Problems in the Use of Indirect Methods of Attitude Measurements BY IRVING R. WESCHLER / Cpf1^ee INSTITUTE OF INDUSTRIAL RELATIONS INwTaTun OF IUSOXIAL RELATIONS UBRARY UNIVERSITY OFLCORNIA LOS ANG 24, CAUFRIA Problems in the Use of Indirect Methods of Attitude Measurements BY IRVING R. WESCHLER. Social scientasts are mating increasing use of various indirect techniques of attitude assessment. These techniques, ranging from projective devices to hidden intelligence tests, are intended to elicit rdeeplying", attitudes or personality characteristics which might not otherwise be accessible to the investigator. In this article, the author argues that while these techniques may have considerable scientific value, their use and susceptibility to misuse raise serious problems of both an ethical and a practical character. Dr. Weschler suggests that the formulation of a code of procedure governing the use of such instruments may ultimately be necessary. This article is an expanded version of a paper presented to the annual convention of the American Psychological Association in September, 1950. The thor is a Research Assistant of the Institute of Industrial Relations and Lecturer in Psychology at the University of California in Los Angeles. au- The use of indirect methods of attitude measurement has recently come into vogue, and a number, of new techniques have been deveoped which supposedly get at those "deeper level" attitudes which a person may be interiorizing and unwilling to reveal.' These, indirect devices conceal from the individual the intent of the measurement and allow him to produce responses which would not be freely forthcoming if he were fearful of becoming personally involved. The purpose of this paper is to raine several as yet unresolved questions relating to the use of these new techniques.2 Indirect methods of attitude measurement may be constructed and used for a number of purposes. They are used to explore and test various psychological theories, especially those related to problems of learning and perception. The work of Murphy, Bruner, Postman and many others is dedicated to this particular type of interest. They are used to I See Campbell D. T., "*The Indirect Assessment of Social Attitudes," Psychological Bulkein, 47, I, January 1950, 15-38; also I. R. Weschler and R. Bernberg, "Indirect Methods of Attitude Measurement," Int. 1. of Opinion and Atitude Research, VoL 4, Summer 1950, pp. 209-259. 2 The author is indebted to Professors Franklin Fearing and Robert Tannenbaum and to Dr. William Schutz, Dr. Eugene Cogan and Mr. Murray Kahane for suggestions and helpful criticisms. Reprinted by permission of the Public Opinion Quarterly, Princeton University. PUBLIC OPINION QUARTERLY, SPRING 1951 134 construct reliable and valid test instru- union and management people, and was ments which may be valuable in the later incorporated as part of the test clinical situation as part of a test bat- battery in a study on "The Personal tery for the assessment of the total per- Factor in Labor Mediation." sonality. The work of Murray and his The results were, in general, as exassociates might be mentioned in this pected. Union members and students particular context. Finally, indirect at- who classified themselves as "pro-labor" titude measurement devices are applied scored high in the "pro-labor" direction, in the actual field situation for the as measured by the test, while the manmeasurement of attitudes held by the agement representatives and students members of various groups. declaring themselves to be "pro-management" scored low. When the test was A Case Study administred to the labor mediators, Typical of these studies, perhaps, is many mediators who were rated high one which the writer completed on "The by their colleagues in terms of their abilPersonal Factor in Labor Mediation," ity to do the job tended to score in the utilizing the "error-choice" technique "neutral zone," that is, near the sample for the measurement of attitudes to- population mean, while those who were ward labor and management.' For pur- rated as "poor" by their colleagues poses of illustration, I would like to scored either in the "pro-management" refer briefly to this study because some or in the "pro-labor" zones of the attiof the ethical and public relations prob- tude range. lems which arose during its progress are There was no difficulty in validating probably encountered in any kind of in- the first form of our "Labor Relations vestigation using indirect methods of Information Inventory" with the help attitude measurement. of UCLA students and various labor The "error-choice" technique, devel- and management groups, especially oped for attitude testing by Professor since the hypotheses and workings of Hammond,4 utilizes an information test the technique were explained after each which forces the respondent to choose administration of the test. Trouble came between two alternative answers, each for sundry reasons from the labor mediof which is by intent factually wrong, ators who were not ready to accept the or controversial, or of such a nature results which had been obtained. A lathat the correct answer is not easily ac- bor mediator usually sees himself as a cessible. This kind of test situation pro- "neutral" agent who, through his pervokes the respondent to select pseudo- sonal skill, is able to bring labor and facts from memory, and the "direction" management together in a settlement of of the error is measured as an indica- mutual satisfaction. Although many "bition of the respondent's attitude. ased" mediators, as measured by the Using the "error-choice" technique, ' See Weschler I. R., 'The Personal Factor the writer developed a test which was Mediation," Personnel Psychology, designed to measure both information inVol.LaborSummer 1950, pp. 113-133. 