| Home | Audio Magazine | Stereo Review magazine | Good Sound | Troubleshooting |
|
Since the 1930's there has been an increasing concern with the basic psychological processes underlying attitudes and their modification. The new trend got under way with the work of Hovland and his colleagues in the Information and Education Division of the War Department during World War II. Their program included a great variety of studies; of special interest here are those which employed controlled variation. These experiments were among the first to show how specific content transmitted by specific communicators affects particular audiences. Let us begin by describing some of the work done in this program and then broaden our discussion by including later work which stems from the initial investigations and which bears on similar and new theoretical and empirical problems. From Arthur R. Cohen, "The Communication," Attitude Change and Social Influence (New York: Basic Books, Inc., 0 1964), 1-16. Reproduced with permission of the publisher. THE EFFECTS OF ONE-SIDED VERSUS TWO-SIDED COMMUNICATIONS What role does the organization of arguments play in the effectiveness of a persuasive appeal? In attempting to answer this question, writers have in the past dealt with methods of refutation, problems of emphasis, number of repetitions, and so forth. One of the central problems, however, is whether it is more effective to present only one side of an issue or to present both sides. Should a communicator concentrate only on the points supporting the position he advocates or should he also discuss opposing arguments? Which strategy is the more effective? In Experiments on Mass Communication, the volume in which they report the results of their wartime studies, Hovland, Lumsdaine, and Sheffield (1949) investigate whether, when the weight of evidence supports the main thesis of a communication, it is more effective to present only the materials supporting the point at issue or to introduce opposing arguments as well. These investigators presented to two experimental groups of 214 soldiers and to a control group of 197 soldiers communications on whether there would be an early end to the war with Japan after the surrender of Germany in 1945. All of the soldiers were tested some time before the communication on their beliefs about whether Japan would surrender. One experimental group was given a fifteen-minute talk that presented only the arguments supporting the idea that the war with Japan would be a long one; the talk included much factual material stressing Japan's strength. The other experimental group was given a communication which contained the same material plus an additional four minutes of in formation, woven into the presentation, which stressed the United States' advantages and Japan's weaknesses. The investigators' hypothesis was that those soldiers who were given only a one-sided argument would distrust a presentation that had failed to include op posing arguments and would be stimulated to rehearse their own position and to seek new ways of supporting it. After the presentation of the material, the soldiers were again measured on their beliefs about the probable length of the war with Japan, and a measure of change from before to after the presentation was computed. The effectiveness of the program was evaluated by comparing the average change in each of the experimental groups with the changes in a control group which had heard no communication but had merely been given the "before" and "after" attitude measures at the same time as the experimental groups. Both experimental programs were found to be extremely effective in producing change in the men's opinions, but neither program had any advantage over the other for the audience as a whole. Depending upon the initial position of the listener, however, the net effects were different for the two ways of presenting the material. The program giving both sides was more effective for those men initially opposed to the position advocated (those who expected a short war), whereas the program giving the one-sided picture was more effective for men initially favoring the stand taken (those who expected a long war). The investigators also expected that an obviously one-sided communication would be less effective with well-educated men and that these men would be more likely to take seriously arguments that seem to take all the factors into account. Less well-educated men, on the other hand, with undeveloped skill in critical thinking, might be more impressed with the strength of a one-sided argument, without thinking of objections. These expectations were borne out: The program which presented both sides was more effective with the better-educated men, and the program which presented one side was more effective with the less educated. When initial position and amount of education are considered together, the two-sided communication turned out to be more effective with better-educated men, no matter what their initial position, and the one-sided presentation most effective with those less-educated men already convinced of the position advocated. Thus, to decide the most effective type of presentation requires information about the educational level of the audience and the beliefs that audience al ready holds. Do these two types of communication reveal differences in resistance to counterinfluence? A further experiment (Lumsdaine and Janis, 1953) bears directly on this point. In this experiment, a week after having been exposed to either a one-sided or a two-sided communication on Russia's inability to produce atomic bombs for many years to come, half of the