Melbourne Law School - SSRN papers

Melbourne Law School Legal Studies Research Paper No. 94

Defamation Law’s Chilling Effect: A Comparative Content Analysis of Australian and US Newspapers Chris Dent and Andrew T Kenyon

This paper can be downloaded without charge from the Social Science Research Network Electronic Library at: http://ssrn.com/abstract=586684.

Dent and Kenyon, ‘Defamation Law’s Chilling Effect’ (2004) 9 Media & Arts Law Review 89

DEFAMATION LAW’S CHILLING EFFECT: A COMPARATIVE CONTENT ANALYSIS OF AUSTRALIAN AND US NEWSPAPERS CHRIS DENT1 AND ANDREW T KENYON2

ABSTRACT [89] This article reports on a comparative content analysis of more than 1400 Australian and US newspaper articles. The study suggests that in the US — where defamation plaintiffs face much heavier burdens than under Australian law — defamatory allegations are made more frequently against both political and corporate actors than in Australia. The US articles contained apparently defamatory allegations at nearly three times the rate of the Australian sample. In particular, the Australian media appeared to be less comfortable making allegations in relation to corporate affairs than its US counterpart. In addition, some US articles included far more extreme commentary than the Australian sample, which suggests a less restrained style of public debate may be fostered under US law. Through introducing comparative content analysis to Australian media law research, the article supports the idea that Anglo-Australian defamation law has a chilling effect media speech.

Introduction Legal and media commentators frequently argue that defamation law ‘chills’ media speech. Defamation law is commonly said to regulate the publication of material harmful to reputation through balancing two interests — the protection of reputation and the protection of free speech. But some commentators suggest the risks of defamation liability deter publications. The quality of public debate about political and public interest matters is thought to be limited by the media’s fear of lengthy, complex and expensive defamation litigation. 1

Research Fellow, Centre for Media and Communications Law, University of Melbourne:

; PhD, LLB/BA (Hons) (Murdoch). 2

Director,

Centre

for

Media

and

Communications

Law,

University

of

Melbourne:

; PhD (Melb), LLM (Dist) (London), LLB (Hons) (Melb). This research has been supported by Australian Research Council: Kenyon and Marjoribanks, Discovery Project DP0343258, ‘Defamation Law in Context’. Thanks to Tim Marjoribanks for helpful suggestions in preparing the paper.


[90] The term ‘chilling effect’ has long been used to describe this position in the US.3 It also has currency in Australia.4 The term certainly has been used as a rhetorical label in calls to reform defamation law, and it has been used when judges develop traditional law.5 The suggested ‘chilling effect’ has spawned a raft of research, including projects led by Eric Barendt;6 Randall Bezanson;7 Richard Ericson;8 and Russell Weaver.9

Existing research does suggest the concept has value. As Barendt and his coresearchers concluded in their UK research: ‘[J]ournalists, particularly those with

3

For an historical analysis of the term see Frederick Schauer, ‘Fear, Risk and the First Amendment:

Unravelling the “Chilling Effect”’ (1978) 58 Boston University Law Review 685. 4

Eg, Editorial, ‘Time to Deregulate Freedom of Opinion’, The Australian (Sydney) 24 January 2002;

NSW Law Reform Commission, Defamation, Report 75 (1995) 25. 5

New York Times v Sullivan 376 US 254 (1964); Theophanous v Herald & Weekly Times (1994) 182

CLR 104, 155 (Brennan J); Loutchansky v Times Newspapers [2002] QB 783, 817. 6

Eric Barendt, Laurence Lustgarten, Kenneth Norrie and Hugh Stephenson, Libel and the Media: The

Chilling Effect (1997). 7

Randall Bezanson, Gilbert Cranberg and John Soloski, Libel Law and the Press: Myth and Reality

(1987). See also Brian C Murchison, John Soloski, Randall P Bezanson, Gilbert Cranberg and Roselle L Wissler, ‘Sullivan’s Paradox: The Emergence of Judicial Standards of Journalism’ (1994) 73 North Carolina Law Review 7. 8

Richard Ericson, Patricia Baranek and Janet Chan have produced a number of books based on their

research including Visualizing Deviance: A Study of News Sources (1987), Negotiating Control: A Study of News Organization (1989), and Representing Order: Crime, Law and Justice in the News Media (1991). 9

Much of Russell Weaver’s research has been collaborative, eg, Russell L Weaver and Geoffrey

Bennett, ‘Is the New York Times “Actual Malice” Standard Really Necessary? A Comparative Perspective’ (1993) 53 Louisiana Law Review 1153. Smaller projects have been conducted by David Barrett, ‘Declaratory Judgments for Libel: A Better Alternative’ (1986) 74 California Law Review 847; Richard Labunski and John Pavlik, ‘The Legal Environment of Investigative Reporters: A Pilot Study’ (1985) 6 Newspaper Research Journal 13; Michael Massing, ‘The Libel Chill: How Cold is it Out There?’ (1985) Columbia Journalism Review 31; Susan P Shapiro, ‘Libel Lawyers as Risk Counselors: Pre-publication and Pre-Broadcast Review and the Social Construction of News’ in James F Short, Jr and Lee Clarke (eds), Organizations, Uncertainties, and Risks (1992) 131; Stephen M Renas, Charles J Hartmann and James L Walker, ‘An Empirical Analysis of the Chilling Effect’ in Everette E Dennis and Eli M Noam (eds), The Cost of Libel — Economic and Policy Implications (1989) 41.


experience, are well aware of the libel regime in which they are working and, therefore, shape their research and their writing to meet its requirements’.10 Two smaller 1980s studies had similar conclusions: David Barrett found publishers said they withheld stories because of legal concerns,11 and Richard Labunski and John Pavlik suggested most journalists believe some stories are not covered because of defamation concerns.12 But in what ways can a chilling effect be seen to exist? To a large extent, research on the ‘chilling effect’ has used interviews and surveys of journalists, editors and lawyers to discuss the extent of defamation law’s influence on media content. Very little work has examined media product in order to consider the existence of any chilling effect.13 This article forms part of a project that seeks to investigate media production practices under differing Australian and US legal, institutional and social contexts by using a variety of empirical research strategies, content analysis of media products.14

[91] This article aims to introduce and demonstrate comparative content analysis to Australian media law research. Using content analysis is unremarkable in media studies, although it is rare to see legal issues as the focus of study.15 Here a 10

Barendt, Lustgarten, Norrie and Stephenson, above n 6, 68.

11

Barrett, above n 9, 860.

12

Labunski and Pavlik, above n 9, 15. Some other interview-based research has suggested that

professional standards rather than law are more influential: Weaver and Bennett, above n 9. 13

For example, the two most significant earlier studies did not use content analysis: Barendt,

Lustgarten, Norrie and Stephenson, above n 6; Bezanson, Cranberg and Soloski, above n 7. 14

For early results from qualitative interviews with journalists and their legal advisers, which form

another strand of the ARC-funded project, see Tim Marjoribanks and Andrew T Kenyon, ‘Negotiating News: Journalistic Practice and Defamation Law in Australia and the US’ (2003) 25 Australian Journalism Review 31. 15

One Anglo-Australian study examined a number of media releases from the Australian Competition

and Consumer Commission to investigate their role in regulation: Karen Yeung, ‘Is the Use of Informal Adverse Publicity a Legitimate Regulatory Compliance Technique?’ (Paper presented at the Regulatory Institutions Network, Australian National University, Canberra, 12 September 2002) (copy on file). Only one earlier study was found that examined the impact of a specific legal doctrine on newspaper content: Todd F Simon, Frederick Fico and Stephen Lacy, ‘Covering Conflict and Controversy: Measuring Balance, Fairness, Defamation’ (1989) 66 Journalism Quarterly 427. It was a small study of one week’s content from 21 newspapers, which focused on questions of fairness and balance. The study’s authors argued the negligence standard, which many US states apply to private


comparative content analysis of more than 1400 US and Australian newspaper articles was chosen to investigate the impact of law on journalistic writing and publishing.16 It would be difficult to examine law’s impact on media content within a single jurisdiction as so many variables may affect published material, such as journalistic ethics, particular outlets’ editorial choices and individual journalist’s decisions. A comparative approach can assist in focusing on a specific issue of difference — in this case defamation law — and its impact on media content.17 The US makes a good site for comparison because, although the news production cultures are fairly similar to Australia, the defamation law regimes are significantly different.18

This article sets out differences in Australian and US defamation law that are most relevant to the study. It then provides some introductory material on content analysis as a method and outlines the selection and coding processes used in this study before going on to discuss the results. Because of the criteria used in searching for Australian and US material,19 the samples of articles were thought likely to contain defamatory material about political or corporate affairs. The purposive samples of articles

figure defamation plaintiffs, involves determining a journalistic ‘standard of care’ including ideas about journalistic ‘balance’. They also said balance may be a contributing factor to finding ‘actual malice’ for public figure plaintiffs: at 427. But there have been a number of cases in which failing to investigate an allegation (eg not getting the accused’s side of the story) has not constituted actual malice: Murray v Bailey 613 F Supp 1276 (ND Cal 1985); Robertson-Taylor Co v Sansing 10 Media L Rep 1489 (DDC 1984); and Johnson v The Herald Co 116 Mich App 523 (1982). The study’s small size, and its unusual doctrinal approach, mean its suggestion that libel laws have little direct impact on media content can be doubted. For more detail on law’s construction of journalistic standards see Murchison, Soloski, Bezanson, Cranberg and Wissler, above n 7. 16

There has been a great deal of sociological work done on the many factors affecting media product;

for example, Pamela Shoemaker and Stephen Reese, Mediating the Message: Theories of Influence on Mass Media (2nd ed, 1996). But the sociological literature has paid little attention to law’s role. 17

The value in comparative content analyses is evident in Myra Marx Ferree, William Anthony

Gamson, Jurgen Gerhards and Dieter Ruchtet, Shaping Abortion Discourse: Democracy and the Public Sphere in Germany and the United States (2002). For a passionate discussion of the benefits of comparative analysis, albeit in a historical context, see Theda Skocpol, ‘Doubly Engaged Social Science: The Promise of Comparative Historical Analysis’ in James Mahoney and Dietrich Rueschemeyer (eds), Comparative Historical Analysis in the Social Sciences (2003) 407. 18

See eg Marjoribanks and Kenyon, above n 14.

