Correction Policies of Online Publications

Correction Policies of Online Publications

Keywords: correction, content, policy, Internet, online, electronic, publications, newspapers, media, regulation.

Program Track #2: Media/Information Industry

Peng Hwa ANG

School of Communication Studies

Nanyang Technological University, Singapore

Nanyang Avenue

Singapore 639798

Tel: (65) 790-6109

Fax: (65) 792-7526

E-mail: <mailto:>

Berlinda NADARAJAN

Department of Communication

Stanford University, USA

Stanford, CA 94305-2050

Tel: (1-650) 497-6102

E-mail: <mailto:>

Abstract

On the Net, anyone can be a publisher, so it is said. But as the number of online publications increases, it is clear that the quality of the content varies. In the offline world, such quality is regulated by industry and internal codes of practice. No such codes exist for Internet content.

The capacity for error is magnified on the Internet because information moves at greater speeds and in larger volume. This creates added time pressures for breaking stories and constant updates.

This study surveyed 319 online newspapers. Questions asked include how the policy reflects that used for the print version of the publication, how the procedures are carried out, and the practical problems that arise.

The survey found that two-thirds of the 94 respondents had some form of a correction policy. However, most of them only had their policies developed recently. And most of these policies were not written, only oral.

The errors found were similar to that found in the hardcopy counterpart. Objective-type errors--typos, misspelled names, technical and fact errors--formed the bulk of errors. The minor objective errors can be corrected quickly and even without a notice on the page where the error occurred. Subjective-type errors--omission, misquotes, misleading headlines, and under- or over- emphasis--formed a smaller percentage of errors. But corrections of these may require notices. The paper discusses some implications for editorial policy in implementing some of these correction notices.

A set of recommended guidelines for the development of such correction policies is included at the end of the paper.

As an epilogue, and yet another tribute to the speed of change of the Internet, the study should be inaccurate by the time it appears at INET99: as a result of the survey questions, many online editors said they would put more thought into developing a correction policy for their website.

Contents

Introduction

Literature Review: The Practice of Correction Policies

The legal dimension

The journalistic dimension

The shift to online media

Methods

Analysis of Findings

Presence/absence of correction policy

Key elements and procedures of online correction policies

Problems in applying correction policies online

Comparison of correction policies in print and online media

Implications for Editorial Policy

Conclusion

Epilogue

References

Appendix A: List of Questions Used

Appendix B: Possible Elements for a Correction Policy in the Online Media

Introduction

Online publications are on the rise, tapping the Internet as a publishing tool and a way to reach wide audiences. Yet online material is perceived, perhaps not wrongly, as inaccurate and unreliable (Johnson & Kaye, 1998; Sundar, 1998). Capacity for error is magnified on the Internet, where information moves at greater speed and volume. This creates added time pressures for breaking stories and constant updates. Also, many journalists and writers are using the Internet as a research tool and sourcing mechanism. Such exchanges are informal and chain-like in effect, implying that rumor and fact often intertwine.

Given this scenario of a high possibility of error, the question becomes one of what can be done to ensure accuracy and reliability of information on the Internet. In the journalistic world, one way of doing this has been through the use of correction policies and procedures. This study examines to what extent online publications are applying correction policies to the digital environment, and how they might be doing so. The study looks at how correction procedures are carried out and the problems that arise. It then highlights issues that have to be addressed in applying correction policies to the online world and points to some elements that might be covered in such policies. If well thought-out and used, correction policies might be one way of enhancing the reliability of a new medium that is both celebrated and criticized for the broad range of unfiltered information that it offers.

