1 E.g., Miss Wyoming's case against Penthouse ($26,500,000 verdict, reduced on posttrial motion to $14,000,000), N.Y. Times, Mar. 28, 1981, at 6; Carol Burnett's case against the National Enquirer ($1,600,000 verdict, reduced to $800,000), Jury's Judgment for Carol Burnett Cut in Half by Judge, to $800,000, N.Y. Times, May 13, 1981, at 7, col. 5; Guccione v. Hustler Magazine, Inc. ($39,600,000 verdict reduced to $4,000,000), N.Y. Times, Apr. 18, 1980, § II, at 5; and a suit by two police inspectors and a former assistant district attorney against the San Francisco Examiner and two reporters ($4,560,000 verdict entered), $4.6 Million Awarded to Officials for Libel by a Hearst Newspaper, N.Y. Times, Apr. 20, 1979, at A16, col. 1. All are at varying stages of posttrial procedure or appeal as of May 1981.Google Scholar

2 See note 21 infra.Google Scholar

3 Franklin, Marc A., Winners and Losers and Why: A Study of Defamation Litigation, 1980 A.B.F. Res. J. 455 (hereinafter study I).Google Scholar

4 The only media cases used in study I that are excluded here are those that were decided in 1976—before the starting date of this follow-up study.Google Scholar

5 Wolston v. Reader's Digest Ass'n, Inc., 443 U.S. 157 (1979).Google Scholar

6 Hutchinson v. Proxmire, 443 U.S. Ill (1979).Google Scholar

7 In Hutchinson, the Court was “constrained to express some doubt” about the idea that summary judgment was the appropriate stage for deciding defamation cases. 443 U.S. at 120 n.9. See also Wolston, 443 U.S. at 161 n.3. Elsewhere the Hutchinson Court suggested limitations on the category of “public official.” 443 U.S. at 119 n.8. Both cases used a narrow approach to the question of who was a “public figure.” Later in this study, we consider the influence of these cases. See text at pp. 821–23 infra. A third footnote in Hutchinson, observing that the Court had never considered whether constitutional protection in defamation cases extended to defendants who were not related to media, 443 U.S. at 134 n.16, is beyond the scope of this study. Nomenclature is something of a problem here. Hutchinson is not included in the study because it was not brought against a media defendant. Yet the major discussions appear in Hutchinson rather than in Wolston, which is in the study. Nonetheless, throughout the tables and sometimes in the text, the critical pair of cases is referred to as before and after Wolston alone. We intend Wolston to refer to both cases. In “pre/post” tables, Wolston itself is treated as a “pre-Wolston” case.Google Scholar

8 We excluded from our definition of “media” those publications intended only for the members of a defined group, such as a union newsletter. We did, however, include specialized publications that were available to the general public on newsstands, such as high fidelity magazines. No more than five or six cases were excluded as a result of the drawing of this line. We included cases in which the defendants asserted privileges not to reveal confidential sources but excluded those in which a nonmedia defendant was being sued and a reporter-third party was being asked to reveal a source or information. Again, this exclusion affected only a handful of cases. One case involving presuit discovery of a potential defendant was excluded as was one involving solely the question of staying discovery pending the outcome of a related matter.Google Scholar

9 Twelve were 1977 cases. The others were spread through 1978 and early 1979. Most of these cases involved jurisdiction or venue questions and did not raise substantive law questions. Of the 25 cases, 4 were at the trial level.Google Scholar

10 Although the cases cover a period of almost four years, we generally treat them as providing a snapshot rather than as exemplifying trends within the period. This is due primarily to the relatively small numbers of cases in some years, and the number of trial-level cases, which offer limited opportunity for analysis. See note 11 infra. The only possible changes being analyzed within the period are those that might be due to the Hutchinson and Wolston decisions. See text at pp. 821–23 infra. But see note 17 infra (success rates by year).Google Scholar