subjects in each experimental group were exposed to a counter-argument before being asked to state their opinions again. The counterpropaganda consisted of a playing-up and an elaboration of the arguments in the two-sided communication, as well as some new material. When the scores for change were examined, it was found that while there was no difference for those not exposed to the counterpropaganda, those who had been exposed to the counter propaganda and who had heard the two-sided program were more resistant to the counterpropaganda than those who had heard only the one-sided argument. In summarizing the results of these two experiments, Hovland, Janis, and Kelley conclude that a two-sided communication is more effective in the long run when, no matter what its initial opinion, the audience is exposed to subsequent counterpropaganda or when, regardless of subsequent exposure, the audience initially disagrees with the position advocated by the communicator. The two-sided communication is less effective if the audience agrees with the initial position and is not exposed to later counterpropaganda. With respect to the second experiment, the two-sided argument apparently prepares the listener to meet counterarguments; it would thus seem especially effective in "inoculating" the listener against subsequent counterinfluence. "Inoculation" relates also to the problem of building up resistance to pressures toward attitude change . . . . The concept of inoculation also raises questions about the way in which two-sided communications present opposing arguments. Although the investigators are most interested in resolving the issue of which type of communication is the more effective, their results make it clear that further work should take ac count of the manner in which opposing arguments are introduced, the character of the arguments, and the extent to which they are explicitly refuted. . . . The foregoing conclusions about the relative effectiveness of one-sided and two-sided communications have been extended by other investigators (Thistle thwaite and Kamenetsky, 1955). Attitude change tends to be greater for those subjects whose comprehension of the communicator's conclusion is greater or who show fewer and less intense discounting reactions to the communication. Introducing facts in support of the "other side" leads to less change of attitude when the facts are unfamiliar to the subjects, but failure to include well-known facts on the "other side" also weakens the appeal. Thus, to the degree that the facts included in the different communications are comprehensible and familiar, the appeals will be differentially effective in producing changes in attitude. The experiments cited take little account of the fact that the persons subjected to attempts at influence are not merely members of an audience but also persons with active social lives within their social groups. Festinger (1955) has commented on the fact that knowledge about the effect of interpersonal and group processes implies that such processes may be modifying the effects obtained by one-sided versus two-sided arguments. It is possible that the two-sided presentation, by stressing the controversial aspect of the problem, may set off considerable discussion among the listeners in the week-long interval after hearing it. Thus, rather than showing the effectiveness of a one-sided versus a two sided presentation, the results may show the resistance to counterinfluence under one set of conditions, where opinions are not anchored in a membership group, as compared to the resistance under conditions where they are firmly anchored in a membership group. . . . THE EFFECTS OF STATING A CONCLUSION Whether a conclusion should be stated conspicuously in a persuasive communication is a question which has long been argued by propagandists, educators, and public speakers. Is it more effective to let an audience draw its own conclusion, or is it better to make the conclusion explicit? If we assume that indirect suggestion is more effective or that decisions are more effective when reached independently, we might expect the presentation of an implicit conclusion to be the more powerful. But some persons, especially the less intelligent, will miss the conclusion entirely if it is not stated explicitly. Hovland and Mandell (1952) designed an experiment for studying this problem systematically. They compared two types of communication which were identical in every respect except one: in the first, the communicator drew the conclusion at the end, while in the second the conclusion was left to the audience. Under the first set of conditions, the general principles of the relevant topic (cur rent economic issues) were presented, together with a statement of the existing situation (the bad financial status of the U.S.). From the principles and the statement of the situation, a logical conclusion could be drawn by the individual listener (desirability of devaluation of American currency). Under the second set of conditions, the implication and conclusion were stated explicitly b y the communicator. The results showed that more than twice as many listeners changed their opinions in the direction advocated by the communicator when the conclusion was explicitly drawn as did when it was left to the audience. While these results favor stating the conclusion explicitly, it does not follow that this strategy will invariably be superior. A number of factors are involved: the kind of communicator (whether he is trustworthy, an expert, or seen as biased), the kind of audience (their intelligence, sophistication, types of personality), the kind of issue (whether or