THE JOURNAL OF COMMUNICATION F ol. 19, September 1969, p. 239-247 A Primacy Effect in Decision-Making by Jurors VERNON A. STONE Abstract An experiment varied the order of presentation of ostensible trial testimony. Prior to the reading of strong final evidence from the prosecution, 65 university students taking the role of jurors read other arguments which were identical for all recipients except that they came from either (a) the prosecution first and the defense second or (b) the defense first and the prosecution second. Two tentative verdicts and a final verdict were marked anonymously following the respective presentations. The sequence of tentative verdicts, but not the order of presentation of arguments, affected the final verdict, which tended to be the same as the first tentative verdict. Regardless of which side presented its case first, those jurors who committed themselves to a verdict of innocent following the first presentation tended to vote acquittal at the end. Conversely, those initially marking guilty tended to convict the defendant on the final vote. The results indicate that initial cornmitmcnts, even when tentative and private, tend to influence final commitments by decision-makers in roles such as that of the juror. One of the most significant forms of persuasive communication in terms of consequences is that which takes place in a jury trial. Justice and the fate of a defendant rest upon the responses of the jurors to the arguments, testimony, and evidence communicated to them on behalf of the prosecution and the defense. Jurors are asked to keep open minds and to refrain from reaching any decision until both sides have presented their cases. However, it has been argued that such an open-minded approach to decisionmaking may be more an ideal than a reality, and that trials regularly are biased in favor of the prosecution by reason of a procedural artifact, namely that the prosecution presents its case first. Possible biasing effects of this order of presentation have been of concern to communication researchers at least since 1925, when Lund wrote: Vernon A. Stone is an Associate Professor of Journalism at the University of Wisconsin, Madison, where he received his Ph.D. in 1966. He is currently investigating source-message orientation and attitude change under a grant from the National Science Foundation.
240 The journal of Communication, Vol. 19, September 1969 While the lawyer of the plaintiff is reviewing his case and making his appeal, the belief of the jurors is already in the process of formulation, and they are not to be dissuaded from their position by an equal amount of evidence or persuasive appeal on the part of the defendant s lawyer, according to the law of primacy.l Thus the so-called Law of Primacy in Persuasion, as formulated by Lund, holds that the side of an issue presented first will have greater effectiveness than the side presented subsequently. However, other research has often failed to support a general law of primacy. Numerous exceptions have been found; indeed, in some experiments, the most recent rather than the first communication has been found the more effective.2 Reviewing a number of primacy-recency studies, Hovland has concluded : For important social situations in which primacy effects have been conconsidered a danger, for example, legal trials... the issue is clearly defined as controversial, the partisanship of the communicator is established, and different communicators present the opposing sides. These factors should give rise to relatively little primacy effect. Our concern might then be concentrated on preventing premature commitment on the basis of the first presentation alone.3 The variable of commitment, as related to primacy effects, has been considered by several investigators. Lund suggests that the commitment deriving from the very act of responding to an opinion questionnaire may tend to freeze opinions and thereby contribute to a primacy effect. He reasons that the respondents, having committed themselves after the reading of the first discussion, will remember this rating and will tend to be influenced by their desire to be consistent when asked to make another rating after reading the opposed disc~ssion. ~ Hovland and his colleagues conclude from their studies that commitment is an important factor only when it is public. In one study, answering a F. H. Lund, The Psychology of Belief. IV. The Law of Primacy in Persuasion, Journal of Abnoml and Social Psychology 20: 191, 1925. 2 See Carl I. Hovland et al., The Order of Presentation in Persuasion. New Haven, Conn.: Yale University Press, 1957. It should be noted that not only is the prosecution favored by any primacy effect, but to some extent stands to gain from any recency effect, since the final summary generally comes from the prosecution. Carl I. Hovland, Summary and Implications, in Hovland et al., op. cit., p. 155. *Lurid, op. cit., p. 190.