11 As noted in study I, supra note 3, at 464, trial judges are more likely to write reportable opinions when making decisions that finally conclude cases, such as when granting motions to dismiss or granting summary judgment rather than when denying such motions. We think it dangerous to use these cases for any conclusions concerning ultimate success or failure. Also, 31 of the 53 state trial court opinions are from New York courts. Overall, New York state and federal trial courts account for 37 of the 101 trial court opinions. The explanation for this dominance is probably that the media bar is centered there and more likely to be aware of opinions at the trial level and to send them to the Media Law Reporter. Moreover, the presence of the New York Law Journal, with its daily reporting of vast numbers of trial court opinions, brings many more opinions to light than are generally known in other states. No other state accounts for more than 4 state court trial opinions or than 8 trial opinions overall. The dominance of New York cases would not be cause for great concern if we believed that they were typical of state court opinions around the country. But we know they are not. First, in New York denials of motions to dismiss and for summary judgment are appealable. Perhaps judges who deny these motions and expect immediate appeals write at least short opinions to indicate the bases for their decisions. In other states, where appeals from such orders are rarely reviewable directly, the judge may decide to withhold any explanation of interlocutory orders until a later stage in the litigation. Further, New York is among the very few states that have imposed a liability test more stringent than that required by Gertz. Gertz v. Robert Welch, Inc., 418 U.S. 323 (1974). Thus, even the law being applied is atypical. See note 54 infra. Whatever the reasons, trial court cases look suspiciously different from appellate cases in other ways. For example, only 5 of the 101 trial court opinions involve posttrial motions compared to 35 of the 190 appeals (18 percent). Also, defendants succeeded in 80 percent (81 of 101 cases) of the trial court cases, compared to 66 percent on appeal. Finally, only 1 of the 101 trial court plaintiffs succeeded at trial—a judgment for $1 not appealed. See note 19 infra. The appellate figure was 10 of 190. This comports with our understanding that defendants are unlikely to accept defeat at the trial level. Plaintiffs, on the other hand, apparently do not persevere to the same extent. Thus, final success figures using trial cases would necessarily be badly skewed. On the other hand, we feel secure in using the trial-level cases when discussing such questions as who brings suit against what media for what kinds of charges. On these, we have not found (and see no reason to expect) any special skewing that would concern us.Google Scholar

12 A preliminary study of one insurer's data indicates that of 118 cases closed in 1979, 25 were dropped by the plaintiff, 30 were settled for some payment, and 55 were litigated. Of those litigated, 20 resulted in summary judgment for the defendant; 22 defendants won on motions to dismiss; 5 won directed verdicts; and 5 won jury verdicts. Three judgments for plaintiff followed jury verdicts. Two were not appealed. The settlements ranged from $300 to $50,000. The median payment was $4,500. This information is being developed for publication by Professor Jerome Skolnick, School of Law, University of California at Berkeley, and Elliot Cahn, manager of the Litigation Support Program at SRI International, Menlo Park, California. One lawyer reported that over the past 5 years the media clients of his office closed 16 cases. Seven were dropped by the plaintiffs; 4 were won on summary judgment and 2 on motions to dismiss; and 3 were settled for a total of $31,000. The lawyer reported that in the preceding 13 years he could recall only one payment that had been made. No cases in this office went to trial during the 5-year period. A second lawyer reported that in 1980 two clients had paid plaintiffs a total of $6,000.Google Scholar

13 The experience of one newspaper with a daily circulation of between 100,000 and 300,000 during this period may be suggestive. It was involved in 14 litigated matters during the years 1976–80. Its liability insurance during this period ran about $3,000 per year for coverage of $5,000,000 with a per-case deductible that rose from $10,000 at the start of the period to $20,000 by the end of the period. The newspaper's legal fees were of two types: (1) litigation fees for work on actual litigation and (2) amounts billed for counseling in the newsroom, including such items as prepublication reviews and handling demands for corrections or retractions. The fees do not include costs of preparing a libel handbook for the newspaper or of holding seminars for the staff. The early counseling figures are estimates. The following table shows the figures:. Litigation Counseling Total. A second lawyer reports that his newspaper client spent $240,000 on legal fees during 1980. The fees could not be divided as in the preceding paragraph. Twelve libel cases were pending against this client during 1980. One payment of $1,500 was made. The client paid $32,500 for libel coverage of up to $20,000,000 per incident, with a deductible of $25,000 per case. Both media clients may have had some of their legal fees reimbursed if the total of fees and other expenses exceeded the per-case deductible amount.Google Scholar