not it is "ego-involving," how complex it is), and the degree of explicitness with which the conclusion is actually drawn. . . . Here we may note that later research (Thistlethwaite, de Haan, and Kamenetsky, 1955) explores the role of intelligence in the effectiveness of stating conclusions. Stating the conclusions in the message proves to be more effective in changing the opinions of the less intelligent members of the audience than it does those of the more intelligent. In general, however, later research has been unable to confirm these earlier findings. One set of investigators used as an issue the wisdom of the United States in following a policy of limited war in Korea. They obtained positive results, but other investigators suggest that these results may have been due to the fact that the measures of attitude emphasized comprehension of what the speaker said rather than personal attitudes about the issue. These critics base this interpretation on their finding that under conditions where the conclusions were stated, the subjects showed an increase in comprehension of the material over those where the conclusions were left implicit, but the two groups showed no difference in agreement with the position advocated. An alternative explanation for these conflicting results (suggested by Krech, Crutchfield, and Ballachey, 1962) is that the complexity of an issue is a central factor in the effectiveness of stating a conclusion (Hovland et a/., 1953). The issue of limited war in Korea was one which had been widely discussed in the mass media, so that the arguments pro and con had already been spelled out for the subjects in understand able terms. By contrast, the issue of currency devaluation is a complex economic problem which most people do not understand very well. In summary, then, we may state that persuasive communications which pre sent a complicated and unfamiliar series of arguments on impersonal topics to less intelligent people are more effective when the conclusion is stated explicitly than when the audience is left to draw its own conclusion. There are, however, many problems that remain to be investigated before we understand fully the conditions under which explicit presentations are more effective than implicit ones in producing attitude change. THE EFFECTS OF ORDER OF PRESENTATION The sequence of presentation of the arguments is another important aspect of a persuasive communication. Should one start with his strongest arguments or save them until the end? Different orders of presentation in persuasive communications are discussed by Hovland, Janis, and Kelley in Communication and Persuasion (1953). They divide the problem into two parts: (1) When only a single side of an issue is presented, is it more effective to utilize the strongest arguments at the outset or at the end? (2) When both sides of an issue are presented successively, does the side presented first or the side presented last have the ad vantage? After reviewing the evidence as to the difference between a "climax" order (important arguments reserved until the end) and an "anticlimax" order (major arguments presented at the beginning and weaker ones at the end), the investigators conclude that it is unlikely that one or the other order of presentation will invariably turn out to be superior. Rather, they feel that different external factors will produce different outcomes, and they suggest what sorts of factor will affect the outcome. Most important are attention, learning, and acceptance. They say further that the presentation of major arguments at the outset (anticlimax order) will be most effective when the audience is initially little interested in the communication, for it helps to catch attention. But when attention and motivation to learn are present, they hypothesize that the climax order will be more effective in gaining acceptance because the anticlimax order fails to fulfill the expectations created by the initial portions of the communication and may produce a letdown that promotes forgetting. Thus the advantages of one order over the other depend on the particular conditions under which the communication is presented, including the predispositions of the audience and the type of material being communicated. Which order is the more effective when both sides of an issue are presented? This second problem of sequence is known in psychology as the "primacy recency" question. It was first stated in its most general form as "Is there a Law of Primacy in persuasion?" An early study ( Lund, 1925) made on college students, in which an instructor communicated two sides of an issue, argued for such a law. This "law" stated that the side of an issue presented first will be more persuasive than the side presented subsequently. In considering the evidence from later experiments, Hovland, Janis, and Kelley (1953) conclude that neither primacy nor recency produces consistent effects. They argue for research on the factors that contribute to the differences in results that have been obtained in the various experiments. These factors may include the complex and interacting roles of learning, attention, and acceptance, and the role of such special factors as the position of a teacher as communicator to his own students in a class room. Following up this line of attack, a later volume by Hovland and his colleagues (The Order of Presentation in Persuasion, 1957) presents replications of earlier experiments and a number of new experiments on the effects of order. In stead of attempting to confirm a general law of primacy in persuasion, they ask what conditions make either primacy or recency effective in persuasion. One set of studies tests the generality and validity of the Law of Primacy; a second deals with the effects of different sequences of appeals and arguments within a single communication. ORDER OF SUCCESSIVE COMMUNICATIONS The series of experiments on the effects of primacy and recency have led to the following generalizations (Hovland, 1957, pp. 130-138). 