has been one of the objects of criticism. Probably I guess because it involves essentially a different kind of test than the test of a skill of an ability. I suppose there's a difference between wondering whether or not we can predict success for a truck driver and for an account executive. For a truck driver, I'm one of the best tests I guess is to give him a truck and watch him handle it. An account executive, however, here you're concerned with not only the things he's acquired, the skills he possesses, but certain personality traits, judgment, and so on. Can these be predicted? Mr. Smith, let me say that I think they can. You don't test for personality. We do not test for personality. One of the main reasons that we don't test for personality is that we hire very few people, practically nobody,

other than those people that we hire for starting jobs in the company. Starting jobs are not as restrictive personality wise as more advanced. Now, we've looked at personality testing. We've looked at it one specific instance. We looked at it on the basis of selecting salesman through personality test. We take people who met the mental requirements that we've set up for a salesman's job and we would give them a personality test to try to determine whether they had the proper motivation and the proper personality characteristics to perform well as a salesman. What happened was that we found in trying to get a criterion or a measure of what is a successful salesman, that the unsuccessful and the most successful had very similar profiles on the test that we used. There was a large group in the center that were pretty much spread all over the map, but we

gave the test and we assessed the ability of the men who were being tested. We had their bosses evaluate them and we found that on that job, which is a more or less specialized job in our business, that the test, which we chose, and it was one of the highly touted tests in the field, did not predict the things that we were looking for. We could not predict whether salesman would be successful or unsuccessful as a result of that test. Mr. Spear, why would such a test have failed? Could you guess? Well, I have to figure out a little background first to answer. I think there are some very obvious reasons why it could have failed. One is that the tests that you use may not have measured traits that were successful in the job. Part of the basic part of using tests to predict success is first a definition of the job itself, what is necessary for success in the job. And a good test, which measures trait A, won't do you any good in

predicting success in a job that depends on trait B, so that you may have used a test which did not measure the factor. No, the tests that we did use measured 15 basic personality tests, which are supposedly inherent in all people in one degree or another and pretty well cover the field. So we didn't say test for aggressiveness and say endurance or order or something like that. We tested for the full range of personality characteristics. A second possibility is that even though you were measuring the right traits with a test which would measure those traits, that your interpretation of the test scores was inadequate. I'm not going to say that that's what happened. This could happen in a situation. A third possibility, of course, and one that I suspect is true in your case is that the job is one which is influenced primarily by factors other than personality. Certainly, this is true in most of the jobs for which you're testing. The personality

range. Well, I think permitted of all jobs. Not only sales. Well, there are numbers. No, personality is not important. No, personality is important. But personality is only one of the things. Well, yeah, I agree. But the degree to which it's important varies from one job to another. Could we just go back a little bit? You raised, it seems to me, a number of points. You started out, first of all, with a question whether traits tested for are really significant in terms of the end result desired. You may have discovered an excellent test for determining, I don't know, puff paste bakers. Only looking for salesmen. Well, how do you know what other particular qualities that are shared by a salesman? In fact, can you even make the assumption that there are, that all good salesmen have certain qualities in common? I don't think you can make an assumption that all good of anything has something common to it. This is a broad generalization. The usual procedure, and I want to suspect the telephone company has made and has followed in studying its jobs, is a study of the people who are successful on the job. A study of the factors in the job itself, for a salesman,

it may be a necessity for a relatively high level. There's one, one thing, pardon me, there's one thing. You not only have to test, or study those who are successful, but those who are unsuccessful. That's right, sure. Which is what we tried to do. We tested a group, a complete district, all a salesman in a district. And we had some in there who were very successful, and some who were not successful at all. This, I find, hard to believe, and is another reason why I think your testing may have fallen down. I suspect that you may not have had a wide enough range of ability or performance on the jobs that you have. This is one of the big problems, because normally, if a salesman isn't any good, you don't keep him on the payroll. Therefore, you don't have a large number of subjects in your population to study to compare with a successful one. That's true. Well, that is true, and I think in any situation you're going to get. What

