New revelation: Almost 98 per cent of errors in US newspapers go uncorrected

Almost half of the articles published by daily newspapers in the US contain one or more factual errors, and less than two per cent end up being corrected.

The findings are from a forthcoming research paper by Scott R Maier, an associate professor at the University of Oregon’s School of Journalism and Communication. The findings challenge how well journalism’s “corrections box” sets the record straight or serves as a safety valve for the venting of frustrations by wronged news sources.

The average US newspaper should expand by a factor of 50 the amount of space given to corrections, says Scott R Maier’s research. Maier, an associate professor at the University of Oregon’s School of Journalism and Communication, describes in a research paper his findings that fewer than 2 per cent of factually flawed articles are corrected at dailies.

The study’s central finding is sobering: 98 per cent of the 1,220 factual newspapers errors examined went uncorrected. The correction rate was uniformly low for each of the 10 newspapers studied, with none correcting even 5 per cent of the mistakes identified by news sources. While it is not plausible or arguably even desirable for every newspaper error to be detected and corrected, Maier noted, the study shows that the corrections box represents the “tip of the iceberg” of mistakes made in a newspaper, therefore providing only a limited mechanism for setting the record straight.

Maier’s findings also challenge journalists’ widely held perception that errors, when detected, are commonly corrected. Previous research showed that news sources brought errors to the attention of newspapers in only about 11 per cent of stories in which errors were identified. Newspapers can hardly be expected to correct errors they do not know were made.

This study, however, shows that even when errors were reported by news sources, the vast majority – 98 per cent – remained uncorrected. In fact, the corrections rate for reported errors is only slightly higher than for errant stories apparently found in error by someone other than the story’s primary source. This suggests that news managers should not rely on corrections as safety valve for the venting of frustrations by wronged news sources, Maier has argued.

Further study is needed to understand why errors, even when reported, go uncorrected. Perhaps news sources didn’t know to whom or how to properly report errors, Maier felt. Reporters and editors, understandably reluctant to make a public mea culpa with published corrections, may have ignored reported errors. Though the study examined only factual errors, differences also may exist between a journalist and a new source as to what constitutes inaccuracy, he pointed out.

This cross-market study of 1,220 errant stories is circumscribed in its statistical analysis by the paucity of published corrections. The data suggest that corrections are more likely to be made when stories are riddled with errors, undermined by egregious errors, or involve front-page coverage. But with only 23 corrections to analyse, these differences fall short of statistical significance.

With such a small base of corrections, it is also difficult to discern whether or not it makes a difference if the errors involves sources who tend to be media savvy – those working in government or community activists with extensive media contact – or an ordinary citizen with little or no experience as a news source. An even larger study is needed to assess these variables on a more robust statistical basis.

On a positive note, Maier said, it is heartening to see that nearly all newspapers routinely invite readers to correct mistakes, providing a corrections box that provides information on how to set the record straight. Most newspapers also make corrections a standing fixture, with publication in a predictable placement either on page A-2 or A-3. Nearly all printed corrections were replicated in newspaper electronic archives.

Even these fixtures were not entirely uniform. Conspicuously missing from the Miami Herald, for instance, is a routinely published corrections policy. Unconscionable is the Grand Fork Herald’s practice of burying its corrections deep inside the paper along side its obituaries. One newspaper failed to append the correction in the electronic version of a story, suggesting that extending corrections to all media formats is not yet automatic.

A newsstand salesman bundles day-old unsold newspapers at his stand in New York’s Times Square. The ‘New York Times’, which began publishing a toll-free number and email address every day for reader input, has seen a marked increase in corrections over the years. A 1982 study found that the ‘New York Times’ printed, on average, one correction per day; in 2004, the ‘New York Times’ printed 3,200 corrections – nearly nine corrections, on average, a day. Objective errors dominate the newspaper’s corrections, though subjective errors might be the greater threat to credibility. (AP)

The study underscores, Maier asserted, that routinely inviting corrections and placing them in a predictable location is not sufficient. With 98 per cent of factual errors uncorrected, innovative ways clearly need to be found to proactively curb inaccuracy as well as to encourage corrections when errors are made. Communication scholars can make a vital contribution in helping the news media set the record straight – and to maintain accuracy as the foundation of media credibility so essential to a working democracy.

Over a two-year period, researchers examined the perceived accuracy of 22 newspapers in 20 US markets. A self-administered mail questionnaire was distributed to 400 news sources identified in each of the markets studied. These were chosen to provide a broad cross-section of newspapers in markets with existing data on newspaper credibility available from the Knight Foundation community surveys. The news accuracy study was funded by the John S and James L Knight Foundation as part of a larger investigation of media quality and profitability.

In a four-page self-administered questionnaire, the accuracy questions closely followed the factual error and subjective error classifications. News sources rated the severity of each error on a 7-point scale. In an open-ended question, sources who did not seek correction of errors were asked to explain why they did not contact the newspaper about the inaccuracies.

Researchers clipped every locally produced, bylined news story appearing in the front, metro, business and lifestyle sections of each newspaper examined until a sample of 400 stories had been compiled. Sports stories, columns, reviews, and editorials were excluded. News sources were defined as the first person in the article who, either as witness or participant, had first-hand knowledge of the event.

Researchers used a variety of web search engines to find source addresses, and, when necessary, made follow-up inquiries by telephone or email to track missing and incorrect contact information. To avoid contaminating the results by making reporters aware that the accuracy of their reporting was being monitored, mailing of the surveys did not begin until all 400 stories had been collected.

Ten newspapers, representing diverse metropolitan dailies, were selected for the follow-up study assessing the extent that errors identified by news sources were corrected. Only “hard” objective errors were examined under the premise that these were factual errors about which there should be little dispute, such as names or addresses, while the basis of subjective errors is by definition open to interpretation.

Drawing on the accuracy audit of the ten newspapers, a database was created of 1,220 news stories with factual errors identified by news sources. The sample provided a total of 2,615 factual errors (many stories had multiple errors). Approximately 11 percent of the respondents said they reported the error(s) to the newspaper.