14 The special problem of New York is not so great with appeals because opinions in appeals in all states are generally available and are likely to be reported. Only 25 (13 percent) of the 190 appeals were from New York state courts.Google Scholar

15 Recall that this is close to the 80 percent success rate observed in the cases for which no appeal was found. Note 11 supra.Google Scholar

16 The implications of the differences in results before and after Wolston are discussed in the text at pp. 821–23 infra. The posttrial results are presented in greater detail in tables 9 and 10 infra.Google Scholar

17 Because the success rate of 66 percent was somewhat higher than the 60 percent rate found in study I, we attempted to determine whether a trend might be developing. The annual rates of defense success are as follows: 1976, 55; 1977, 61; 1978, 77; 1979, 53; 1980, 70. (The 1976 rate is from study I, supra note 3, at 471 n.36.) The 1979 rates are made up of a pre- Wolston half with an inexplicably low rate of 43 percent and a post-Wolston rate of 64 percent. The low rate for early 1979 was also found in study I. Id. No trend is apparent.Google Scholar

18 The “other” cases involved interlocutory appeals that were unlikely to yield any final outcome no matter who prevailed on the appeal.Google Scholar

19 In the sixth appeal from a judge's verdict, Sierra Life Ins. Co. v. Magic Valley Newspapers, Inc., 101 Ida. 795, 623 P.2d 103 (1980), the defendant newspaper refused to reveal a confidential source. The trial judge declared the newspaper in contempt and default, struck all defenses, and assessed damages of $1,942,680. On appeal, the judgment was vacated and the case remanded. In the tables in the text the case is treated as a verdict after trial but is identified where relevant. The discrepancy between jury and judge verdicts remains if one considers trial court cases. The defendant won one unappealed jury trial. One appellate case in the study involved a jury trial that plaintiff won. On appeal, defendant accepted liability for compensatory damages but argued, successfully, that punitive damages were improperly awarded. Cape Publications, Inc. v. Teri's Health Studio, Inc., 385 So. 2d 188 (Fla. Dist. Ct. App. 1980). Among the judge-tried cases, one verdict for $1 was not appealed. Handelman v. Hustler Magazine, Inc., 469 F. Supp. 1053 (S.D.N.Y. 1979). Two judge trials resulting in defense judgments have not been found on appeal. Adding in these trial court figures, judges found for plaintiffs in only three of eight cases.Google Scholar

20 The difference is significant at the .10 level. The phrase “statistically significant” as used throughout this study is not being used to assess the likelihood that future samples drawn from a universe will be similar to the sample already drawn. In this study we have the entire universe of cases before us, and we are not concerned about drawing further samples. In this study we are analyzing two-by-two contingency tables with fixed row and column totals (“marginals”). Our question is whether the distributions of numbers in the four cells are what they are because of chance (the null hypothesis) or whether the factor according to which they have been divided makes it appear more likely that something in the process explains the difference. Specifically, we were led to conclude that the difference in table 7 was significant by the following process. First we arranged the contested appellate cases in the following two-by-two box:Google Scholar