I think you did was to compare the very successful with the average successful, and therefore to have insufficient spread or differentiation in your pattern, to be able to predict from, for the individual who is being considered for entry into the job. How do you make the selection? Is that on the basis of test? Well, you're partly right on that, but you're not fully right. We give these men quite a long, I won't say apprenticeship, but a fellow who doesn't work out isn't immediately discarded from you. Well, could we do make a pretty good selection prior to the time that we bring people in? How do you make the selection? Is that on the basis of tests? Yes, and no. All of the people we hired are tested. At the time we hired, and they're tested for these people are all tested for mental, the eye -cute type of test. So we know, when they come on the job, that they have

enough mental capacity, or they can learn how to do the job. The things you're not testing for, however, are whether they'll want to learn, and whether once they've learned, they'll continue to work it up, and so on. How do you do this? We have pretty well -trained interviewers, who I admit that there isn't any interviewer who's completely objective. Every interviewer has certain types of people he likes better than others, just like you like certain people better than others, and so do I. But they get so that they have a pretty good idea, the kind of person who will fit into the business, and who has been successful in similar situations, and there's probably two or three points that they're looking for. They're observing, and they're looking for, but they're not testing. Mr. Trying to find the other way. If you had to choose between interviewers and tests, what would you use? Well, I think we'd use the interviewers. The

test would never, I don't care how proficient we'd become in testing, and what I would call a foreseeable future. We will not be able to do away with the interviewer. Mr. Spear, doesn't the test, though, have certain advantages over the interviewer? Oh, yes. I'm a little surprised that Mr. Smith would make the answer that he did. Frankly, I would say if I had to choose between them, I couldn't. I don't think that you can do a really complete job without both tests and interviews. Certainly, you can't depend solely on tests unless you have a large population from which to choose, and you are making your selection primarily on the basis of a numerical score, as in a scholarship examination, where you're testing several thousands. You can take the top 10 percent, and well, you may miss some. You're close enough so that there isn't too much difficulty. Just a parenthetical question here. Would you say that the testing for industry is best employed, the most successfully employed, on a large level industry? Testing in general is most

successful statistically when it's used in large groups. The plant hiring 40 people will be foolish for such a plan to have a test device specifically for a particular problem. I'm saying that the testing alone for large populations, but there are very few situations of that kind, other than a state wide scholarship examination, or the draft deferment service, something of that kind, where you're testing thousands, hundreds of thousands. But for industry, the test should be a part of a total personnel selection, which starts with recruiting. Where do you look for people? And what kind of people do you look for? How do you interview them? What do you want to know about them? What kind of an application blank do you get? And then what kinds of tests are you going to apply? And then what are you going to do with this total pattern? Anyone who depends on tests alone to select his personnel is going to make an awful lot of mistakes. Aren't there testing associations which offer tests to supply all other activities? Sure.

I can't encourage that. I can't encourage that kind of thing, but there are. Sure. Well, I wonder this, of course, this total testing procedure involving tests, interviews, and so on is ideal, assuming that the results of all of these inquiries coincide. That's when this conflict, the interviewer says this man has splendid. He's just what I want. He reminds me of my dear departed father. On the other hand, the test says, mm -hmm, won't do it all. He should look elsewhere. I'd say if there's a discrepancy between the interviewer's impression or the past employment record and a testing program. And you're talking about people who have passed the employment record. Yes, or if there are young ones who haven't worked at all, then you have the interview and school record that's about all you can get. But if the past record, whatever it may be of the individual, and the interviewer's impression, disagree with the indications of a well -selected testing program. What I'd say the first step is to go back