19

The sampling criteria are discussed below in ‘Method: Selection Process’ (see p 98–9).


suggest three points of difference between Australian and US newspaper content. First, the sample of US articles contained allegations, which appeared to be defamatory, at nearly three times the rate of the Australian sample. Second, the balance between allegations against political or corporate figures was very different in each country, and the Australian media may be less comfortable making allegations against corporations and corporate figures than its US counterpart. Third, the range of commentary within US articles included far more extreme material than the Australian sample, which suggests a less restrained style of public debate may be fostered under US law. All these differences support the idea, common in Australian defamation law and commentary, that the Australian law may well have a chilling effect on media content.

[92] Defamation Law Australian and US defamation laws strike different overall balances between reputation and free speech, while within each country there are variations between individual states.20 In Australia, for example, some states operate under a defamation ‘code’,21 others rely on the precedent-driven common law,22 while in New South Wales defamation is common law based but has substantial statutory modification.23 Given this study’s scope and purpose, the law is described in broad terms here,24

20

This lack of uniformity creates a ‘confusing morass of uncertainty’ for the Australian media: Des

Butler and Sharon Rodrick, Australian Media Law (2nd ed, 2004) 26. This uncertainty may, in itself, be considered to act as a ‘chilling effect’. 21

Eg, Queensland: Defamation Act 1889 (Qld).

22

Eg, Victoria and South Australia.

23

Defamation Act 1974 (NSW). For more detail about the Australian jurisdictions see, eg, Sally

Walker, Media Law: Commentary and Materials (2000) 101–2. 24

For more detailed discussion of Australian defamation law, see Michael Gillooly, The Law of

Defamation in Australia and New Zealand (1998); Butler and Rodrick, above n 20; Michael Chesterman, Freedom of Speech in Australian Law: A Delicate Plant (2000). For detail on the US law, see Robert Sack, Sack on Defamation: Libel, Slander and Related Problems (1999); and Libel Defense Resource Center, 50 State Survey 2002–03, Media Libel Law (2002). (The Libel Defense Resource Center is now known as the Media Law Resource Center).


which probably accords with the general legal understanding of most journalists and editors in each country.25 Australian Law Cause of action A traditional commonwealth approach to defamation law requires the plaintiff to prove only three things: first, the defendant published material; second, the material identified the plaintiff; third, the material could be expected to harm the plaintiff’s reputation. There are a number of ways in which ‘defamatory’ can be defined in Australian law, including: ‘a publication, without justification or lawful excuse, which is calculated to injure the reputation of another by exposing [the person] to hatred, contempt or ridicule’;26 words that are ‘calculated to injure … character or reputation in the opinion of right thinking members of the community’;27 and ‘an imputation concerning a person … by which the reputation of that person is likely to be injured … or other persons are likely to be induced to shun, avoid, ridicule or despise that person’.28 Despite the range in definitions, there is no need for any publication to actually damage the plaintiff’s reputation. Once the three elements of publication, identification and defamatory meaning are established, damage to reputation is presumed and the publisher will be liable unless it can establish a defence.

Defences There is greater variation between jurisdictions about available defences. Defences most relevant to the media involve truth, comment, qualified privilege and fair reports.29

[93] (a) Truth and comment 25

In-house and external lawyers assisting the media have far better understandings of defamation law,

but not all, and depending on the organisation perhaps not many, articles are checked by lawyers prior to publication. Thus many articles included in this study would never have been seen by any legally trained person: see Marjoribanks and Kenyon, above n 14, 42–3. 26

Parmiter v Coupland (1840) 6 M & W 105, 108 (Parke B).

27

Slatyer v Daily Telegraph (1908) 6 CLR 1, 7 (Griffith CJ).

28

Defamation Act 1957 (Tas) s 5(1).

29

Another significant defence, absolute privilege, is less relevant to the media as it relates to speakers

within parliaments or courts, see generally Gillooly, above n 24, 149–68; Butler and Rodrick, above n 20, 59–65.


Apart from some defences such as those for fair reports of what is said in parliament or court, the main media defences have been truth or comment. These generally require media publishers to prove that a publication is true,30 or that it is an opinion of a type that a publisher could honestly hold and one based on true facts that the publication also contains.31 So for both defences of truth and comment, establishing the truth underlying the publication’s factual content is a central requirement. This differs from the US position, discussed below, where many plaintiffs must prove actual malice with convincing clarity. In Australia, if the defendant cannot prove material true, by admissible evidence, it often will be liable.

(b) Qualified privilege Generally speaking, the common law defence of qualified privilege reflects an understanding that it ‘may sometimes be necessary to make defamatory statements in order to properly discharge one’s legal or moral obligations or to protect one’s own legitimate interests’.32 Common examples of the privilege include employment references and complaints to investigative bodies, including the police. Usually the privilege only applies to relatively small, identifiable groups that share a common duty and interest in disseminating information.33 This means it is not of great relevance to the media. In NSW, statute has modified the common law position on qualified privilege, while in Queensland and Tasmania it has been replaced by statutory qualified protection.34 Section 22 of the NSW Defamation Act apparently aimed to increase the protection for media publications, but it is widely seen to have

30

In the ACT, NSW, Queensland and Tasmania the defence also requires the publication to have been

in the public interest or for the public benefit: see eg Butler and Rodrick, above n 20, 51–3. 31

See, eg, Defamation Act 1974 (NSW) ss 29–35; Defamation Act 1957 (Tas) s 14; Andrew T Kenyon,

‘Defamation, Artistic Criticism and Fair Comment’ (1996) 18 Sydney Law Review 193. 32

Gillooly, above n 24, 169.

33

The statement’s publisher and all its recipients must have the requisite duty and interest, if not,

qualified privilege is not an available defence: ibid 172. 34

Defamation Act 1889 (Qld) s 16(1); Defamation Act 1957 (Tas) s 16(1). As Gillooly notes, qualified

protection is broader than qualified privilege, and the Acts list circumstances under which the protection will apply: ibid 208.


failed because of the stringent way in which the reasonableness test of s 22(1)(c) has been applied.35

During the last decade, a series of constitutional cases have remodelled Australian defamation law, leading to the Lange qualified privilege defence.36 Australian law now provides a general defence for media publications that cannot be proven true, if the publications are about ‘government and political matters’ — or come within the concept of ‘political communication’ — and the circumstances of publication were ‘reasonable’.37 Generally, before publishing material the defendant must take proper steps to verify the material’s truth and have reasonable grounds to believe it to be true (and not actually believe it to be untrue). In addition, the publisher usually should have sought and published a response [94] from the person or entity defamed in the publication.38 The High Court did not provide a comprehensive discussion of what could be considered to be reasonable,39 and the subsequent case law has done little to

35

Eg Gillooly, ibid 202; Butler and Rodrick are more circumspect, considering that s 22 ‘has not

proved to be a particularly attractive defence for the media’: above n 20, 84. 36

See particularly Nationwide News v Wills (1992) 177 CLR 1; Australian Capital Television v

Commonwealth (1992) 177 CLR 106; Theophanous v Herald & Weekly Times (1994) 182 CLR 104; Stephens v Western Australian Newspapers (1994) 182 CLR 211; Cunliffe v Commonwealth (1994) 182 CLR 272; Lange v Australian Broadcasting Corporation (1997) 189 CLR 520; Levy v Victoria (1997) 189 CLR 579; and eg, Chesterman, above n 24, 15–77. 37

See Lange v Australian Broadcasting Corporation (1997) 189 CLR 520, 574. Lange was an

expansion of qualified privilege, but it arguably covers less than the constitutional defence previously discussed by the High Court in Theophanous and Stephens, see, eg, Andrew T Kenyon, ‘Lange and Reynolds: Lessons for Ireland?’ (Paper presented at the Symposium on Freedom of Expression, Dublin, 5 December 2003). 38

Lange v Australian Broadcasting Corporation (1997) 189 CLR 520, 574.

39

Unlike Reynolds v Times Newspapers [2001] 2 AC 127, which expanded English qualified privilege

and in which Lord Nicholls listed ten factors that are indicative as to the privilege’s availability. After NSW reforms in late 2002, many of the Reynolds factors are raised for consideration under Defamation Act 1974 (NSW) s 22(2A).


clarify the test.40 As Chesterman and others have pointed out, the courts may well apply these tests so that it is very hard for the media to succeed.41

(c) Fair reports The defence for fair and accurate reports is often considered to fall under the law of qualified privilege,42 but there have been suggestions to consider it separately.43 The defence is available only to re-publishers of defamatory statements who must establish the report was fair and accurate. Reports must ‘substantially record what was said and done’44 in the proceedings and cannot be one-sided or biased. In addition, the proceedings reported must be of a nature covered by the defence.45 The types of proceedings covered include courts and parliaments.46

The brief discussion above makes clear that it is not a defence under traditional defamation law for the publisher to show that it was not careless, or that publishing the material served a general public interest.47 This approach to defamation law makes it easy for publications that criticise particular events or entities to convey defamatory 40

See Kenyon, above n 37.