The null hypothesis was that the identity of the trier of fact was irrelevant to the trial results in these cases. Stated another way, the null hypothesis was that the two proportions (among jury-tried cases and judge-tried cases) were both estimates of the same quantity. We used a x² test with one degree of freedom and a correction factor that reduced the larger of two numbers (20) by .5 and increased the smaller one (2) by the same amount. Although such a correction factor need be used only when the total number of cases is below 20, we have used the factor in all calculations of statistical significance in this study. E.g., George W. Snedecor & William G. Cochran, Statistical Methods 217 (6th ed. Ames: Iowa State University Press, 1967). We might also note that the total of 29 cases in this table is one of the smallest in the study. In most, the total number of cases involved is well over 100.Google Scholar

21 The five awards that were upheld were: (1) jury, $20,000 general and punitive; (2) jury, $50,000 general; (3) jury, $60,000 general; (4) judge, $10,000 general; (5) judge, $350,000 general. The five altered verdicts all occurred in jury trials: (6) $75,000, reduced to $45,000 by trial judge because a third party had already paid $30,000; (7) $88,000, reduced by the trial judge to $75,000 general and punitive; (8) $132,500, reduced on appeal by eliminating the $100,000 punitive award; (9) $17,550, reduced on appeal by eliminating the $2,550 punitive award; (10) $36,000 in general and punitive damages for four claims. One was reversed on appeal and the judgment was reduced to $26,000. Note that three plaintiffs clearly obtained punitive damages. Six clearly did not. The result in case 6 is uncertain.Google Scholar

22 See note 21 supra, cases 6 and 7.Google Scholar

23 Alioto v. Cowles Communications, Inc., 623 F.2d 616 (9th Cir. 1980), cert, denied, 449 U.S. 1102(1981).Google Scholar

24 Study I, supra note 3, at 473.Google Scholar

25 See note 1 supra.Google Scholar

26 This percentage was obtained by dividing the number of affirmances of rulings on motions to dismiss and for summary judgment (103) by the total number of those cases (144) as shown in table 5.Google Scholar

27 This should be no surprise since the critical events, if not the suits, all occurred before Wolston.Google Scholar

28 In study I, we had two large groups of persons in business: (1) owners and managers and (2) lower-level employees. As that study showed, the owners and managers tended to sue media much more frequently than did the lower-level employees. That pattern has been repeated here. We found only three lower-level business employees suing media defendants and we included them in the category “other.” Throughout this study, therefore, we use “manager” or “owner or manager” to indicate business or commercial plaintiffs.Google Scholar

29 Generally the “public official” category is composed of elected officials not running for office (16) and incumbents involved in election campaigns (4). Nonincumbents running for office are listed separately.Google Scholar

30 The 44 “other” included the 3 lower-level employees mentioned in note 28 supra, 2 retired persons, 4 students or children, 2 prison inmates, 5 business groups, 4 social groups, 10 persons who could not be coded more specifically than as “alleged criminal,” and 14 “others.” Plaintiffs in 43 cases were not codable.Google Scholar

31 See study I, supra note 3, at 481 table 12 col. 2.Google Scholar

32 These figures are significant at the .05 level. See note 20 supra for a description of the statistical process we used. In table 12 we derived our two-by-two contingency table by isolating the elected public officials and treating all other groups as “other” in the following manner:Google Scholar

Again, the null hypothesis was that this distribution could have occurred without regard to the nature of the plaintiff. Our tests rejected that hypothesis at the .05 level, strongly suggesting that the identity of the plaintiff was important in some way. The same process was followed for the statement about corporations, noted in the text.Google Scholar

33 See text at note 54 infra.Google Scholar

34 The public officials' success rate is significant at the .05 level. See notes 20 and 32 supra.Google Scholar

35 This can be seen by comparing columns 3 and 5 of table 11. The point is even stronger if we break the group of public officials into those running for office and those who are not. Of the latter, 15 of the 16 officials were involved in appellate cases compared with 2 of the 4 incumbents running for office. It is commonly thought that most defamation cases that arise during heated election campaigns do not survive long after the election. Here 50 percent of the cases brought by incumbents were appealed, compared with 94 percent for the noncandidate office holders.Google Scholar