1. The Law of Primacy is not general. When two sides of an issue are presented successively by different communicators, the side presented first does not necessarily have the advantage. In a study by Hovland and Mandell, successive communications advocated first one and then the other side of an issue. The order of presentation was counterbalanced-half of the subjects received the "pro" arguments first and the other half the "anti" arguments first. The subjects were given opinion questionnaires after the first side had been presented and then again after the second side. The results failed to replicate the findings by Lund (1925): although some groups showed primacy effects, most groups showed recency effects. The authors speculate whether the differences between Lund's results and theirs may be due to differences in learning and acceptance. The motivation to learn the first communication may have been stronger in Lund's experimental group, and the subjects' commitment to the statement of their first opinion may have reinforced their acceptance of the first communication. 2. Public commitment is a significant factor in the effects of primacy and recency. If, after hearing only one side of a controversial issue, a listener makes a response which publicly indicates his position on the issue, the result is a primacy effect. Without telling the subjects that they would later hear the other side of a controversial issue, the investigators presented one side and asked half of the subjects to write for publication their opinions on the issue, while the other half were asked to write their opinions anonymously. Then the other side of the issue was presented to both groups and their opinions recorded again. The investigators found that the public expression of opinion tended to fix opinions on the first side and to make the presentation of the second side less effective in changing attitudes-a primacy effect. Where there is no public commitment (as in the Hovland and Mandell study), the mere statement of one's opinion anonymously on a questionnaire after hearing only one side of an issue does not re duce the effectiveness of the second side. The experimenters believe that the effect of public commitment is due to social rewards and the need for social approval. Having placed his opinions on record, the subject feels that he cannot alter his views if he is to be regarded as consistent and honest by those with whom he expects to interact. 3. Primacy effects may occur when one communicator presents contradictory information in a single communication, but this effect can be reduced by interpolating other activities between the two blocks of information and by warning the subjects against "the fallibility of first impressions." In a related series of experiments, Luchins prepared two blocks of information describing the personality characteristics of a person not known to the subjects. One block contained information about an introverted person, the other block contained items characteristic of an extroverted person; for some subjects the order was introversion extroversion and for the other half the reverse. Subjects were then asked to select adjectives reflecting their impressions, to write brief descriptions of personality, and to make predictions about the later behavior of the person they had read about. The material presented first proved to be considerably more influential in determining what the subjects thought to be the chief characteristics of the person described. When subjects in a comparable experiment were forewarned about the possible fallibility of first impressions before any information was presented, however, the second block of information tended to exert a relatively greater influence than the first on the final impressions formed. Recency was most effective in the group that was forewarned, next most effective in another group in which the warning was interpolated between the first and second blocks of in formation, and weakest in a third group for whom arithmetic tasks were interpolated between the two blocks of information. In all three groups, however, the recency effect was stronger than primacy effect. ORDER WITHIN A COMMUNICATION What is the best way to organize the material when a complex argument for one side of an issue is to be presented? For example, should you begin with those arguments which favor your side of the issue and then refute the opposing arguments, or should you take care of the opposition first? 1. // a communicator first arouses the subject's needs and then presents in formation that tends to satisfy those needs, the information will be accepted more readily than if the arousal of need follows the presentation of the information. An experiment by Cohen compared two situations: one, presenting to college students a threatening communication about problems of grading and necessary re forms, was followed by the information that grading "on the curve" would solve the current situation; the other reversed the order of the presentation of the threat and the information about grading on the curve. Attitudes toward grading on the curve were measured before and immediately after the experimental introduction of the threat and the information, and then again three months later. It was found that the first sequence (need-information) was more effective, probably because the second order (information-need) did not make much sense to the subjects. Information presented after a need is aroused can operate in a direct fashion; less effort is required on the part of the listener to see it as relevant to his needs. Those who receive the information first do not see the point of the information as they are getting it. They see its relevance only afterward, and by then they may already have lost much of it because they have not been paying close attention; in any case, they now have the job of reconstructing what they have heard in a manner relevant to their needs. 