3, as well as attitudes in the field of labor 4 See Hammond K., "Measuring Attitudes relations. The test was validated on a by Error-Choice: An Indirect Method," I. Abgroup of students, as well as on active norm. Soc. Psychol., Vol. 43, I948, pp. 38-48. INDIRECT METHODS OF ATTITUDE MEASUREMENTS 135 test, were rated "good" by their col- direct attitude measurement techniques leagues on performance, the fact that keep the respondent in the dark about most of them scored far from neutral the true purpose of the test; or, putting on the test apparently became a threat it in less elegant terms, they use an eleto their personal security. The following ment of deceit to trick the respondent excerpt from a letter by one mediator is into answers which the experimenter an indication of the feeling which many considers more honest. The widespread use of various indiothers may have shared: methods of attitude measurement rect "I tend to agree with those media- creates a series of problems with ethical tors who participated in your survey as well as public relations implications, who feel that our confidence was vio- and any investigator who makes a decilated and abused when condusions sion about using these indirect methods were reached and publicized which might well consider these two related were based to some degree on 'load- aspects. Although many of us may have ed' questions. Those of us who agreed rationalized our use of these indirect to be 'guinea pigs' in your survey methods with the maxim "truthregard were assured that our replies would less of consequences," the time has come be held in strict confidence, and al- to analyze the consequences that are inthough I may not disagree very much volved. with the condusions which you have From. an ethical point of view, the reached, I am questioning the pro- social scientist must be concerned pripriety of using the materials which marily with the interests of his subjects you have collected." and should view with suspicion any It should be mentioned in passing that attitude investigation which endangers none of the individuals participating in the subjects' security. This is, in essence, the survey could in any way be identi- a "client-centered" point of view, which places the investigator under a moral fied. obligation to protect the goals and obTrickery or Scientific Method? jectives of his subjects and not to unWhen the attitude surveyor presents dertake any course of action which is his subjects with materials and instruc- harmful to their social, economic or tions which are not related to the stated psychological well-being. From a public relations point of view, purpose of his investigation, it becomes difficult to distinguish between honesty the social scientist is obliged to consider of purpose and deception. Members of only those practical aspects of his inthe public who are misled into offering vestigation which concern the smooth a glimpse into "the hidden crevices of functioning of relations with his subtheir soul," to use one of the time-worn jects or the general public. Public reladiches, are not likely to appear enthusi- tions minded, he has an "experimenterastic on discovering the hoax, even centered" point of view which looks though it may have been carried out primarily to the creation of a permissive for the noble scientific goals of obtaining atmosphere that makes possible the orknowledge and learning truth. The "er- derly progress of long range research. ror-choice" method and many other in- A research activity may prove to be un- PUBLIC OPINION QUARTERLY, SPRING 1951 136 wise from an ethical point of view, but public relations rather than an ethical if the researcher concerns himself main- problem. Without full participation by ly with the public relations aspects of his subjects, the social scientist cannot his investigation, he should be pre- hope to get results which accurately repared to deal with some of the follow- flcct the attitudes of his total population. ing problems: How to treat subjects His job, therefore, is to encourage parwho discover that they have been duped ticipation through an active educational into revealing their attitudes on one of program among the general public. The investigator who is unable to the new measurement devices, How to deal with the rising distrust of the pub- get the subject's permission to test his lic toward the techniques as well as the attitude may find it appropriate to utifindings of social research, How to reach lize some of the indirect techniques to the public which discovers the manner which reference has been made. In this of operation of these indirect devices, instance, the investigator gets cooperaHow to prevent possible misuse of the tion by involving the subject in a situatechniques which he is inventing, How tion which does not reveal to him the to encourage the public to participate true intent of the investigation. The in the increasing number of projects subject fails to give the experimenter which he is contemplating for the fu- permission to examine his attitudes, but ture. agrees to participate because he is unable to discern the true nature of the Questions That Need Asking investigator's intentions. Every experimenter who considers The Propriety of Deception. This utilizing indirect methods of attitude raises a second vital question that might measurement in his investigation might be posed by the social scientist at this profitably ask himself a series of ques- time: "Do I have the right to deceive tions whose answers will help him to people in order to get at their attideal with some of the ethical and public tudes?" It must be understood that the relations problems which he may have social scientist is in a different situation to face. The Right to Investigate. The first than the clinician who uses a variety of question, basic to any kind of attitude projective techniques for the purpose of investigation, might perhaps look some- helping the individual make a better, thing like this: "Do I have the right more healthy adjustment. The projecto investigate other people's attitudes?" tive tools which the dinician applies are A democratic society presumably pro- part of his diagnostic kit, similar to the tects the right of the individual to his many other devices which the regular personal privacy and there is no law, physician uses in his practice. ITe cliother than the Census law or perhaps ent knows the' intent of the therapist, some local ordinance, which forces him and even though he may not understand to participate in a polling activity. If or be convinced of the validity of the the respondent who recognizes the in- various projective techniques, he feels tent of the investigator refuses to take that the therapist has his best interests a stand on an issue which the social re- at heart. searcher is interested in, it illustrates a This relationship of trust and confi- INDIRECT METHODS OF ATTITUDE MEASUREMENTS 137 dence is usually not the case when the the right to deceive people in order to social scientist uses indirect methods to get at their attitudes, there is still an get at the attitudes of individuals who additional point which should be conare quite likely unwilling to divulge