36 See notes 20 and 32 supra.Google Scholar

37 See study I, supra note 3, at 499.Google Scholar

38 Although this print-broadcast difference was not statistically significant (partly because of the absence of any successes against broadcasters), the difference is at least suggestive. See notes 20 and 32 supra.Google Scholar

39 Study I, supra note 3, at 481.Google Scholar

40 This allocation occurred in 23 cases—15 combining moral failings in business and incompetence; 4 combining crime in business and moral failing; and 4 combining crime in business and incompetence. Eight of the 15 combinations occurred as well in the appellate cases, as did both groups of 4.Google Scholar

41 This is derived by adding the 47 moral failing cases to the 12 combinations that include moral failings, making a total of 59 of the 190 appeals that involve moral failings. We are reallocating here by content rather than arbitrarily dividing each case in half, because the question of success allows inferences to be drawn and arbitrary allocations may distort. Although the five victories in cases involving charges of moral failings are suggestive, they are not statistically significant. See notes 20 and 32 supra.Google Scholar

42 The text statement is accurate if the reallocation explained in note 41 supra is accepted.Google Scholar

43 The aberration is significant at the .05 level. See notes 20 and 32 supra.Google Scholar

44 Study I, supra note 3, at 488.Google Scholar

45 Although it is beyond the scope of this study, one wonders whether the 80–20 split is reflected in the allocation of news space or in the expenditures for news gathering and reporting.Google Scholar

46 Books are unusual when it comes to the recency of the events being reported.Google Scholar

47 Why are media sued for these repetitions? First, the original source may be immune, and the plaintiff may feel the need to sue someone, particularly if the charge is false. Second, there is doubt about just what reports may be safely quoted. See Medico v. Time, Inc., 643 F.2d 134 (3d Cir. 1981). Finally, the quotation may be inaccurate or misleading and create potential liability. See study I, supra note 3, at 488.Google Scholar

48 These differences were not significant. See notes 20 and 32 supra.Google Scholar

49 These figures were obtained by combining the 14 appeals involving named sources and the 6 involving confidential sources.Google Scholar

50 Recall from table 6 that the general defense success rate after trial is 59 percent.Google Scholar

51 This difference is significant at the .01 level. See notes 20 and 32 supra.Google Scholar

52 New York Times Co. v. Sullivan, 376 U.S. 254 (1964); Gertz v. Robert Welch, Inc., 418 U.S. 323 (1974).Google Scholar

53 See notes 20 and 32 supra.Google Scholar

54 We have grouped as “state” defenses those that were available long before federal constitutional defenses emerged. We have treated separately two defenses that, although technically available under state law, have undoubtedly been influenced by recent federal constitutional developments. One is the defense that the defendant's utterance was an “opinion” and thus is not actionable. This has been crystallized, if not created, by the dictum in Gertz that there “is no such thing as a false idea.” Gertz v. Robert Welch, Inc., 418 U.S. 323, at 339 (1974). The second arises in the states that have rejected the opportunity offered by Gertz to impose liability for actual injury damages in private citizen cases so long as some fault is shown. This group, of which New York is far and away the most important both doctrinally and quantitatively, requires a private plaintiff to show something substantially more demanding than negligence before recovering any damages in media defamation cases. This state defense is built above a floor created by the federal constitutional defense.Google Scholar

55 The apparent drop in the role of state defenses before and after Wolston is not only totally irrelevant to anything suggested in that case but is also not statistically significant. See notes 20 and 32 supra.Google Scholar

56 The difference between stages of disposition for state and federal defenses is significant at the .01 level. See notes 20 and 32 supra.Google Scholar

57 The Edwards privilege, protecting accurate reports of public statements of some private organizations and persons, might operate at the motion to dismiss stage if the complaint showed the source of the statement and did not allege that it was unfair or incorrect. Edwards v. National Audubon Soc'y, Inc., 556 F.2d 113 (2d Cir.), cert, denied, 434 U.S. 1002 (1977). This might explain the two cases decided at the motion to dismiss stage in which “both” defenses prevailed. Times and Gertz cases were probably the ones at the summary judgment stage.Google Scholar