41

Chesterman, above n 24, 98–102; Butler and Rodrick, above n 20, 81. In addition, and in accordance

with qualified privilege generally, proof of malice on the part of the publisher defeats a claim of Lange privilege: Lange v Australian Broadcasting Corporation (1997) 189 CLR 520, 574. 42

Patrick Milmo and W V H Rogers (eds), Gatley on Libel and Slander (10th ed, 2004) 364; Stephens

v Western Australian Newspapers (1994) 182 CLR 211, 246–53. 43

Bellino v Australian Broadcasting Corporation (1996) 185 CLR 183, 215.

44

Waterhouse v Broadcasting Station 2GB (1985) 1 NSWLR 58, 65.

45

Differences between common law and statutory defences for reports in the Australian jurisdictions

have little impact on this study. For discussion of the defence generally, see Gillooly, above n 24, 221– 38. 46

As with qualified privilege, if the report is actuated by malice, the defence is unavailable.

47

Some Australian jurisdictions come closer to these sorts of tests: eg the code defence of qualified

protection may come closer to a general public interest coverage, while the Australian Capital Territory established a defence of ‘non-negligent publication’ in 2002, which as yet has received no judicial interpretation: Civil Law (Wrongs) Act 2002 (ACT) s 66; and see, eg, Matthew Collins, ‘New Defamation Law for the ACT’ (2001) 6 Media & Arts Law Review 335.


meanings and for publishers to face liability. It would seem that under Australian law the media must carefully consider how it would defend material criticising individuals or companies.

US Law As in Australia, US defamation law is principally a matter of state law. But there are broad similarities between the states, in part due to the national reach of the US Constitution and its Amendments.

Cause of Action There are many similarities between Australian and US defamation laws in terms of the basic cause of action. Plaintiffs must show that defamatory material was published about them, with the tests of publication and identification being substantially the same in each country. ‘Defamatory’ in the US also has a similar, wide meaning to Australia. For example, in California a defamatory publication is one ‘which exposes any person to hatred, contempt, ridicule, or obloquy, or which causes him to be shunned or avoided, or which has a tendency to injure him in his occupation’.48 In New York, to defame someone [95] is to expose the person to ‘hatred, contempt or aversion or to induce an evil or unsavory opinion of him in the minds of a substantial number of people in the community’.49 And the Second Restatement of Torts comments:50 ‘a communication is defamatory if it tends so to harm the reputation of another as to lower him in the estimation of the community or to deter third persons from associating or dealing with him’.51 This type of similarity underlies the viability of a comparative study. But a striking difference with US law is that prospective plaintiffs face different burdens depending on their position in society. That is, a

48

Civil Code (California) s 45 .

49

Libel Defense Resource Center, above n 24, 750, citing Mencher v Chesley 297 NY 94, 75 NE 2d

257 (1947). 50

The Restatements of Law are produced by the American Legal Institute and compiled by academics,

judges and practitioners. The Restatements themselves have no formal legal weight, however, they are persuasive as secondary authority and some US states adopt specific provisions of the Restatements as the law: Toni Fine, American Legal Systems: A Resource and Reference Guide (1997) 11. 51

American Law Institute, Second Restatement of Torts (2002) § 559.


public figure plaintiff faces a much higher burden in establishing her or his case than a private figure plaintiff.

(a) The Sullivan rules The US approach to defamation can be called the ‘Sullivan rules’52 because it has been developed in a long series of cases since the classic decision of New York Times v Sullivan.53 Under the Sullivan rules, it is much more difficult to sue than under the traditional approach. This suggests there will be a broader area for public debate under the Sullivan rules and it will be easier for the media to publish critical material. That is, any chilling effect of the law should be less than under the traditional approach. The Sullivan rules require plaintiffs who are public officials or public figures to prove the three elements mentioned above: publication, identification and defamatory meaning. But these plaintiffs must also prove the defendant published with ‘actual malice’.54 This entails proving the publication conveys factual material that is actually false, which the publisher believed to be false when it was published. The plaintiff must prove the publisher actually knew the material was false, or at least that the publisher had a ‘high degree of awareness’ of the publication’s ‘probable falsity’ and recklessly disregarded that danger.55 An additional requirement adds to the difficulties of the public official or public figure plaintiff. Such plaintiffs must establish actual malice with ‘convincing clarity’.56 This is a standard of proof substantially higher than the usual standard in civil litigation.

It is important to note that, although the Sullivan rules apply to most people in the public eye, their precise coverage varies under each state’s law. All elected people and election candidates, as well as public officials of sufficient seniority, will be covered by the Sullivan rules. Whether lower level public officials are included depends on the state in question. For example, all police officers in some states appear

52

Chesterman, above n 24, 155.

53

New York Times v Sullivan 376 US 254 (1964).

54

Ibid 280.

55

St Amant v Thompson 300 US 727, 731 (1968).

56

New York Times v Sullivan 376 US 254, 270 (1964). It is sometimes referred to as the ‘clear and

convincing’ standard.


to be covered,57 while in other states only officers of a high enough rank would have to meet the Sullivan hurdles.58 Similarly, public figures are subject to the constitutional requirements under Sullivan in all states, but who qualifies as a public figure can differ. The characterisation of corporate officers as public figures varies between states. For example, in New York a ‘prominent businessman’59 was held to be a public figure,60 whereas in New Jersey the sole owner of a construction company that [96] only carried out public contracts was not necessarily a public figure.61 This US approach applies to people or entities who are public officials or public figures. Its rationale is that public interest speech will tend to concern these people and entities, and so it is only appropriate to make it harder for them to sue. Instead of suing, they can respond to media attacks with their own speech. The law means the media can criticise public officials and public figures without liability for defamation, so long as the media does not know its publication is false. The US Supreme Court has said common law defences that rely on truth are insufficient to protect public debate, because a ‘rule compelling the critic of official conduct to guarantee the truth of all [its] factual assertions’ inevitably leads to self-censorship62 — that is, it leads to a chilling effect. The Sullivan rules are the way that US defamation law supports the country’s ‘profound national commitment to the principle that debate on public issues should be uninhibited, robust and wide-open’.63

If a plaintiff is not an all purpose public figure — or the publication’s subject matter does not relate to the controversy which made the person a limited purpose public figure — then the plaintiff is considered to be a private figure plaintiff.64 Then the 57

Eg in California — Gomes v Fried 136 Cal App 3d 924 (1982) — and in Massachusetts —

Rotkiewicz v Sadowsky 730 NE 2d 985 (2000). 58

Eg a State Trooper was held not to be a public figure in Louisiana: Trentacosta v Beck 703 So 2d 552

(La 1997). 59

Libel Defense Resource Center, above n 24, 754.

60

Cera v Gannett Co 40 NY 2d 415 (1976).

61

Schiavone Construction Co v Time Inc 847 F 2d 1069 (3rd Cir 1988).

62

New York Times v Sullivan 376 US 254, 277 (1964).

63

Ibid 270.

64

Case law recognises three categories of public figures — all purpose public figures, limited purpose

public figures and, the much rarer, involuntary public figures. Which category a plaintiff may fall into depends on the circumstances of the public controversy in question.


required standard of fault varies depending on the state.65 If a private figure sues over a matter of public concern, the plaintiff still must prove falsity.66 There is much less litigation for purely private concerns and for actions involving non-media defendants. Most States, however, have adopted the negligence standard for private figure plaintiffs.67

(b) Opinion In Milkovich v Lorain Journal,68 the US Supreme Court built on Sullivan to hold that statements of opinion will only be actionable if they contains ‘false facts’.69 This means the mere expression of someone’s point of view will not be actionable. For example, the following statement was made about a lawyer: ‘[H]istory reveals that he will say or do just about anything to win, typically at the expense of the truth’. The trial court held this was opinion, considering that in context, the statement was a ‘collection of opinions, colorfully expressed, which renders [it] … simply more rhetorical hyperbole’. The comment was ‘not provably true or false because there is no core of objective evidence’.70 The breadth of protection for opinion means US law goes far beyond the scope of ‘fair comment’ under Australian law.

[97] Defences

65

The US Supreme Court decision of Gertz v Robert Welch Inc 418 US 323 (1974) held states could

define for themselves much of the law regarding libel suits by private figure plaintiffs. 66

Philadelphia Newspapers v Hepps 475 US 767 (1986). The Supreme Court has avoided commenting

on whether private figures suing over matters that are not of public concern must prove falsity. 67

Eg in California — Brown v Kelly Broadcasting Co 48 Cal 3d 711 (1989) — and in Massachusetts

— Stone v Essex County Newspapers 367 Mass 849 (1975). 68

497 US 1 (1990).

69

Milkovich v Lorain Journal Co 497 US 1 (1990). This was apparently intended to clarify the earlier

ruling in Gertz v Robert Welch Inc 418 US 323 (1974). But Milkovich itself generated much controversy: see eg Kathryn Dix Sowle, ‘A Matter of Opinion: Milkovich Four Years Later’ (1994) 3 William & Mary Bill of Rights Journal 467; M Eric Eversole, ‘Eight Years After Milkovich: Applying a Constitutional Privilege for Opinions under the Wrong Constitution’ (1998) 31 Indiana Law Review 1107. 70

Cochran v NYP Holdings 58 F Supp 2d 1113, 1124 (CD Cal 1998), affirmed on appeal, 210 F 3d

1036 (9th Cir 2000).