2. Attitudes change more when communications highly desirable to the subject are presented first, followed by the less desirable ones, than when the less desirable ones come first. McGuire found that the communicator elicited more agreement with his views when his earlier communications were rewarding for the subject. McGuire's hypothesis is that, with this sequence, the subject starts out by becoming progressively more responsive to the communicator in terms of paying attention and being willing to learn; later the communicator capitalizes on these responses in the less rewarding part of the message. On the other hand, the communicator who first presents undesirable content to the recipient excites responses leading to nonacceptance (withdrawing attention) because agreeing with these undesirable issues is unpleasant; by the time he gets to the rewarding part of his message he has lost the subject's attention. 3. The "pro-con" order is superior to the "con-pro" order when an authoritative communicator plans to mention "pro" arguments and also nonsalient "con" arguments. This generalization comes from an experiment by Janis and Feierabend. One experimental group received a version of a pamphlet on civil defense in which the arguments for civil defense were presented first, followed by the arguments against it; the other group received a version which reversed the order. To gauge the relative effectiveness of the two forms of the communication, the subjects were given a post-communication questionnaire on their willingness to volunteer for civil defense. The data showed that the group which received the "pro" arguments first found the communication more persuasive than did those who received the "con" attitudes first: they were more ready to volunteer for civil defense. Janis and Feierabend interpret their results as an example of the resolution of an approach-avoidance conflict: arguments in favor of a position agreed with initially strengthen the approach tendency, and arguments against the position, provided they are not too strong, can then be handled without causing a reversal of the initially favorable attitude. More recent research by Miller and Campbell (1959) and Anderson and Barrios (1961) is also relevant to the problem of order effects. Miller and Camp bell call attention to the fact that the rate of forgetting diminishes over time; we forget most rapidly immediately after learning, and in successive equal time intervals we forget proportionately less and less of what is left. Thus, of two associations equally strong at a given moment, the older will decay less rapidly. From the properties of the curve of forgetting, certain predictions can be made and tested. Miller and Campbell varied the order in which they presented opposing arguments (pro-con, con-pro) to college students. The communications were on an issue about which the subjects had little prior information and had formed no opinions (a trial involving a suit for damages), so that there was little contamination from previous learning, a condition not generally met in the earlier two sided studies. In addition to varying the order, the investigators varied both the time interval between the opposing arguments (none and one week) and the time of testing the effect of the arguments (immediately after the second argument and one week later). The measure of attitude toward the defendant's case (the "con" side) and toward the plaintiff's (the "pro" side) also showed a significant recency effect, no matter which information was presented first, under the conditions most favorable to the emergence of a recency effect as predicted from the curve of forgetting. The last-heard argument is most effective when there is a long delay between the first and second communications (with maximal loss of memory for the first) coupled with an immediate measurement after the second communication (with minimal loss of memory for the second). The last-heard argument is least effective when the two presentations are contiguous (so that the temporal advantage for the second is minimal) and the testing is delayed (so that the relative lapses of time since the initial exposures become more and more nearly equal). On the basis of the learning curve, however, one would not expect that the first-heard argument would have the advantage even under the latter conditions, which are those least unfavorable to it. Nevertheless, under these conditions the first-heard argument turns out to be the stronger one. In other words, there is a primacy effect when it is given a chance to show itself. Miller and Campbell point out that most of the earlier experimental studies involved conditions of contiguous presentation and immediate testing. Immediate testing is a condition in which the relative time lapse is at its greatest for the first-heard argument, so that any primacy effect there may be is canceled out by the time-advantage given to the second argument. The reader familiar with the relative effects of what are known in