58 Recall that in cases tried under the Times rule the plaintiff must prove his case with “convincing clarity.” New York Times Co. v. Sullivan, 376 U.S. 254, 285–86 (1964).Google Scholar

59 In table 30, of the 24 Times cases that were won (by defendant) on summary judgment, 14 were among the 115 appeals decided before Wolston, and 10 were among the 75 appeals decided after it. The posttrial figures are 10 pre-Wolston and 5 post-Wolston. These figures suggest that, as should be expected, cases applying the Times rule were decided comparably before and after Wolston.Google Scholar

60 This includes Dixson v. Newsweek, Inc., 562 F.2d 626 (10th Cir. 1977), in which the court used Colorado's rule for private citizens, a more rigorous rule than Gertz demands.Google Scholar

61 Although table 29 shows only 6 such cases, the other 2 wins are to be found under-“both” defenses. Only one plaintiff prevailed in a “super-Gerfz” case. See note 60 supra. One wonders whether a state such as New York, which applies its “super-Gertz” rule so rigorously, might find it easier than states using Gertz to characterize plaintiffs as “private.” The success rate might show no change from Times cases—though the trial rate might more closely approximate Gertz cases than Times cases.Google Scholar

62 The difference here is significant at the .10 level. See notes 20 and 32 supra.Google Scholar

63 Although the dictum in Hutchinson suggested that it was wrong to think that the definition of “public official” included everyone on the public payroll, that cannot be the explanation here. We found no contested case in which a public employee was permitted to proceed under the Gertz rule. Still, the fact remains that courts granted Times protection less often in contested cases after Wolston than before it.Google Scholar

64 See notes 20 and 32 supra.Google Scholar

65 Although we did not code for this, we can tell that in at least eight cases (and probably many more) the defendant asked only for Gertz protection—and, of course, obtained it. In addition, we know that some defendants did not raise points after Wolston that they would have raised earlier. See Medico v. Time, Inc., 643 F.2d 134 (3d Cir. 1981) (defense withdrew part of motion for summary judgment based on Times protection after Wolston was decided).Google Scholar

66 See note 7 supra.Google Scholar

67 Note that the “other” and “not available” categories are quite large and that state defenses were raised more frequently against those categories of plaintiffs. Note also that the two groups that were barred by state defenses—managers and corporations—were also the groups against which the “super-Gertz” defenses operated almost exclusively.Google Scholar

68 Study I, supra note 3, at 494.Google Scholar

69 Id. at 495.Google Scholar

70 Both the difference in plaintiffs' wins and the difference in defendants' wins are statistically significant at the.05 level. See notes 20 and 32 supra. Lawrence Worrall, with Media Professional Insurance Co., of Kansas City, a large libel underwriter, is quoted as talking about a group of three—California, Oklahoma, and South Carolina. Burnett Win May Spark More Libel Suits, L.A. Times, Mar. 28, 1981, § II, at 1. The figures in table 35 would support Worrall's view.Google Scholar

71 Supra note 3, at 497 table 25 note c.Google Scholar

72 Id.Google Scholar

73 See note 12 supra.Google Scholar

74 See notes 20 and 32 supra.Google Scholar

75 Outside the realm of reported litigation, it is difficult to assess the impact of Hutchinson and Wolston. One insurer stated that reported claims against its insureds actually dropped from 692 in 1979 to 651 in 1980. No Rush to Judgment, Am. Law., Mar. 1981, at 16 (quoting executive of Employers Reinsurance Corp., a major libel insurer, outside Kansas City). Cramer, Acts of Malice? Colum. Journalism Rev., July/Aug. 1980, at 17, reports that in the years just before 1979, the claims against clients of Employers Reinsurance had tripled to reach the 1979 figure. See generally Picard, Litigation Costs and Self-Censorship, Freedom of Information Center Report No. 434 (Feb. 1981).Google Scholar