and look at this record more carefully. Mr. Smith, Mr. Spears presentation is theoretical. Yours is practical. Have you ever been faced by this situation? I mean, we're a test, we'll say, one thing, an interviewer will say something. I think in the type of testing we do, and you're working, we're many times faced with that. Because what we're testing for are those qualities which the interviewer cannot assess. We're talking about, we're testing in the areas of basic mental ability, the ability to solve certain problems and the ability to apply a certain mental capacity to the job that we're hiring. Well, the interviewer, and I think this is pretty generally agreed, cannot assess that quality. We have many people, and I can think of a good instance. I was going to say we had many people come in who are very well liked, and who take the tests, and we find they haven't got quite what we thought they

had, and we turn them away from that reason. We had an instance of a college graduate who came in. He took the General Intelligence Test that we give. I think he scored very, very low on that. If he had any mathematical knowledge, it wasn't apparent who the test. The department that was looking at him was looking for people with primarily verbal ability, and he apparently was a good looking young fellow. He'd done fairly well in school, which is possible without having enough intelligence to do well in an industrial situation. Depends on the school. Well, it depends on the school and the course, and a lot of factors, which we look at too. But the department was crazy about it. They thought he was wonderful. They talked to other people that knew him. They all thought he was wonderful. But he was so far down on the measurement,

or the criterion that we'd set up, the base under our mental ability measurement. We don't go to him and say, well, you can't have him, but we go to him and say, well, now, maybe you better think again, and they think again. Still come up with the same answer. We get a little more explicit. We don't ever say that you can't hire this fellow because he isn't smart enough. We say that we don't think he's quite as smart as the people that we're bringing into the business, and he ought to have some other pretty good qualities to be able to do the job. Well, there is some possibility of placement, also, in many cases. Incidentally, Jules, my previous comments may have been theoretical, but they're not based purely on theory. They're also based on practical experience. In large organizations, you can test an individual for one job and discover that he doesn't fit that job, but that he might be very successful in another one. You don't have to necessarily have a high degree of mathematical ability to succeed

in a sales job, but if you're tested as an engineer, the test will turn you down if you don't have a bad ability. And I'm sure that this might well be true in this particular situation, and you're thinking, well, I think that would be true. But he might go across the street and be a big success selling paper products or some of that kind. A matter of prediction is more than a general thing. It's a specific thing. You're trying to predict, is he going to be successful in this particular job? In other words, when you have to start out, if you're going to construct a test, you have to have a pretty good idea of what you mean by success in a particular situation, and build this into the test itself. Yeah, it's a pretty good idea of what you're trying to predict, as well as what is important for the success in that. I think we're probably the same as many other industries in this line. We hire people to come to work for us, and we anticipate that they're going to work for us for a long time. Women, of course, come in, and the turnover is rather high, but if you get into men

and you get any of the more important jobs in the company, which are filled principally by men, those people are not hired off the street. Those people have practically all come up through the ranks of the business. Now, when we bring somebody into a management job in almost any department, it's anticipated that he will meet the requirements, not only of that job, but of other jobs, if he's going to be very successful. So, while we might say we have a manager -level job in a certain situation, and this man could specifically fill that job, he has a requirement, say it's a job that requires very high mathematical ability, and he has a lot of mathematical ability and very little verbal ability. We'd say, well, he appears to have the qualifications for this job, but this job is a very exceptional job, so that he probably won't fit in elsewhere. Well, I don't know if we were testing that we'd ever promote a man under those conditions, because we want a

man that can move. At very best, kind of the test is a tool, which helps you in any kind of an act of total judgment. I think testing is a tool that is in addition to the application blank, the references that the man has, the interviewer's judgment. I think that on certain basic abilities that we should set a floor under which we will not go. We set a base and say, man, must have this much of this quality, and I'm talking about abilities or skills. I think Jules, I think Mr. Smith is talking about one particular kind of test situation, however, that is an entry job. You're testing people that come into the organization, and you're interested in a minimal level of ability. They have enough ability to succeed. Then you promote from within on the basis of what you know about the man after he's been on the job. You're not in the situation where you have to predict from test the qualities of