the general psychology of memory and for getting as retroactive and proactive inhibition will realize that from the point of view of these effects the first-heard argument is probably under the greatest disadvantage when the presentations are contiguous. It is not surprising, therefore, that different experiments have shown inconsistent results (Hovland, 1957). Miller and Campbell (1959) conclude that the evidence cited by Hovland and his colleagues could be interpreted as consistent with an ever-present primacy effect, though the strength of this effect would necessarily vary with the experimental conditions peculiar to each of the studies and would frequently be masked by a recency effect. Anderson and Barrios (1961) raise an interesting issue: what happens when a subject receives a series of communications on a variety of topics from the same source, each communication being divisible into "pro" and "con" segments? Although the primacy effects are striking at first, they diminish as the series of communications progresses. The investigators point out that there are several possible explanations for this result. For one thing, the particular pat tern of good and bad adjectives used in the later sets may have stimulated an increased tendency to take account of all the words in each set. Also, a progressive loss of interest in the task might decrease the primacy effect. Finally, the subjects may simply become more skillful in handling the total communication as a single integrated unit, so that the primacy-recency distinction is less relevant. In any case, it is clear that it is not safe to generalize to situations involving many communications from the results of studies which employ only one or two communications. Taken as a whole, the findings regarding primacy and recency (see also Lana, 1961) seem to rule out any universal principle of primacy in persuasion, but they have led to specifications of some of the sets of conditions which affect primacy. These factors include time of measurement, similarity of issues, contiguity of presentation, number of separate issues, earlier positive experiences with the communicator, interpolated activity, warnings against premature commitment, encouragement toward commitment, ambiguity inherent in the sequence of communications, and arousal of needs before proffering information. The effects of these factors are not due in any simple manner to learning or memory of evidence or arguments or to such factors as set, reinforcement, or attention. If anything, factors of acceptance may be the most critical. Thus, coming first makes a statement no more likely to be remembered, but does make it more likely to be believed; one side of an argument tends to be persuasive provided we have not heard the other side, and hearing one side after we have heard the other makes us more critical and skeptical. As is true for the entire area of persuasive communication and attitude change, future study of the order of presentation should be based on the development of more elaborate theoretical models which take into account the laws of learning, perception, and motivation and on the conducting of crucial experiments which pit one theoretical approach against another. By invoking different theoretical formulations, we uncover sets of considerations which might never emerge as factors affecting the order of presentation if we relied on only one theoretical system. REFERENCES ANDERSON, N. H., and A. A. BARRIOS (1961). Primacy effects in personality impression formation. Journal of Abnormal and Social Psychology, 63, 346-350. FESTINGER, LEON (1955). Social psychology and group processes. In C. P. Stone and Quinn McNemar, eds., Annual Review of Psychology, Vol. 6. Stanford, Calif.: Annual Re views. Pp. 187-216. HOVLAND, C. I ., ed. (1957). The Order of Presentation in Persuasion. New Haven, Conn.: Yale University Press. HOVLAND, C. I ., I . L JANIS, and H. H. KELLEY (1953). Communication and Persuasion. New Haven, Conn.: Yale University Press. HOVLAND, C. I., A. A. LUMSDAINE, and F. D. SHEFFIELD (1949). Experiments on Mass Communication. Princeton, NJ.: Princeton University Press. HOVLAND, C. L, and WALLACE MANDELL (1952). An experimental comparison of conclusion drawing by the communicator and by the audience. Journal of Abnormal and Social Psychology, 47, 581-588. KRECH, DAVID, R. S . CRUTCHFIELD, and E. L. BALLACHEY (1962). Individual in Society. New York: McCraw-Hill Book Company. LANA, R. E. (1961). Familiarity and the order of presentation of persuasive communications. Journal of Abnormal and Social Psychology, 62, 573-577. LUMSDAINE, A. A., and I . L . J ANIS (1953). Resistance to "counter-propaganda" produced by one-sided and two-sided "propaganda" presentations. Public Opinion Quarterly, 17, 311-318. LuND, F. H. (1925). The psychology of belief. IV. The law of primacy in persuasion. Journal of Abnormal and Social Psychology, 20, 183-191. MILLER, NORMAN, and D. T. CAMPBELL (1959). Recency and primacy in persuasion as a function of the timing of speeches and measurements. Journal of Abnormal and Social Psychology, 59, 1-9. SHERIF, MUZAFER, and C. I. HOVLAND (1961). Social Judgment. New Haven, Conn.: Yale University Press. THISTLETHWAITE, D. L ., HENRY DE HAAN, and JOSEPH KAMENETSKY (1955). The effects of "directive" and "nondirective" communication procedures on attitudes. Journal of Abnormal and Social Psychology, 51, 107-113. THISTLETHWAITE, D. L, and JOSEPH KAMENETSKY (1955). Attitude change through refutation and elaboration of audience counterarguments. Journal of Abnormal and Social Psychology, 51, 3-12. Also in Part 4:
|
Prev. | Next Top of Page | Index | Home |