Stone: Primacy Effect in Decision-Making 241 questionnaire anonymously failed to produce a primacy effe~t,~ and in another this relatively private commitment was found to be less conducive to primacy effects than was public commitment.6 On the other hand, Bennett found private commitment to be as effective as public commitment in yielding primacy effects. Hovlands rationale for holding that commitment must be public to produce any considerable primacy effect is based on an assumption that knowledge of the social approval granted to consistency prevents one from altering his views.8 A somewhat broader approach, on the other hand, would hold that the need for consistency within oneself may be strong enough for internal as well as external commitment to contribute to primacy effects. In the case of jurors, assuming that they do not discuss the issues among themselves during the trial, the social approval aspect should be irrelevant in any decision-making prior to their final group decision. As for the effects of even tentative internal, or private, commitment, the conflicting results reported by previous investigators do not enable clear prediction. The problems of order of presentation and commitment appear relevant to strategies of inducing resistance to persuasion by the opposing side, or as recent investigators have labelled the process, conferring immunization against persua~ion.~ Suppose, for example, that in a jury trial one side, hoping to utilize any recency effects, waited until the last minute to introduce its strongest evidence. Would the jurors be more resistant to persuasion by this 5Carl I. Hovland and Wallace Mandell, Is There a Law of Primacy in Persuasion? in Hovland et al., op. cit., p. 13-22. Carl I. Hovland, Enid H. Campbell and Timothy Brock, The Effects of Commitment on Opinion Change Following Communication, in Hovland et al., op. cit., p. 23-32. Edith B. Bennett, Discussion, Decision, Commitment, and Consensus in Group Decision, Human Relations 8: 251-73, 1955. Hovland, op. cit., p. 132. For descriptions of various approaches to immunization against persuasion, see: William J. McGuire, Inducing Resistance to Persuasion: Some Contemporary Approaches, in Leonard Berkowitz (Ed.), Advances in Experimental Social Psychology, Vol. 1, New York: Academic Press, 1964, p. 191-229; Percy H. Tannenbaum, Jacqueline R. Macaulay and Eleanor L. Norris, Principle of Congruity and Reduction of Persuasion. JOUTWZ of Personality and Social Psychology 3: 233-38, 1966.
242 The Journal of Communication, Val. 19, September 1969 evidence if the opposing side in the prior proceedings had presented its case first or last? Or does order of presentation matter? This is one of the questions to which the present study was addressed. The main problem investigated was whether a belief is more resistant to attack when it has been held from the start (Direct Belief) or when it has been arrived at by changing from an opposing belief ( Conversion). The belief chosen for investigation was that the defendant in a murder case was not guilty. The attack on this belief consisted of strong evidence presented by the prosecution at the end of the trial. Commitments (expressions of belief) prior to the attack were private and supposedly tentative. For purposes of the experiment, an ostensible case was devised in which the defense effectively refuted the prosecution case as presented up to the point of the final prosecution attack. This was intended to lead all or most subjects taking the role of jurors to the same belief (not guilty) immediately prior to the final damaging presentation by the prosecution. From various existing studies, support was available for the prediction of: (1) a primacy effect, (2) a recency effect, or (3) no effect of order. If first impressions hold the danger that Lund suggests, O then those jurors who initially received and accepted the prosecution s arguments should be more likely to revert to guilty verdicts following the final strong attack by the prosecution. On the other hand, a recency prediction is suggested by the conversion concepts of Aronson and Linder.ll The essential conversion notion is that A s positive attitude toward B will be stronger if A has converted from a negative attitude toward B than if he has felt positively toward B from the start. Having initially seen the defendant as guilty and then having been converted to a verdict of innocent, the jurors should be more difficult to sway back to a verdict of guilty than if they had perceived the defendant as innocent from the start. Finally, for reasons noted earlier, particularly because commitment was private rather lolund, op. cit. llelliot Aronson and Darwyn Linder, Gain and Loss of Esteem as Determinants of Interpersonal Attractiveness, Journal of Experimental Social Psychology 1: 156-71, 1965.