As the above outline suggests, defences of truth, comment and qualified privilege do not arise in same manner in US media cases as in Australia. Notwithstanding the requirements for plaintiffs to prove falsity, however, truth remains a complete defence in the US. In the language of a New York court, ‘[i]t is, of course, fundamental that truth is an absolute, unqualified defence to a civil defamation action’.71 Comment will not be actionable unless conveys a false fact, while the Sullivan rules largely supplant arguments in relation to qualified privilege for media cases. But there is a great deal of similarity between the US and Australian defences for fair reports, with the US defence being available if a defamatory statement is contained in a ‘fair and accurate report’ of, for example, court proceedings.72

Contrasting the Laws The similarities and differences between the Australian and US law promote the use of comparative analysis. This study would not be effective if there were no similarities between the laws. The most important is the congruence in the meaning of ‘defamatory’.73 Other similarities include the use of defences such as ‘fair comment’ as it applies to criticism or review, and ‘fair report’. These similarities assist the content analysis. The correspondence in the meaning of defamatory allows articles to be judged according to the same standard in each country. The similarities in defences such as criticism and fair report means articles within these categories can be treated similarly. An additional, minor similarity is that, generally, corporations are entitled to bring defamation actions. In most instances, corporate ‘legal’ persons have similar legal recourse to ‘natural’ persons. But two details are worth noting. First, recent amendments to the NSW Defamation Act74 mean corporations employing ten or more persons cannot bring defamation actions in NSW, although these corporations could

71

Commonwealth Motor Parts v Bank of Nova Scotia 44 AD 2d 375, 378 (1974). In many states, as in

Australia, ‘substantial accuracy is the test for truth or falsity’: Libel Defense Resource Center, above n 24, 753, citing Rinaldi v Holt, Rinehart & Winston Inc 42 NY 2d 369 (1977). 72

For a detailed examination of fair report privilege in the US see David A Elder, The Fair Report

Privilege (1988). 73

The connection between the two countries’ definitions is emphasised by the recognition in Gatley

that the ‘closest to a comprehensive definition [of “defamatory”] is that adopted by the American Law Institute in the Second Restatement of Torts’: above n 42, 29. 74

Defamation Act 1974 (NSW) s 8A.


sue in another state for publication in that state.75 Second, in most US jurisdictions, it would depend on the circumstances as to whether corporations would be deemed to be public or private figures.76

There are at least four significant differences between Australian and US law that may affect media content. One can be seen in the above discussion of Lange and Sullivan. The second relates to different ways in which opinion is treated in each country, while the third concerns burdens and standards of proof. A fourth difference involves two financial matters. The first of these is legal fees. In Australia, winning parties in litigation usually have substantial parts of their lawyers’ fees paid by the losing parties. In the US, each party generally bears its own lawyers’ fees. The second financial matter is damages. Damages awards can be far higher in the United States than in Australia, but they are far less frequent, especially after appeals, under the Sullivan rules.77 Australian awards are lower, but still controversial. Each financial aspect may suggest contrary influences in terms of any chilling effect. The US regime for [98] legal fees may make it substantially more expensive for US media outlets to defend defamation actions, and so chill the media.78 Any such tendency could be lessened because there are markedly fewer defamation trials in the US than in Australia,79 but the tendency could be increased by the comparatively very small risk

75

A corporation’s office holders can still sue personally in NSW if they can satisfy the normal

identification requirement of the cause of action. 76

See, eg, Trans World Accounts v Associated Press 425 F Supp 814 (ND Cal 1977); Reliance

Insurance v Barron’s 442 F Supp 1341 (SDNY 1977). In Massachusetts, however, corporations routinely seem to be considered to be public figures, eg, Foley v Polaroid Corp 400 Mass 82 (1987). 77

For an overview of damages awards in the United States, see Libel Defence Resource Center, Report

on Trials and Damages (2004). 78

The ‘chilling effect’ of damages has long been recognised in Australian-related commentary, eg Alex

Samuels, ‘Problems of Assessing Damages for Defamation’ (1963) 79 Law Quarterly Review 63; John Fleming, ‘Retraction and Reply: Alternative Remedies for Defamation’ (1978) University of British Columbia Law Review 15; NSW Law Reform Commission, Defamation (Discussion Paper 32, 1993) 13; Andrew T Kenyon, ‘Problems with Defamation Damages?’ (1998) Monash University Law Review 70. 79

Eg there appear to have been only five US trials against the media in 2002, which related to

defamation or privacy claims: Media Law Resource Center, Press Release: Trial Records Set in 2002 (2003) .


of paying millions of dollars in damages.80 Overall, US media defendants face unrecoverable legal fees to defend claims, and may well feel compelled to defend those claims vigorously because the quantum of potential damages runs to many millions of dollars.81 Against those possible deterrents to publication, this study examines whether the doctrinal differences related to Lange, Sullivan, opinion and burdens of proof may affect content. As the legal burdens are greater for defendants in Australia, journalists and editors may be more hesitant in printing stories if they have doubts about whether evidence would establish ‘truth’ to the satisfaction of a court. Not only is the US media aware that plaintiffs must prove falsity, public figure plaintiffs also must prove actual malice to the ‘convincing clarity’ standard. All these differences between US and Australian law may contribute to differences in the two countries’ newspapers, which may be observable through comparative content analysis.

Method Content analysis is quite common within media studies,82 although very few studies use law as their focus. In research on media law, empirical sociological methods are not often explored. Both these situations make it useful to outline some general characteristics of the approach.

Content analysis is a ‘flexible’ technique ‘well suited for analysing and mapping key characteristics of large bodies of text’.83 Contemporary uses of the method have 80

Of the five trials cited in the MLRC Press Release, the media won four: ibid.

81

But the cost to them probably should not be seen in as threatening terms as some of the earlier

academic literature suggests. David Anderson, using LDRC statistics, argues that while ‘from 1980 through 2000, the total amount of final libel judgments against media was less than $21 million … this is 0.01 per cent of the media revenues for one year’: David A Anderson, ‘Freedom of the Press’ (2002) 80 Texas Law Review 429, 484 and n 298 (emphasis in original). The $21 million figure was in $US and represented an average award of approximately $260,000. 82

Up to 20 per cent of dissertations and published articles in ‘mass communication research’ use

content analysis techniques: Daniel Riffe and Alan Freitag, ‘A Content Analysis of Content Analyses: Twenty Five Years of Journalism Quarterly’ (1997) 74 Journalism and Mass Communication Quarterly 515, 515–6. 83

Anders Hansen, Simon Cottle, Ralph Negrine and Chris Newbold, Mass Communication Research

Methods (1998) 123. ‘Text’ here can be understood broadly, with content analysis techniques also


examined news, drama and entertainment content in light of various social and cultural issues. The technique is well suited to gain [99] a perspective on law’s current impact on media content. Content analysis also ‘lends itself well to the systematic charting of long-term changes and trends in media’ products.84 This suggests its use for comparative studies across jurisdictions, in addition to studies of a single country or media outlet over time.

One of the classic definitions of content analysis states it is a ‘research technique for the objective, systematic and quantitative description of the manifest content of communication’.85 But other researchers pursue ‘ethnographic content analysis’86 or ‘narrative analysis’87 using similar techniques to ‘document and understand the communication of meaning, as well as verify theoretical relationships’.88 This tension between researchers who consider content analysis to be purely quantitative and those who conduct qualitative studies can be overstated, as Clive Seale has explored in detail.89 Quantitative and qualitative approaches certainly can diverge over their level

being applied to images: at 189–224. Much content analysis of printed text is devoted to the print media, but other sites of study have included National Addresses by the US President: Elvin Lim, ‘Five Trends in Presidential Rhetoric: An Analysis of Rhetoric from George Washington to Bill Clinton’ (2002) 32 Presidential Studies Quarterly 328; and content analysis studies themselves: Riffe and Freitag, ibid. Content analysis focuses on the texts chosen for study, but most content studies do not go as far as Foucault did in that respect. His archaeological method ‘transforms documents into monuments’: Michel Foucault, The Archaeology of Knowledge (1994) 7. This approach leaves the ‘author’ in a questionable position: Michel Foucault, ‘What is an Author?’ in Paul Rabinow (ed), The Foucault Reader (1991) 101. 84

Hansen, Cottle, Negrine and Newbold, ibid 123. Foucault’s archaeological method, ibid, is

ostensibly an historical method. As with content analysis, it has been applied to the law infrequently: see Chris Dent, Reflecting on Continuity and Discontinuity in ‘The Law’: An Application of Foucault’s Archaeological Method in a Reading of Judicial Decisions in Negligence (PhD Thesis, Murdoch University 2002). 85

Bernard Berelson, Content Analysis in Communication Research (1952) 18.

86

David Altheide, Qualitative Media Analysis (1996) 13.

87

Works using this approach are discussed in Bruce Berg, Qualitative Research Methods for the Social

Sciences (3rd ed, 1989) 225. 88

Altheide, above n 86, 16.

89

Clive Seale, The Quality of Qualitative Research (1999). See also Berg, above n 87, 224.


of comfort with the ‘objective’ nature of quantitative methods.90 But here it is more appropriate to emphasise that ‘counting and quantifying’ should benefit from the supplement of ‘interpretive procedures’.91 Thus both quantitative and qualitative data is drawn on in this study to investigate the wider value of content analysis as a method for media law research. Claims made here aim to be based on materials and procedures that are ‘reliable’ and ‘replicable’.92 Necessarily, coding processes rely on varying degrees of reader interpretation, and some of the results reported below are not quantifiable. But as long as the procedures are transparent, then the replicability of the study should enable validation of any results.

Selection Process This study focuses on the apparent impact, if any, of defamation law on media content. Two significant variables to consider were the news medium to be examined and the study’s scale. Limits were needed in terms of the period studied and the sources examined. Any news medium could have been chosen: defamation law applies equally to newspapers, television and the internet. The print media was focused on because of its numerous, accessible outlets and for the ease of comparison — it is much easier to compare two newspaper stories than two television news stories.93 Of the print media, only non-specialist regular publications printed in hard copy in their local areas were considered. That is, daily broadsheet and mass circulation newspapers were included, as were local weeklies and independent monthlies. But specialist publications such as Chemical News & Intelligence, online only publications and wire services were not included. These limits reflect an intention to investigate the print media as generally ‘consumed’ in society. A wide range of newspapers was accessed in each jurisdiction to minimise the impact of editorial policy. The inclusion of independents and local papers reduces the potential impact of some larger media organisations such as News Corporation or Gannett. 90

See eg Altheide, above n 86, 15; and Hansen, Cottle, Negrine and Newbold, above n 83, 94–5.