Literature Review: The Practice of Correction Policies

The legal dimension

Correction policies have been grounded more in practice than in theory, one of these aspects being in the practice of law. Correction statutes have their origins in defamation law, where the statutes spell out the steps a potential plaintiff must take to demand correction of allegedly defamatory material (Burke, 1997). These statutes usually state that a potential plaintiff must give written notice of correction to the publisher of the material. The timeliness of the written notice may determine the type or amount of damages that a plaintiff may receive. The statutes also usually have specifications on the timing, content and placement of the correction in the publication. In the US at least, most of these statutes currently apply only to print, and radio and television broadcasting, although it is being debated whether such statutes should be extended to electronic communications through the US Uniform Correction or Clarification of Defamation Act (ibid.). The proposed Act aims to encourage corrections and clarifications of false statements in all written, broadcast and electronic communications that could hurt a person's reputation (American Society of Newspaper Editors, 1998). The Act was passed in North Dakota in 1995 and other jurisdictions are considering it (ibid.). Thus, current correction statutes have mostly been applied in cases of defamation and do not reflect technological changes that have implications for online publishing. They also do not address other correction issues beyond defamation, such as more general concerns about accuracy of information and the presentation of news.

The journalistic dimension

In journalistic practice, most newspapers publish correction notices to acknowledge errors and have a standardized location for such notices. Editorial policies differ in terms of what constitutes an error, what errors are acknowledged and how to correct them (Barkin & Levy, 1983; Fowler & Mumert, 1988; Wamsley, 1988; Cremedas, 1992). Some of the major studies on corrections are reviewed below.

In a comparative study of two major newspapers, the New York Times and the Washington Post, Barkin and Levy (1983) found that each newspaper averaged about one correction a day for the time under study. Most of the errors in both newspapers (87.7 per cent) were "objective-type errors", i.e. factual and typographical mistakes about times, dates, locations, titles and numbers, while 12.3 per cent were "subjective-type errors" to do with omission, misquotes, misleading headlines, and under- or over- emphasis. About two-thirds of the corrections (64.9 per cent) did not give explanations for the errors. The average time between publication and correction was slightly more than three days, and more than half the corrections (53.3 per cent) were ran within one day of the original, erroneous item.

Whitney (1986) studied the link between correction policies and corrections in 12 newspapers, comprising six national dailies and six smaller dailies. He found that all the newspapers had made corrections during the study period, that the number of corrections was related to the size of the newshole, that reporters and staff spotted more than half of corrected errors, that objective errors were corrected most often, that the newspapers corrected these errors fairly quickly, and that most papers did not have written policies.

Fowler and Mumert (1988) did a survey of 35 daily and 102 weekly Arkansas newspapers. They found that all but one of the 70 who responded reported publishing corrections. About 70 per cent reported having a correction policy and 24.2 per cent reported having a written policy. Some 34.2 per cent reported publishing a correction in the last 30 days, while 30 per cent reported publishing no corrections in the last 30 days, and the others reported publishing at least two or more corrections in the last month. Certain types of errors were more likely to be corrected than others—errors of omission (48.5 per cent), wrong figures (37.1 per cent), wrong dates (34.2 per cent) and typographical errors (32.8 per cent).

In a survey of 223 editors of US dailies and 35 members of the Organization of Newspaper Ombudsmen, Wamsley (1988) found that newspapers with an ombudsman were more likely to have a written, formalized corrections policy. The managing editor was the most likely person to decide whether to run a correction, and often, the writer involved was asked to help in writing the correction notice. The main reasons for running a correction were: identification error (74 per cent), missing information (55 per cent), reader's request (16 per cent), to avoid legal action (11 per cent) and wrong slant or emphasis (4 per cent).

While most of the research on corrections has been done on print journalism, Cremedas (1992) looked at correction policies in local television news by surveying 150 news directors at broadcasting stations. Most of the news stations (47 per cent) averaged one correction every other month, and 15.3 per cent averaged one per month. One positive correlation was found between corrections and market size—the larger the television market, the more likely the station was to correct errors. Almost two-thirds (60 per cent) reported that they did not have a formal correction policy, while just 12 per cent had a written policy. Respondents also reported that they were more likely to correct objective-type rather than "subjective-type" errors, although there were indications that larger TV markets were more likely to correct "subjective-type" errors than smaller markets.