Stone: Primacy Effect in Decision-Making 243 than public, Hovland would predict no difference due to order of presentation or commitment.12 METHOD An experiment was conducted in which the subjects, 29 male and 36 female students in introductory psychology courses at the University of Wisconsin, were asked to take the role of jurors. They received credit points toward their course grades for participating in the study. All experimental materials, including instructions, the ostensible major testimony from a trial and questionnaires to assess verdicts at various stages, were contained in a single mimeographed booklet. Ss were instructed to read the booklet in sequence, concentrating on one page at a time and not looking ahead or back at other pages. Two differently arranged booklets representing the two experimental conditions were distributed in such a manner that alternate forms went to alternate Ss by random assignment. The experiment was represented as a Legal Science Foundation study of the process of decision-making by jurors. Instructions at the beginning of the booklet explained that the major testimony and evidence from a trial would be presented one part at a time and, at the end of each part, the Ss would be asked as jurors to give what would be their verdicts if the trial were ending at that point. The item used to assess the verdicts read: Given what you now know, if you had to make a decision at this moment, what would it be? ss placed a check on the appropriate side of a vertical line dividing GUILTY and INNOCENT. 13 Materials were identical except for their arrangement in two different orders of presentation intended to yield different sequences of tentative verdicts prior to a final strong attack by the prosecution. In the Conversion ( GUILTY-INNOCENT) condition, Ss: (1) read the regular prosecution case. (2) marked a l2 Hovland, op. cit., p. 132, 155. l3 Degrees of guilt or innocence were indicated on an 8-point scale, with a vertical line separating the 4-poht GUILTY and the 4-point INNOCENT sides. However, these data were collapsed to nominal form; it was reasoned that a l-point change from INNOCENT to GUILTY, for example, can hardly be considered the same degree of psychological change as a 1- point shift from one to another level of innocence.
244 The Journal of Communication, Vol. 19, September 1969 tentative verdict, (3) read the defense case refuting the prosecution and adding further testimony to support the defendant, (4) marked a second tentative verdict, (5) read the prosecution s final damaging attack, and (6) marked a final verdict. In the Direct Belief (INNOCENT, INNOCENT) condition, Ss: (1) read the defense testimony refuting the prosecution s regular case and then read that prosescution case, (2) marked a tentative verdict, (3) read further testimony supporting the defendant, (4) marked a second tentative verdict, (5) read the final prosecution attack, and (6) marked a final verdict. The further supportive testimony immediately preceding the second tentative verdict in both conditions was intended to facilitate the verdicts of innocent desired from all subjects at that point. A criterion in constructing the materials was that they lend themselves to the two different ordinal arrangements without seeming contrived in either condition. The ostensible trial proceedings were also designed to enable Ss to feel that they had actually made choices, i.e., testimony and evidence were intended to be convincing but not obvious. Furthermore, the technique of source discreditation was never used. No witness was ever made to appear unreliable and none ever contradicted another. The defense mainly put into context the apparently honest but circumstantial testimony offered on behalf of the prosecution. The ostensible case was that of a woman charged with murder in her husbands death by poisoning. The prosecution contended that she murdered him in jealousy because he wanted to divorce her and marry a younger woman with whom he had had a prolonged affair. The defense held that it was suicide by an emotionally disturbed man being rejected by the younger woman. The prosecution s final strong attack consisted of evidence and testimony that the defendant bought a book on toxicology two days before her husband died by p0is0ning.l~ In order to assure that the tentative and final verdicts (commitments) were made with a feeling of anonymity, the experimenter orally stressed to the Ss at the beginning of the experi- 14Complete sets of the materials used may be obtained from the author at the School of Journalism, 425 Henry Mall, University of Wisconsin, Madison 53706.