91

Barrie Gunter, ‘The Quantitative Research Process’ in Klaus Bruhn Jensen (ed), A Handbook of

Media and Communication Research: Qualitative and Quantitative Methodologies (2002) 220. 92

See Seale, above n 89, 140–58.

93

In relation to content analysis of television, see, eg, Simon Cottle, ‘TV Journalism and Deliberative

Democracy: Mediating Communicative Action’ in Simon Cottle (ed), News, Public Relations and Power (2003) 153.


[100] All ‘types’ of articles returned in the searches were considered. That is, stories written by journalists, editorial columns, pieces written by guest columnists and letters to the editor were included. There could have been grounds to not include letters to the editor and guest columnists as these pieces’ authors were not newspaper employees. But they were included because the newspaper would be named as a defendant in any action, and it would have passed the piece through an editorial process. The Lexis-Nexis database was used to source articles.94 Four search terms were used: political, corporate, fraud and corrupt.95 The first two were used in order to find articles that related to business or political activities. The legal protection for speech about each sector has been a prominent issue in the literature.96 In addition, a large enough sample was aimed for to suggest conclusions as to the breadth of the US and Australian exceptions most related to public debate; that is, were more allegations made against politicians because they more clearly fall within the Sullivan and Lange defences than corporate interests do? The latter two search terms were selected to focus on articles likely to contain potentially defamatory statements. The terms were

94

Other studies using this database include Sharon Bramlett-Solomon, ‘Newspaper Editorials Show

Few Regional Differences’ (2001) 22 Newspaper Research Journal 28; Erica Scharrer, ‘An “Improbable Leap”: A Content Analysis of Newspaper Coverage of Hillary Clinton’s Transition from First Lady to Senate Candidate’ (2002) 3 Journalism Studies 393; Clive Seale, ‘Sporting Cancer: Struggle Language in News Reports of People with Cancer’ (2001) 23 Sociology of Health and Illness 308; Gadi Wolfsfeld, Media and Political Conflict: News from the Middle East (1997). 95

The terms were used with ‘wildcards’; that is, these terms were searched for: ‘politic!’, ‘corporate!’,

‘fraud!’ and ‘corrupt!’. Subsequent references in this paper to the search terms include the wildcard versions. These search terms were used after testing a number of different terms including ‘executive’, ‘bribe!’ and ‘misconduct’. The final four were chosen on the basis they provided enough hits, particularly from Australian publications, to suggest trends without producing so many articles as to be unworkable in an exploratory use of the method. 96

See, eg, NSW Attorney-General’s Task Force on Defamation Law Reform, Defamation Law:

Proposals for Reform in NSW (July 2002); Gavin Anderson, ‘Corporations, Democracy and the Implied Freedom of Political Communication: Towards a Pluralistic Analysis of Constitutional Law’ (1998) 22 Melbourne University Law Review 1; Brian Walters, Slapping on the Writs (2003).


used in pairs to search for political and fraud; political and corrupt; corporate and fraud; and corporate and corrupt.97

Four US jurisdictions were chosen: New York (NY), California (CA), Washington DC (DC) and Massachusetts (MA). New York was chosen for its corporate life, the District of Columbia for its politics, California for the size of its market and Massachusetts as an example of a smaller state. Limiting the study to these jurisdictions meant over 40 print publications were included in the searches.98 All Australia states and territories were chosen for comparison. This meant more than 50 publications were included, though many were suburban weekly publications. Both major Australian print outlets, News Corporation and Fairfax, were included, as was one of the independent dailies, The Canberra Times.99 A side effect of these choices was that a broadly similar number of newspapers (particularly for substantial daily newspapers) were searched in the US as were searched in Australia, which may allow more useful comparison of results from each country.100

[101] Searching was conducted to cover one week from each of four consecutive months early in 2003. Choosing that time of year avoided major national election campaigns and the traditionally quieter media time in each country’s summer months. Four separate weeks were chosen with an aim of minimising the effects of single high

97

A limitation in this method should be noted: an assumption in the search strategy may have affected

the results. That assumption is that journalists in both countries use the search terms in similar ways. While coding the articles did not suggest any substantial difference in how the terms are used, it may be that ‘fraud’, for example, is only used in reference to a criminal charge of fraud in one country whereas it may be used more broadly in the other. 98

They included newspapers from major owners such as News Corporation, Knight-Ridder and Hearst.

Prominent dailies on the database included New York Times, Los Angeles Times, Washington Post, Boston Globe and the Boston Herald. Non-daily publications such as the Village Voice, LA Weekly and the New Yorker were also included in the searches. 99

Another independent daily, The West Australian, was not included on the Lexis-Nexis database at the

time of searching. 100

It should be noted the four US jurisdictions had a much higher total population than Australia,

although it may not be more useful to compare numbers of people within the searched territories than numbers of newspapers. But these factors underline the exploratory nature of the results obtained through a relatively confined content analysis, such as the one reported here.


profile stories.101 The weeks chosen were those of 9 to 15 March, April, May and June.102

Coding With the search parameters established, articles fitting the criteria were downloaded and coded.103 Two stages of coding were used.

Stage One In the first stage of coding, articles were classified into a number of categories.104 Each article returned by the search criteria was placed into only one category. As outlined above, the plaintiff must be identified in a successful defamation action. For coding purposes, identification was considered established if a person was named, or if her or his position was identified. For example, if an allegation was made about an unnamed CEO of a named company, or about an unnamed holder of a named public office, then the identification requirement was considered to be satisfied.105

The main intention of coding was to establish whether each article contained statements that could be considered defamatory by an ordinary reader. Every article was read to see if it contained allegations harmful to the reputation of the person or

101

As it turned out, each week, particularly in the US, featured many stories on a single topic, eg, June

had many articles featuring the Sam Waksal/ Imclone/ Martha Stewart issue. 102

Including March in the study, when the Iraq war was most active, did raise a concern there would be

less space available for statements defamatory of people other than Saddam Hussein. But the monthby-month figures in Table 1, below, do not show any substantial difference and suggest the war in Iraq had minimal impact on the use of the search terms in articles. 103

Downloading occurred between September and November 2003. All articles were coded by Chris

Dent, with a number being checked for coding consistency by Andrew Kenyon. 104 105

See below Figure 1 for a summary of stage one of the coding. As mentioned above nn 74–5, corporations employing 10 or more persons cannot sue for

defamation under NSW legislation. But this does not prevent corporations bringing actions in other states, so the amendment was not considered in the coding process.


company concerned.106 In court, it is usually a jury or judge who gets to operationalise the standard of the ordinary reader.107 For the purposes of coding, the standard may have been more sensitive. If there was doubt as to whether an ordinary reader would have found an allegation defamatory, then the article was coded as containing such an allegation. All material tending to lower someone in community estimation was included. It is important to note that the US constitutional requirement for a false fact to exist in order for a defamation to be actionable is a separate element of US law. As discussed above,108 the US tests for what is defamatory are very similar to the Australian law. Overall, the majority of apparently defamatory allegations related to criminal activity, so there was little doubt as to their defamatory status.

[102] Some articles made no defamatory allegations, even though they contained terms such as ‘political’ and ‘corruption’. For example, articles may have discussed movies depicting political corruption, causing them to be retrieved by the search terms although they contained no defamatory allegations. These articles were coded as none. If defamatory allegations were made, then their subject was assessed. If the subject was considered to be a resident of another country, the article was coded foreign. If the subject was not foreign then the nature of the allegation was examined.

Allegations contained in letters to the editor were coded letter. This category was included because editorial practices related to letters’ publication differ to those for articles. No matter who wrote the allegation, the newspaper may be sued, but it was thought useful to see how many sample letters contained allegations. Some other allegations apparently fell into an established defence relevant to the media in both

106

If the only defamatory statement in the article was made about a government department or agency,

in such a way that no specific person could be identified, the article was not counted as containing a defamatory allegation. If the statement was written in such a way that a specific person could be identified, for example the head of the department, then the article was coded as an allegation against that person. 107

For a discussion of role that legal pleadings, particularly their use by the parties’ lawyers, play in the

determination of an article’s meaning, see Andrew Kenyon, Word Games: Meaning in Defamation Law and Practice in England, News South Wales and Victoria (PhD, University of Melbourne, 2002). 108

See above nn 48–50 and accompanying text.


countries; namely, fair reports and criticism or review.109 These allegations were coded according to the defence. That is, if articles appeared to be reports of judicial proceedings, with no other information or commentary, they were coded judicial.110 If articles were critical reviews of movies or books, and the potentially defamatory allegations were part of the reviews, they were coded critic.111 Articles not fitting into any of these categories were coded general to reflect the general defamatory allegations they contained. In addition, general was an overriding code; that is, if a single article contained allegations about foreign nationals and local residents then the article was coded general, or if the article contained both a report of a judicial proceeding and a defamatory editorial comment about the litigants, or any other person, then the article was coded general. The first stage of coding is represented in Figure 1.

Figure 1: Flow chart of coding process Did the article include statements that could be read as defamatory? No → Coded None

Yes

Were the defamatory statements about a foreign national? Yes → Coded foreign

No

Were the defamatory statements contained in a ‘letter to the editor’?

109

Because the study focuses on the comparative frequency with which defamatory allegations are

made, not on whether any legal action would succeed, defences are not a primary focus in the coding. Coding for these defences does assume newspapers did not completely mislead their readers; eg, media outlets did not invent stories about court cases, which there is no reason to think newspapers in either country would do. In any event, the numbers of articles in the defence categories are small enough not to affect overall trends drawn from this study. 110

Articles that appeared to be reports of parliamentary proceedings, with no other information or

commentary, were also to be coded separately. At the end of the coding process, however, only one article was classified as a parliamentary report. This coding category was therefore removed and the one article included under the ‘general’ category. 111

Of course, this may not always operate as a successful defence, especially under the NSW approach

to comment: see eg Kenyon, above n 31.