In sum, most news organizations do run corrections, although they tend to correct "objective-type" errors more than they do "subjective-type" errors. Also, correction notices seldom come with explanations as to why and how the errors occurred. Few organizations have a formalized correction policy and even fewer have a written one. While most practitioners agree on the need to correct all substantive errors promptly and prominently, there is less agreement over what constitutes a "substantive" error. In deciding whether to correct an error, editors often consider such factors as local relevance, timeliness, weightiness, impact and context. Conventional practice is not to correct minor errors of deviation, and to correct those that make a difference or need explanation, with some even publishing a correction only if it is requested (Wamsley, 1988).

The shift to online media

If journalism has not had an impeccable record of correcting errors in the print and broadcast media, then it may be said that this record has not been bettered in the shift to online media such as the Internet. This is in part a result of the free-wheeling nature of information on the Internet, and in part a result of how the Internet magnifies seemingly small technical errors.

In terms of content, the Internet has been a fertile ground for various hoaxes (e.g. the Chicago Tribune column by Mary Schmich that was mistakenly circulated as a graduation commencement speech by Kurt Vonnegut) and conspiracy theories (e.g. the Oklahoma City bombing and TWA Flight 800). If one were to pick up the virtual trail and wade through the various pieces of "evidence" on the Web, one is not necessarily left with a clearer idea of what happened—because stories that were apparently refuted do not go corrected and, in fact, continue to stand in opposition to the accepted versions.

Another example of the fluid content on the Internet is the Drudge Report, which promotes a brand of journalism that does not believe so much in accuracy as the tenacity to find the "truth" and publish it. When Brill's Content reviewed 51 "exclusive" Drudge stories between January and September 1998, it found that 31 stories were actually "exclusive". Of these, 10 (32 per cent) were untrue and/or never happened, 11 (36 per cent) were true, and the remaining 10 (32 per cent) were debatable (McClintick, 1998). Yet the Drudge Report will also be remembered for some journalistic scoops that were accurate—it was the first to reveal that Jerry Seinfeld was demanding $1 million an episode for his sitcom, that Connie Chung was being fired as the co-anchor of the CBS Evening News, and that NBC and Microsoft were joining to form MSNBC (ibid.)

The paradox of the Internet is that "fact" and "fiction" can coexist and contradict each other, and that increased coverage does not mean increased understanding. When certain versions of "truth" stand uncorrected, their existence gains weight even as their legitimacy is doubted, and they contribute to the perception that information on the Net is unreliable.

The form and publishing techniques applicable to the Internet do not help either. The rush to break the story has seen errors of a different sort. On November 2 1998, ABC News inadvertently posted its "mockup" US election results during a trial-run on its website, which gave the impression of a sound Democratic victory—a day before polls had actually opened (McKay, 1998). The error was explained as a "technical glitch"--abcnews.com uses a staging server to hold the prepared news before posting it live, but they were apparently unaware that the ticket function uploads information directly. The erroneous material has since been removed, although a copy has been captured by the Drudge Report (http://www.drudgereport.com/abc1.htm). This is not the first time for such slip-ups. In October 1995, the Time-Warner news site Pathfinder pronounced O.J. Simpson "guilty" in his criminal trial moments after the jury had acquitted him. Then in June 1997, abcnews.com posted that Oklahoma City bombing defendant Timothy McVeigh was guilty approximately an hour before the jurors had made their verdict known (Bunn, 1997). While it may be common practice for news organizations to prepare alternate headlines and stories should an event go either way, so that they can get the news out as fast as possible, the Internet context means that technical errors take on an added importance than they do in the physical world.

Errors are not, of course, unique to the digital media—even the traditional wire services have had their share of blunders. United Press International (UPI) "scooped" the world by announcing on its wire in 1918 that the World War was over, when in fact the news are premature—by several weeks. Some observers recall this as one of the worst media mistakes of the century. Then there was the Associated Press (AP) that incorrectly reported that Pope John Paul II had died of his wounds when he was shot in 1981 by a would-be assassin, and had to offer a correction. AP also "killed" Marshall Tito, former leader of Yugoslavia, by erroneously putting his obituary to wire some 20 years ago and causing armies to mobilize in several countries (Adler, 1998). Even the editorial page of the influential Wall Street Journal has been shown to have contained a significant amount of errors (Lieberman, 1996).