Stone: Primacy Effect in Decision-Making 245 ment that they would not be asked for their names or any other identifying information. RESULTS AND DISCUSSION For both orders of presentation, the materials were totally successful in leading the Ss to verdicts of innocent just before the strong final attack by the prosecution. Immediately prior to the final attack, all 65 Ss saw the defendant as innocent. The earlier tentative verdicts, those given after only the first part of the trial proceedings had been read, were overwhelmingly but not unanimously as intended. Seven of the 31 Ss in the Conversion condition found the defendant innocent after reading the prosecution s regular case as the initial part of the trial. Six of the 34 Ss in the Direct Belief condition cast ballots of guilty on their first questionnaire page instead of the verdicts of innocent intended at this point by the experimenter. Table 1 shows that the mere fact that one side was presented before the other made practically no difference in resistance to the final strong attack by the prosecution. Placing the prosecution s case before the defense in the prior proceedings yielded 16 final verdicts of guilty and 15 of innocent. Placing the defense first yielded 18 innocent and 16 guilty. In short, there was no difference due to the order of presentation per se. However, there were clearly significant differences due to the sequences of tentative verdicts or commitments given prior to the final strong persuasive message. Two thirds of the converts Table 1 Final Verdicts According to Side Making First Presentation and Sequence of Tentative Verdicts First Presentation Prosecution Defense Tentative Verdicts Guilty Final Verdicts Innocent Guilty, Innocent 15 9 Innocent, Innocent 1 6 Innocent, Innocent 10 18 Guilty, Innocent 6 0
246 The Journal of Communication, Vol. 19, September 1969 reverted to their initial position following the prosecution s final attack. Of the 30 Ss whose initial verdict had been guilty, 21 returned a final verdict of guilty. Conversely, for the 35 Ss whose initial verdict had been innocent, 24 chose the same final verdict.15 In other words, for two of every three Ss in the role of juror, the final verdict was the same as the first verdict. Thus there was a primacy effect for the side making the first presentation only when it was successful in bringing the desired initial commitment by the jurors. The results support Hovlands concern with premature commitment rather than with order of presentation per se as a possible biasing factor in a jury trial. However, they do not support the Hovland thesis that there need be concern only about public commitments.16 The commitments made by the jurors in this study were not only private, with full anonymity assured, but were made in a setting designed to facilitate their tentativeness. The experiment was presented as exactly what it was-a study of the process of decision-making by jurors-and for the first two verdicts, or commitments, the Ss were asked for tentative decisions, how they saw the defendant at that point in the trial. This would appear quite comparable to the internal commitments which actual jurors might make at various points in a trial, perhaps not always with full awareness that a commitment was being made. No support was given the notion that, having changed sides once, a person is less inclined to change back again than if he has held the same position from the start. It may be that concepts which would explain the resistance to change by certain converts are not applicable to a situation in which commitments are both private and tentative and where ego involvement or self-definition would be expected to play minimal roles. There was also no support for any notion that people strive for consistency to the point of avoiding switching back and forth in the internal process of individual decision-making. This is attested to by the 16Chi-square for the frequencies of final verdicts, divided according to initial verdicts, is 9.61 (p <.01). Chi-square for the full array of eight frequencies shown in Table 1 is 13.29 ( p <.01). 16Hovland, wpru note 12.
Stone: Primacy Effect in Decision-Making 247 fact that we were able to convert most subjects in the Conversion condition from guilty to innocent verdicts and then bring them back to guilty for a final verdict. Had the verdicts been public, the results might have been different. The desire to appear consistent to others might have reduced the rates both of conversion and of reversion, the latter especially. However, the use of public commitments would have been less representative of the jury situation, in which the deliberative process is supposed to be internal for the individual juror prior to the final group decision-making. Furthermore, we would expect jurors generally to approach their decision-making from what Bauer would call a problem-solving rather than a psychosocial 0rientati0n.l~ Their focus should be upon the informational --upon the weighing of facts in a situation and reaching a solution to a problem-rather than upon approval seeking or otherwise interacting with others in a social situation. Psychosocial variables which public commitment might bring into play, then, hardly appear relevant to the process of individual decisionmaking to which the present study was addressed. The finding of a primacy effect in the two approaches to inducing resistance to the final strong persuasive appeal by the prosecution is in accord with the finding by Tannenbaum et al. that protecting a belief by prior refutational immunization is more effective than trying to restore the belief after it has been successfully attacked. As those authors put it, this is akin to locking the barn door after the horse has been stolen.l* The evidence clearly indicates that, other things equal, first impressions tend to have a disproportionate effect on final impressions by decisionmakers in roles such as that of the juror. ACKNOWLEDGMENTS The helpful suggestions made by Dr. Percy H. Tannenbaum and Dr. Eleanor L. Norris are gratefully acknowledged. Raymond A. Bauer, Source Effect and Persuasibility: A New Look, in Donald F. Cox (Ed.), Risk Taking and Information Handling in Consumer Behauior, Boston: Graduate School of Business Administration, Harvard University, 1967, p. 559-78. l8 Tannenbaum, et al., op. n t., p. 237.