Yes → Coded letter

No

Were the defamatory statements only a report of judicial proceedings? Yes → Coded judicial

No

Were the defamatory statements about material offered to the public? No → Coded general

Yes → Coded critic

[103] Stage Two The articles coded general were then re-examined. Stage two categorised the subject of the defamatory allegation: was the subject female, male or a corporation.112 Subjects that were ‘natural’ persons (as opposed to the ‘legal’ persons of bodies corporate) were also coded as to their community position. These codes were: •

politician: any person who is or was a candidate or incumbent in a position voted for by the general public — a much wider category in the US than Australia;

•

official: any person appointed to a position of public authority;113

•

corporate figure: anyone with a degree of management authority in a corporation;114

• 112

convicted: anyone convicted of a crime; and

This category was used inclusively — research was not undertaken to establish whether all entities

coded as ‘corporations’ were public companies or had a different financial structure. 113

Positions included under this category were heads of statutory authorities, general managers of local

authorities, judicial officers (including magistrates, but not elected judges in the US who were coded as politicians) and police officers above the rank of Inspector. Lower ranked police officers and nonmanagement public servants were not coded as public officials. 114

Any person described as an executive, a prominent business person or a director was coded as a

corporate figure, unless the context made it clear that the person was the sole director of a small, nonpublic operation. Mid-level managers of large, well known companies were also coded as corporate figures.

Dent and Kenyon, ‘Defamation Law’s Chilling Effect’ (2004) 9 Media & Arts Law Review 89 •

other: all other subjects.

There was no separate category for statements about people charged or indicted over an offence. Statements about people who were to be subject to criminal proceedings were coded as other. Having been charged suggests there may be some factual evidence to justify defamatory allegations. However, given differences in the contempt laws between the two countries,115 and the effect these laws may have on what is published, it was decided that no separate category would be used. A similar study could be conducted that did code for ‘indicted’ or ‘charged’.

Some exceptions were made to the above coding processes to limit skewing of results because of the sample’s timing. For example, allegations about Saddam Hussein being ‘evil’ were excluded.116 But if a specific allegation was made about a Ba’ath party member then it was coded foreign. In the corporate field, passing references to the Enron scandal were not coded as allegations, but allegations made about named persons or office-holders connected to Enron were included.

Results This study’s quantitative and qualitative results are discussed in turn, with the aim of combining the perceived strengths of both numeric presentation and more interpretive approaches.

Quantitative The number of coded articles was 429 in Australia and 986 in the US. These figures are not the ‘raw’ figures produced by entering the search terms into Lexis-Nexis; for example, they exclude duplicate articles117 and articles from Asia-Pacific newspapers

115

See, eg, Chesterman, above n 24, 255–99.

116

Such vague references would in most cases not be actionable. In the US they would be considered

opinion and in both Australia and the US, with the exception of references to Saddam Hussein, there would most likely be doubt as to whether any plaintiffs were identified. 117

Eg an article referring to ‘the fraudulent behaviour of corrupt politicians’ would be found in both the

corrupt-political and the fraud-political searches. In addition, some downloaded ‘articles’ contained more than one ‘story’; eg the search terms could be contained in a page of ‘letters to the editor’. But letters were only coded when the search terms were contained within a single letter.


outside Australia.118 In part, the different numbers [104] of coded articles may reflect population size (20 million in Australia and close to 60 million for California, Washington DC, New York and Massachusetts) and the related differences numbers of newspaper outlets in each country.119 However, the figures still offer useful trends and provide a basis for more research into defamation law’s effect on media content.

Table 1 shows the breakdown of the search results for each month. In both countries, far more articles were returned in relation to political corruption that corporate corruption, although the difference did not appear in relation to political or corporate fraud. In the US, slightly more articles were returned in relation to political than corporate matters — namely, 517 compared with 415 articles, which is 58 per cent for political and 42 per cent for corporate. In Australia, there were noticeably more articles concerned with political matters — 295 compared with 134 articles, which is 69 per cent political and 31 per cent corporate. Even these general results suggest the Australian media may be more comfortable discussing fraud and corruption in political stories than in corporate ones.

Table 1: Search results, number of articles coded

US

Aust 118

March

April

May

June

Total

Political-Corrupt120

82

86

97

97

362

Political-Fraud

61

36

57

55

209

Corporate-Corrupt

9

16

22

28

75

Corporate-Fraud

70

78

71

121

340

Political-Corrupt

64

41

43

54

202

While Lexis-Nexis searches could be conducted easily on individual US states, Australian searches

were conducted by searching the entire Asia-Pacific region. Searches also returned a small number of articles that only contained one of the search terms (eg ‘political’ but not ‘corrupt’) and these articles were discarded. 119

The US searches encompassed 41 newspapers. The Australian searches returned articles from 40

newspapers, including the higher profile dailies of Fairfax and News Corporation, but also including smaller circulation newspapers such as the Albert & Logan News and the Hornsby & Upper North Shore Advocate. 120

In the Tables, plain English forms of the search terms are used but the results include ‘hits’ for the

wildcard searches: see above n 95.


Political-Fraud

26

16

27

24

93

Corporate-Corrupt

15

12

10

8

45

Corporate-Fraud

21

18

25

25

89

Table 2 shows the results of the first coding process. Broadly speaking, there was a much higher proportion of Australian articles categorised as none: 56 per cent compared with 39 per cent in the US. Australia also had more articles coded foreign: 21 per cent compared with 8 per cent in the US. The US had a greater proportion of articles coded general: 43 per cent compared with 15 per cent in Australia.121

At least one factor may influence the difference in results for foreign defamatory statements. There may be more reporting of foreign news in Australia than in the US, either because Australian newspapers covers less local news (perhaps following from its comparatively small population) or because US newspapers do not see their readership as being interested in foreign news. But an argument against this suggestion is the survey included titles such as the New York Times and the Los Angeles Times with a reputation for quality international news.122

[105] Table 2: Articles with potentially defamatory statements: Australia–US Search

Total None

Allegations Foreign

Aust Political

US 121

295

Letters Judicial

Critic

General

5 (2%)

49 (17%)

171 (58%) 56 (19%) 3 (1%)

11 (4%)

Corporate 134

68 (51%)

18 (13%) -

15 (11%)

Total

429

239 (56%) 88 (21%) 4 (1%)

29 (7%)

5 (1%)

64 (15%)

Political

571

252 (44%) 75 (13%) 2 (0.5)

45 (8%)

15 (3%) 182 (32)

32 (24%) 1 (1%)

The numbers of Australian and US articles coded judicial, criticism and letters were similar and

very small. 122

See, eg, Christopher Beaudoin and Esther Thorson, ‘LA Times Offered as Model for Foreign News

Coverage’ (2001) 22 Newspaper Research Journal 80. Another influence on the difference in results may be the coding process and a possible editorial style in US newspapers. Articles were coded foreign only if they included defamatory allegations about foreign citizens but not about residents of the country of publication. Therefore, if US newspapers’ international stories are more likely to include local ‘angles’, then defamatory allegations may be made against both local and international entities. In these instances, articles would be coded general rather than foreign.


Corporate 415

129 (31%) 7 (2%)

3 (0.5)

32 (8%)

4 (1%)

Total

381 (39%) 82 (8%)

5 (0.5)

77 (8%)

19 (2%) 422 (43)

986

240 (58)

One of the strongest results of this survey is the difference in the articles coded general in each country. In particular, the ratio varied greatly between Australian and US articles coded none — which contained no allegations — and those coded general. In Australia, almost four times as many articles were coded none (56 per cent) than were coded general (15 per cent). But in the US slightly less articles were coded none (39 per cent) than general (43 per cent). This was for a sample of articles thought likely to contain defamatory allegations because they used terms related to political and corporate fraud or corruption. Yet only 15 per cent of the Australian articles included general defamatory allegations against Australian individuals or corporate entities. In the US, 42 per cent of the sample contained such potentially defamatory material. That is, the US sample contained general allegations at nearly three times the rate of the Australian sample. This does suggest that US journalists and media outlets are much more comfortable making statements that may harm reputation.

One explanation, distinct from editorial styles or legal limitations, is that the US may have more corruption and fraud than Australia. Different levels of corruption may impact on the level of each term in media content. The latest report from Transparency International does position the US lower on its perception of corruption scale.123 That is, Transparency International found a worse level of perceived corruption in the US than Australia. The relative rankings, however, do not seem proportional to the differences apparent from the coding process.124 Instead, the results do suggest Australian newspaper content is ‘chilled’ in comparison to the US.

Table 3: Articles with apparently defamatory statements: comparative US States Total None

Allegations Foreign

Letters Judicial

Critic

General

123

For information about Transparency International, see generally: .

124

In the report, Australia scored 8.8 while the US scored 7.5 on a scale of 0-10 with 10 being ‘highly

clean’: Transparency International, Corruption Perception Index 2003 (2003).


MA 79

33 (42%)

4 (5%)

-

NY

427

152 (35%) 42 (10%) 3 (1%)

43 (10%) 8 (2%) 179 (42%)

CA

291

114 (39%) 13 (4%)

23 (8%)

4 (1%) 135 (46%)

DC

189

82 (43%)

5 (3%)

6 (3%) 73 (39%)

2 (1%)

23 (12%) -

6 (8%)

1 (1%) 35 (44%)

[106] Table 3 includes the US results on a state-by-state basis. It suggests the proportion of each type of article is broadly similar in each of the four US jurisdictions, although New York and Washington DC do have higher rates in relation to foreign allegations.125

Tables 4 and 5 contain data from the second stage of coding, where the people about whom defamatory statements were made were categorised. Only articles coded general in stage one were coded a second time. The tally does not reflect the number of individuals apparently defamed in the sample, but the number of times people may have been defamed. That is, potential plaintiffs from each article were totalled, and people defamed in more than one article were counted more than once.