aggressiveness or ambition, determination, adjustment with other people, or emotional stability and that kind of thing, so that he is not in a situation where personality testing is going to contribute very much. The main thing are these skills and intellectual functioning, and if he's got that, we can put him to work. But think in terms of a situation where you are trying to select a man for a sales job or a managerial job to begin with. Here is where personality testing and the more comprehensive and thoroughly evaluation of the individual through systematic testing and use of questionnaires becomes significant. Because you don't have three years to decide whether or not he will get along with somebody, you've got to decide, no, and you use... How much do you have? I know that certain tests, for example, are administered in 12 minutes flight. It's a kind of IQ test. And other tests, well, the three -year variety, Mr. Smith speaking, out of the test and experience. If you want to test a man who's going to go into a situation

such as you describe one that will demand certain personality qualities, how do you test them? What do you test for? What's the battery of tests you're exposing to? Well, that's hard to say, but a general rule would be that the amount of testing and interviewing time by the psychologist would be roughly proportional to the level of responsibility of the job and its complexity. But you would use both then testing and interviewing. Oh, yes. Yes, I see. In most of our jobs, though not necessarily in all of them, and most of the things for which we select people, we interview as well. But our interviewers are our trained interviewers, as yours are. I imagine one ideal of the test is that someday might, and I'm just going to get caught, can. But possibly that it should someday supplant the interviewer, if only in terms of economy. The interviewers can see much fewer people. No, the interviewer has a entirely different function. I don't know what... Ten the function of the interviewer... Oh, I would agree with that. And I don't think... When we start talking about tests, we're not talking about economy. No, no. The interviewer... Test cost money.

...is intended to see the individual as a total personality. And the test is intended to see him as certain slices of his behavior. His ability to respond and to act in particular kinds of situations. I would like to say the tests don't cost money. They save money. You're right. And I think that was very well stated. Your qualification was if they are properly selected correctly interpreted. And this is, of course, the critical question. That's right. I'll have to add a little bit to... I wouldn't want George to think I have no use for personality testing at all, because I do think that there's a place, and I think that in the future there'll be a greater place for it, as we learn more about it. But personality tests are something that must be used in the hands of highly trained people. While we train the people that we have hiring, doing the hiring job, the amount of time... They look at... In our general employment office alone, they look at 20 ,000 people a year. Well, you

can't spend a long time looking at 20 ,000 people, because you don't have that many hours in the day or that many of people... That many people on the job. So I think that personality tests have a place, and in the future we hope that they'll have a place in the telephone business. Do you really have to start them? In the hands... No, in the hands of very highly skilled people. Maybe the right tests aren't predicted, or aren't... Haven't been developed. I am a little inclined to think that the problem is one more of criterion than tests. I think that many of the present tests are transparent, although from our previous conversation I gather, you make people believe that the best thing they can do for themselves and for the job is to answer the test honestly. Certainly, we try to do that. And I think in the long run, that's correct. There's no question it's correct. The only thing is that the man's anxiety for his job is very great, or for this job that he's chosen. For this job. It's very great. He will tend to see himself maybe as he'd like to have you see. I

wonder if we have a very short time now, but there's one area we haven't discussed at all in regard to testing. And that's the possibility of testing young people, as early as possible, to determine their potentials for success, their talent for mathematics, for science. Well, let me say that the best one I've had so far was an engineer who wanted me to test his son to determine whether or not his son could also be an engineer. And I asked how old the son was and he said two weeks. We told him to come back in 16 years. At 16 years, you feel it can be done. We think we can get some significant idea. Do you feel that should be done in the public schools? That's kind of testing. Well, I think it should be done. Whether it should be done in the public school depends on the staff, the time, the amount of taxes that the taxpayers are willing to provide, and so on. That's it. They do cost money. I could tell you quite a bit about that. I think I'm afraid we'll have to put that up for some other time. No, I'd like to thank you both. Mr. Smith, who is the Illinois Bell, Mr. George Spear, Illinois Institute of Technology. Good morning for the American scene.

Mr. Jules Hanger.