sidered. It may not take long before the their opinions through the use of any public in general "catches on" to the of the more direct techniques. The vio- operation of the various indirect techlent anti-Semite, the latent radical, the niques of attitude measurement. When arch conservative usually cannot be this occurs, the usefulness of these techidentified in the experimental test situa- niques will be greatly impeded, because tion unless devices arc used which pcnc- they depend for their effectiveness upon trate the protective cover with which hiding the purpose for which they are these individuals surround themselves. used. Furthermore, unless precautionary It should be kept in mind that in- measures are taken to prepare the pubdirect methods of attitude measurement lic for the type of investigations in vary greatly in the effectiveness of their which the social scientist expects it to disguise. Some of these tools hide only cooperate, it may look upon all operathe purpose of their utilization, and a tions of social science with rising scorn, sophisticated subject can easily see the distrust, or perhaps even fear. Misuse of Indirect Techniques. The many ways in which the results can be utilized. In this respect, the degree of third question which the social scientist indirectness of the attitude measurement should ask himself is: "Do I have the device is a function of the subject's right to report on new indirect attitude sophistication and depends greatly upon measurement devices, at a time when the frame of reference which the sub- these can be misused by unscrupulous ject brings to the testing situation. Thus, politicians or other selfish interests?" The present political and social clithe differential perception of the degree of indirectness produces a variable which mate abounds with instances of witchpartly accounts for the various degrees hunting, smearing of innocent reputaof tolerance and resistance with which tions, and attacking of people because the "duped" subjects react to their dis- of their political and social beliefs. It is covery of the real purpose of these test- quite easy to imagine that some of the ing devices. An analogy to what I have new indirect attitude measurement dein mind can be taken from the field of vices might be discovered by people in mental testing. Although an intelligence various kinds of inquisition movements, test usually uses straightforward direct and used by them for evil purposes. Almanipulations, many subjects are un- though it will undoubtedly take a long aware-at least while taking the test- time before any of these techniques is that their intelligence is being measured. valid for prediction at the individual Through the eyes of the unsophisti- level, I am sure that before too long cated subject, the intelligence test ap- enthusiastic and unscrupulous practipears as an indirect method although tioners may find these techniques idealthe examiner may consider the intent of ly suited for prying into the attitudes of the investigation quite obvious. people whom they regard as dangerous. Even if we decide that we do have The "error-choice" test is a good case PUBLIC OPINION QUARTERLY, SPRING 1951 138 in point. No doubt instruments using Mediation," and followed by a grossly the "error-choice" technique might be inaccurate statement of the findings. constructed which could be applied to This reporter undoubtedly felt that he eliminate allegedly "disloyal" citizens had given an accurate account of the from jobs of confidence and trust, to mediator study; his misinterpretations spot so-called "troublemakers" and "agi- may have been due to a lack of clarity tators" in industrial concerns, or, in ef- with which the assumptions were origfect, to discover "non-conformers" in inally expressed. many other important social areas. The The reliance which the public still writer hopes that this pessimism is not seems to place upon the published rewarranted, although he feels we should sults of attitude surveys may have a consider all possible ramifications. bearing on the question. As Edward L. Misinterpretation of Results. Finally Bernays has writen: there is a fourth question whose perti"There is a great danger in the nence is not limited necessarily to the new kind of leadership which polls use of indirect methods of attitude meashave produced in the United Statesurement: "What is my responsibility leadership of obedience to polls.... for seeing that the findings which I reThe people who pin their faith on the port are properly interpreted?" permanency of attitudes as shown by The danger of misinterpretation is polls, and therefore believe they are especially great in those investigations accurate forecasts, are often misled.... which utilize the various indirect methThe present belief that polls show a ods of attitude measurement. The genpermanent public opinion helps to eral public, unfamiliar with the backmaintain the status quo ... the danground and assumptions of these methger to society is self-evident."5 ods, is likely to read something into the results which may not even be implied This paper has, it is hoped, suggested in the investigator's formal report. An some of the considerations which should illustration of this sort of thing comes be looked into before the investigator again from the study of the personal decides on the use of indirect methods factor in labor mediation. A reporter of attitude measurement. As the applifor one of the large industrial trade pub- cation of these techniques becomes more lications learned, through personal con- widespread, some codes will have to be tact, about the "error-choice" test which established to provide a guide for the was devised to test the nmediator's handling of the ethical as well as the knowledge as well as his so-called "im- public relations problems to which I partiality." He asked permission to see have referred. At the moment, the writthe study, and to quote from it prior to er is not in the position to provide these its publication in one of the profes- codes; he will be satisfied if this presensional journals. The request was granted, tation is found useful as a lead for furbut he was warned to check any con- ther discussion and possible action. clusions or statements of which he 5 Bernays E. L., "Should Pollsters Be Limight not feel sure. No more was heard censed?", Journal of Opinion until his story appeared in print, with and AttitudcInternationat Research, Vol. 3, Spring I949, the headline "Heads-or-Tails Odds Beat P. 9.