Table 4 has the breakdown of male, female and corporations. Many previous studies have found that more males than females are defamation plaintiffs. For example, Michael Newcity examined at all defamation actions active during a two year period in Sydney between 1979 to 1981, and found 8 per cent of defamation plaintiffs were women.126 Andrew Kenyon examined almost every defamation case issued during 1998 in Melbourne and Sydney, as well as most 1997-issued Supreme Court defamation claims. Those files involved 134 plaintiffs in cases against the media, of whom 12 per cent were female.127 The US Iowa Libel Research Project, primarily conducted during the 1980s, is the largest earlier research.128 It used various data 125

It also could be noted that despite substantially higher populations in each of New York and

California than Australia’s total population, the searches did not find correspondingly more articles that raised allegations in these states; see above n 100. 126

Michael Newcity, ‘The Sociology of Defamation in Australia and the United States’ (1991) 26

Texas International Law Journal 1, 13. 127

The 1998 files were drawn on extensively in Kenyon, above n 107.

128

See Bezanson, Cranberg and Soloski, above n 7; Murchison, Soloski, Bezanson, Cranberg and

Wissler, above n 7; Randall P Bezanson, ‘The Libel Tort Today’ (1988) 45 Washington & Lee Law Review 535; Randall P Bezanson, ‘The Libel Suit in Retrospect: What Plaintiffs Want and What


sources, including reported US defamation cases from 1974 to 1984. In that sample, which included a disproportionate number of appellate decisions, plaintiffs were female in 11.5 per cent of cases.129 The number of defamatory statements about men and women in the content analysis reported in this paper suggests one reason for the earlier results. That is, media coverage of the political and corporate spheres does not give women a prominent place, as others have noted.130 In this sample, 10 per cent of defamed entities in the US were female, while in Australia the figure was 7 per cent.

Table 4: Categorisation of entities ‘defamed’ in articles coded general Articles

Total ‘defamed’

Corporations

Male

Female

MA

35

95

17 (18%)

73 (77%)

5 (5%)

NY

179

447

115 (26%)

289 (65%)

43 (10%)

CA

135

334

96 (29%)

201 (60%)

37 (11%)

DC

73

187

34 (18%)

133 (71%)

20 (11%)

US

422

1063

262 (25%)

696 (65%)

105 (10%)

Aust

64

118

14 (12%)

96 (81%)

8 (7%)

[107] Table 5 gives a more detailed categorisation of potential plaintiffs in each country. Coding aimed to identify potential plaintiffs’ public positions, reflecting the focus on political and corporate material, rather than match each country’s legal

Plaintiffs Get’ (1986) 74 California Law Review 789; Randall P Bezanson, ‘Libel Law and the Realities of Litigation: Setting the Record Straight’ (1985) 71 Iowa Law Review 226; Gilbert Cranberg, ‘Fanning the Fire: The Media’s Role in Libel Litigation’ (1985) 71 Iowa Law Review 221; John Soloski, ‘The Study and the Libel Plaintiff: Who Sues for Libel?’ (1985) 71 Iowa Law Review 217. 129

Bezanson, Cranberg and Soloski, above n 7, 7. An earlier US study, that looked at reported cases

rather than all actions, found that 14 per cent of plaintiffs were women: Marc Franklin, ‘Winners and Losers and Why: A Study of Defamation Litigation’ (1980) American Bar Foundation Research Journal 455, 478. 130

See, eg, Newcity, above n 126, 16 who also notes the low rate of female plaintiffs in civil litigation

generally.


categories. For example, people likely to be public figures under US law, such as movie stars and other celebrities, were coded other in this study.131

Table 5: Categorisation of entities ‘defamed’ in articles coded general Total Corp/ns

Corporate figures

Officials

Politicians

Convicted

Other

MA

95

17 (18%)

13 (14%)

18 (19%)

14 (15%)

16 (17%)

17 (18%)

NY

447

115 (26%) 122 (27%) 38 (9%)

22 (5%)

50 (11%)

100 (22%)

CA

334

96 (29%)

44 (13%)

25 (7%)

43 (13%)

39 (12%)

87 (26%)

DC

187

34 (18%)

53 (28%)

21 (11%)

32 (17%)

12 (6%)

35 (19%)

US

1063

262 (25%) 232 (22%) 102 (10%) 111 (10%) 117 (11%) 239 (22%)

Aust 118

14 (12%)

16 (14%)

13 (11%)

42 (36%)

18 (15%)

15 (13%)

In Australia, the highest rate for people ‘defamed’ was for politicians. On these political and corporate search terms, there were more allegations made against political figures (36 per cent) than for corporations and corporate figures combined (26 per cent). The Australian total for politicians and officials was 47 per cent, while in the US it was only 20 per cent.132 Irrespective of the Lange defence’s limited courtroom success, the media may have received some benefit from the High Court decision.133 In any event, the Australian media may be more comfortable criticising political figures.134

131

In addition, someone charged with an offence, but not convicted, would be coded other in Table 5,

unless she or he was a corporate figure, politician or public official. But Table 5 only considers articles coded general at stage one, thus excluding fair reports of court proceedings which were coded judicial in stage one: see above Table 2. 132

These figures refer to potential Australian plaintiffs. As such a large proportion of Australian stories

coded for this study included allegations against non-Australian residents, the status of these potential plaintiffs was also coded. The combined percentage of foreign politicians and public officials was 19 per cent and of foreign corporations and corporate figures was 40 per cent. It should be noted also that almost half of the potential non-Australian plaintiffs were US citizens or corporations. 133

Australian journalists and media lawyers have recently reported that complaints from politicians

have ‘plummeted’: Marjoribanks and Kenyon, above n 14, 38. 134

A wider contemporary or historical content study could investigate this possibility further.


In the US, it appears journalists may be more comfortable making allegations about corporations and corporate figures.135 On this study’s search terms, the combined figure for corporations and corporate figures is 47 per cent. That is, almost half all ‘defamed’ entities came from the corporate sector.136 When compared with the combined total of 20 per cent for politicians and officials, this [108] suggests the US media may be more focused on watching corporate interests than politicians.137 The comparative results in each country are the interesting aspect of these results: with the US prominence of defamed corporations and related people substantially exceeding Australian results. There is doubt under most US states’ laws, and variation between the states, about whether corporations and corporate figures will be classified as public figures and have to meet the Sullivan rules.138 But it is fair to suggest that such entities are far more likely to be covered by Sullivan than their Australian equivalents are likely to be within the extended qualified privilege under Lange. These results suggest the comparative weakness of Australian defences for defamations of corporate interests may relate to much weaker media scrutiny of the sector. This possibility is an important finding of the content analysis.139

A final trend from Table 5 is evident in the figures for other people apparently defamed. The number appears low in each country, considering that in the US it 135

In all four US jurisdictions studied, corporations are either considered to be public figures — eg

Foley v Polaroid Corp 400 Mass 82 (1987) — or the test of whether a corporation is a public figure is the same as for a natural person — eg Reliance v Barron’s 442 F Supp 1341 (SDNY 1977). 136

This figure agrees with statistics about the number of claims against the media. Figures from the

Libel Defense Resource Center show that ‘businesspersons or corporations were the most frequent plaintiff category for every type of media defendant except books’: ‘2001 Complaint Study and Supreme Court Report — 2001 Term’ [August 2002] LDRC Bulletin 8. The LDRC figures, however, include privacy and other claims as well as defamation. 137

The relationship between the media and corporations has spawned a great deal of research. This has

included works criticising the media’s actions: eg, Edward Herman and Noam Chomsky, Manufacturing Consent: The Political Economy of the Mass Media (1988); and work examining the journalistic value of news departments with respect to their parent corporate conglomerates: Dmitri Williams, ‘Synergy Bias: Conglomerates and Promotion in the News’ (2002) 46 Journal of Broadcasting & Electronic Media 453. 138

See, eg, Andrew T Kenyon, ‘Defamation and Critique: Political Speech and New York Times v

Sullivan in England and Australia’ (2001) 25 Melbourne University Law Review 522. 139

It accords with some recent Australian literature such as Walters, above n 96.


includes celebrities and in both countries people charged with, but not convicted of, offences. There is a wide range of people who may be considered at least limited purpose public figures in the US, which could suggest the low other figure is somewhat surprising. But it may result merely from the search terms used. Most public figures who are likely to be defamed in articles found with these search terms may well be politicians, public officials or corporate figures. To support this suggestion, it was notable that few stories in the sample concerned celebrities. They may be defamed within other sorts of material.

B

Qualitative

Qualitative differences could be seen in two general areas — the reporting of court proceedings, and the use of language when making ‘editorial comment’. First, differences were noticed during coding in the way court proceedings were reported in the US and Australia. These differences may be produced by variations in the contempt laws.140 Generally speaking, under Australian law, a media organisation could commit contempt by breaching a suppression order; publishing material with a real tendency to interfere with pending proceedings — which is usually called sub judice contempt; undermining public confidence in the court system more generally than in relation to a specific proceeding; or publishing details of jury deliberations.141 Although coded articles were not read with contempt as the focus, it did not appear that any Australian pieces transgressed the contempt laws. No US articles discussed jurors’ deliberations, while (as for Australia) there was no way of knowing if any suppression orders were broken.142 But the articles did contain discussion of judicial behaviour and a great deal of discussion of court proceedings.143 The articles’

140

See above n 115.

141

For more detail on contempt, and its variation in different Australian jurisdictions, see Butler and

Rodrick, above n 20, 225–97. 142

US judges do seal indictments and other aspects of court files occasionally; eg court records relating

to the search of Michael Jackson’s Neverland Ranch were sealed early in 2004: Tim Molloy, ‘News Groups

Seek

Jackson

Records’,

The

Australian

(Sydney)

8

January

2004

. 143

In terms of judicial behaviour, there were articles discussing the investigation of judges’ financial

records: Denise Buffa, ‘Probe Eyes 21 Brooklyn Judges’, New York Post (New York), 10 May 2003, 5. A fairly typical article on a court proceeding was Tom Robbins, ‘Ex-Giuliani Aide Admits Theft’,


editorial commentary meant they often did not amount to fair [109] reporting of court proceedings. Some articles also included details of the accused’s past convictions. It appears many US articles coded general would not have been published in Australia for fear of committing sub judice contempt.144

Second, there were different approaches to editorial comments in the US and Australia. The relative freedom of such comment in US newspapers was notable. It appeared to follow from the greater protection for opinion under US law.145 Clearly, some statements in US articles could not be published in Australia without a very high defamation risk. Some of the more extreme US statements were: [Californian] Governor/Rotting Carcass/Democrat Gray Davis … An LA Times poll shows that … three out of four people in this decidedly Democratic state believe Davis to be not only incompetent and lacking in leadership but also bereft of principles.146

Van Breda Kolff [head basketball coach at St Bonaventure College] has been an utter embarrassment for Bona. His ridiculous policy to silence the players after losses blew up in his face. He has run amok, alienating people at every turn, treating Olean like some cow town where he’s too slick to get caught. Well it’s VBK who looks silly now. Bonaventure should fire him and refuse to pay him another cent. If he sues, so be it. Better to see him in court than to let him coach one more game on it.147

Village Voice (New York) 13 May 2003, 24. This piece described accused’s demeanour and plea, and referred to allegations and ongoing court cases against people connected with the accused. 144

The greater freedom to report court proceedings in the US is sometimes justified in terms of

protecting the accused’s rights. The argument is that more detailed reporting increases the chances of a person coming forward with information that may benefit the accused; eg the reporting may jog the memory of a witness who had not talked to police. For a discussion of the US system, see Nancy Jean King, ‘The American Criminal Jury’ in Neil Vidmar (ed), World Jury Systems (2000) 93. The justification for the lower level of court reporting in Australia is usually explained in terms of not wanting to prejudice juries through publishing material that may not be legally admissible as evidence. Michael Chesterman has led Australian empirical research in the area of contempt and jury trials, see, eg, Michael Chesterman, Janet Chan and Shelley Hampton, Managing Prejudicial Publicity (2001). 145

See above nn 68–70 and accompanying text.

146

Steve Lowery, ‘Diary of a Mad County’, OC Weekly (Costa Mesa, California), 14 March 2003, 14.

147

‘Bona’s Ax Can’t Spare the Coach’, Buffalo News (Buffalo), 12 March 2003, C1.

Dent and Kenyon, ‘Defamation Law’s Chilling Effect’ (2004) 9 Media & Arts Law Review 89 Some people still insist on seeing Bulger [University of Massachusetts President] as a kind of tragic hero. Others see him as the personification of evil and corruption. Emotions run deep, and neither side is likely to soften its view as long as he still lives. I do not judge his soul, only his actions. And they render him a flawed, crippled messenger at a time when the state college system needs and deserves a much purer champion.148

These highly defamatory statements are probably not actionable in the US because they would be classified as opinion and not be found to convey any false fact. Further, at least two of the targets would be considered public figures. Davis, the elected Governor of California, and Bulger, President of the University of Massachusetts, would be public figures. The third target, the head coach of a college basketball team, may well also be a public figure.149 Therefore, even if these statements were held to convey factual material that was proven false, the plaintiffs would need to establish they were published with ‘actual malice’. The quotes illustrate how the strength of protection for opinionated commentary is far stronger under US law. Certainly, it is unclear that any of the three articles would pass Lange’s reasonableness test. None of the stories included a response from its subject, and suggested no attempt to provide an [110] opposing point of view. The articles appear to convey the journalists’ opinions,150 but it remains unclear and problematic how the Lange defence applies to comment.151

The quotes also highlight differences between the scope of protected public discourse in Australia and the US. While the US law encompasses public officials and public figures, Australia’s approach under Lange extends only to ‘government and political matters’. This may be as narrow as information directly relevant to Australian

148

Joan Vennochi, ‘UMass Needs an Untarnished Champion’, Boston Globe (Boston), 13 March 2003,

A13. Massachusetts newspapers returned much discussion about William Bulger in the searches. The shortest, and perhaps most extreme, comment was: ‘You’re Billy Bulger, the corrupt midget’: Howie Carr, ‘Bunker Hill Day Blues for Bulger’, Boston Herald (Boston), 15 June 2003, 17. 149

The US Supreme Court in Milkovich v Lorain Journal Co 497 US 1 (1990) found that a head high

school wrestling coach was a public figure in the circumstances of that case. 150

The reference to the poll results in the quote about Gray Davis is likely to be defensible with ‘truth’.

151

See, eg, Roy Baker, ‘Defamatory Comment on the Judiciary: Lange Qualified Privilege in Popovic v

Herald & Weekly Times’ (2002) 7 Media & Arts Law Review 213.


electors.152 The quote about Davis could well be within the scope of Lange because it discusses his suitability for the elected office of Governor. But it is much less likely the references to Bulger would be included as governmental discourse, although there may be an outside chance depending on how the position involves public funding and accountability.153 As the law currently stands, defaming a college sports coach could not be brought within Lange qualified privilege.

Although the quoted articles usefully illustrate issues about opinion, it should be noted that such strong comments were fairly rare. Overall, the tone and style of writing did not differ greatly between the two countries. US newspaper journalists did not tend to use extreme language, despite having greater legal protection if they did. For example, a 500 word article reporting charges against an executive stated: ‘federal prosecutors in Houston filed fraud and conspiracy charges last October against former Enron chief financial officer Andrew S Fastow … He pleaded not guilty and is free on $5 million bail’.154 No further editorial comment was made about Fastow. This form of allegation is similar to that found in the Australian press. For example, a story about a former NSW Inspector-General alleged the Inspector-General ‘found “reasonable grounds to suspect corrupt conduct”, including a conflict of interest’ by a named government appointee.155 The article stated the appointee denied the allegations. In most such cases, reporters in both countries repeated allegations made in court, or by investigative bodies, and included the accused person’s response.

V

CONCLUSION

Overall, this study’s results should encourage the use of content analysis for research into law’s impact on the media. Though the scope of research reported here is clearly limited, relating to quite particular samples of newspaper articles, the analysis does 152

See, eg, Herald & Weekly Times v Popovic [2003] VSCA 161.

153

Recent case law suggests that discussion of matters of public administration is capable of falling

within Lange, however, material about the private sector, including not-for-profit entities, are likely to fall outside its scope: see Kenyon, above n 45. 154

Peter Behr, ‘US Charges Loom over Enron Venture’, Washington Post (Washington DC), 11 March

2003, E1. 155

Paolo Totaro, ‘The Execution of a Watchdog Who Dared to Bite’, Sydney Morning Herald

(Sydney), 12 June 2003, 10.


suggest that law makes a difference in terms of what is published in Australian and US newspapers. Quantitative results suggest US journalists more frequently make statements that law treats as harmful to reputation: US articles contained apparently defamatory allegations nearly three times as frequently as the Australian sample. Such allegations are more likely to concern a corporation or corporate executive in the US, while Australian newspapers appear to make fewer defamatory allegations, focused more on politicians and noticeably less on the corporate sector. The study’s qualitative results reinforce the suggestion that law impacts on media content. US law’s greater protection for opinion appears to allow stronger critical commentary than in Australia. Though much of each country’s sample had a similar tone, the US extremes in commentary were much more defamatory than in Australia.

This study has explored the use of comparative content analysis in media law research. It approached Australian and US law in general terms, focussing on what appear to be key differences between the countries. The study has not discussed other journalistic practices that may affect the [111] number of published defamatory statements — for example, self-censorship,156 journalistic ethics,157 or attitudes within particular media organisations.158 The study’s results do suggest the approach’s value for further work, such as historically comparing media content before and after Lange in Australia or Sullivan in the US; or investigating the types of corporations targeted by allegations (for example, are more of them large, publicly-owned companies?). In the sample coded for this study, no substantial difference was apparent in the number of investigative reports in each country. The vast majority of articles in the sample were general news reports about what had happened in the day or week preceding publication. A separate content study is underway, focussed on leading investigative reports in each country, to investigate whether content differences can be seen that may be linked to differences in defamation law.

156

In the words of an Australian journalist: the ‘problem with defamation… is self-censorship. People

consistently don’t write things, don’t write elements, don’t write the guts of it’, quoted in Marjoribanks and Kenyon, above n 14, 38. 157

Ibid 46.

158

Issues such as these are developed in other parts of the full project, see above n 2.


For many commentators and media professionals the existence of a chilling effect would interfere with the media’s believed social role. The press is often identified as the ‘Fourth Estate’,159 with an important ‘watch-dog’ role. The media is seen to carry out an essential task in the processes of accountability.160 If this is an important job for the press, then it is useful to investigate whether laws are limiting the media’s capacity to fulfil these obligations. This study has not aimed to argue that the media’s function is best understood as a check on public power, but it has aimed to establish whether media products appear to be affected by the legal rules relating to defamation. Its conclusion is that newspaper content does suggest the Australian media is chilled in comparison to the US, and unlike the US may well be more hesitant to scrutinise the corporate sector than politicians and public officials.

159

For example, Julianne Schultz, Reviving the Fourth Estate: Democracy, Accountability and the

Media (1998). 160

For example, a ‘strong media is an important counterweight when the political system does not

generate its own checks and balances’: Rodney Tiffen, Scandals: Media, Politics and Corruption in Contemporary Australia (1999) 255.