Newspaper accuracy: A new approach
Mitchell Charnley concluded his pioneering 1936 study of newspaper accuracy with the hope that future research would provide “a reliable body of data concerning newspaper dependability…,” which eventually would mean that “newspapers–and teachers of journalism–could concentrate intelligently on bolstering the weak points in reportorial and news writing techniques.(1) During the more than 50 years since Charnley wrote, both the newspaper industry and journalism scholars have concentrated on building the body of data and on bolstering accuracy.
However, neither effort has reached satisfactory conclusions. The extensive research has revealed as many complications as solutions. And the industry has satisfied neither itself nor its readers that its reports are adequately accurate. Indeed, a national survey, sponsored by the American Society of Newspaper Editors a half-century after Charnley, found perceived inaccuracy at a level strikingly similar to that he found. The authors of that study, which was intended to assess the broader concept of credibility, offered as their first recommendation for improving credibility: “Be fair, unbiased, accurate, complete, factual, professional, aggressive and compassionate.(2) Similar injunctions to be accurate are found in all the profession’s codes of ethics.
From the beginning, scholars and journalists have agreed that the primary tool for measuring accuracy should be the assessments by sources themselves. Thus, in his first study Charnley sent selected articles to people identified in the articles as sources of information and asked those people to point out and describe any errors in the published work. Along with the identification of inaccuracy, the study also identified a problem that continues to plague scholars and practitioners. The most common type of error cited by the sources was “errors in meaning.(3) The problem was, and is, that meaning lies largely in the mind of the beholder. Context, purpose and perspective all affect meaning. Agreement about meaning is far less likely than is agreement on such objective content as spelling of names, titles or numbers. (Errors in names and titles were the second- and third-most common types Charnley found.)
The significance of this distinction between objective and subjective errors was well illustrated in William Tillinghast’s effort to explain the sources of mistakes. Tillinghast followed generally the well-established procedure of asking news sources to assess the accuracy of newspaper articles. Then he went back to the reporters who had gathered the information and written the articles. He found that reporters were willing to concede mistakes in about half the cases identified by sources as objective errors. However, when sources claimed subjective errors–such as misleading context, misplaced emphasis or inadequate explanation–reporters agreed in only about 5 percent of the cases.(4)
Newspaper efforts to concentrate intelligently on improving accuracy have relied for years on a tool whose own reliability is, in fact, dubious. That tool is an adaptation of the original Charnley methodology, now usually called the accuracy check. Typically, published articles are clipped from the paper and mailed to one or more of the information sources identified within the articles. The clippings are accompanied by a letter from an editor and, usually, a form on which the source is invited to indicate any errors the source finds. The forms also usually ask about the fairness and completeness of the coverage and solicit open-ended comments. These checks are popular with editors, who regard them as serving multiple purposes. At a national conference of newspaper ombudsmen, for example, accuracy checks were described as “a good opportunity to remind the public and reporters and editors that credibility and fairness are the cornerstones of our payer.(5)
However, just as definitional problems undermine the validity of scholarly accuracy checks so do methodological problems call into question the reliability of accuracy checks administered by newspapers. Gilbert Cranberg conducted an accuracy check of the Des Moines Register and compared his results to those obtained in a check by the Register itself. The difference in result was striking. The Register received complaints of inaccuracy from only 14 percent of sources. Cranberg got such complaints from 63 percent.(6)
In his report to editors, Cranberg concluded that the discrepancies are so great that “editors ought to embark on accuracy-checking with caution… . If accuracy-checking is worth doing, it’s worth doing: in a way that assures editors and staffers get all the bad–as well as the good–news about accuracy.”
The most elaborate effort to date t achieve that assurance was conducted by another journalist/scholar, Philip Meyer. Meyer sought to overcome the problems identified by Tillinghast and Cranberg. His goal was to create a feasible management tool for newspapers to use in measuring their own accuracy. His methodology is described in detail in the Newspaper Research Journal.(7) He attempted to bridge the gap between source and reporter by introducing a third-party assessor to determine, after weighing source complaints and reporter responses, which of the complaints had merit. He concluded that such an approach, in any one of several variations, could be workable.
THE MISSOURI EXPERIMENT
Since 1975, the Columbia Missourian, a daily community newspaper published by the University of Missouri School of Journalism, has followed a different kind of accuracy-check policy, one designed to catch and correct mistakes before publication. This policy may be unique among dailies in North America. The policy is simple at least in concept. Reporters are instructed that every story is to be checked for accuracy by recontacting its sources and confirming with them the accuracy of facts and direct quotes. Standing instructions are that the sources should be read enough of the story so that they can judge the fairness of the context in which their information has been used as well as what Meyer called “simple factual accuracy.” The problem of change during the editing process is addressed by having reporters conduct their accuracy checks after the story has been read by a city editor.
Typically, the accuracy checks are conducted by telephone. However, especially in the cases of particularly complicated or sensitive stories, reporters are encouraged to have sources read for themselves the pertinent portions of the story. This policy is followed with particular care in investigative stories. One example may illustrate.
A graduate student spent several months investigating allegations that a local for-profit mental hospital followed a policy of discharging patients on the expiration of their insurance payments rather than basing discharge decisions solely on medical judgments. The reporting began after it was learned that two teenaged boys who committed suicide within days of each other had both been recently discharged from drug treatment programs in the hospital. The reporter examined records, interviewed former patients and family members, found experts and persuaded several former officials of the hospital to talk with him. Then, when his story was written and edited, it was given to top administrators of the hospital for their pre-publication response. Predictably, they were angry. They requested a meeting with the reporter and the Missourian’s managing editor.
During a two-hour session, the officials presented more than 20 objections to the story. In two instances, they had found factual errors (in the job title of one source and in a paraphrase of a hospital policy). Neither was central to the story, but either would have raised doubts about credibility had they been published. Both were corrected on the spot. In every other case of objection, the reporter was able to document his work to his editor’s (if not the officials’) satisfaction. The officials also offered explanations in response to several criticisms in the story. Those were incorporated.
Shortly thereafter, the story, was published with complete confidence in its accuracy and fairness. Hospital officials, who had during the meeting threatened to bring legal action and to protest to administrators of the university, did neither.
This story was of unusual complexity and sensitivity, but the process and the benefits of this type of accuracy check are similar for the full range of stories. Both process and benefits parallel but exceed the recommendations that have been offered by Cranberg Meyer and Tillinghast. The latter’s advice to journalists was typical: “Many of the reporter-acknowledged errors might be prevented if reporters took more care in double-checking information in the interviewing stage and in their paraphrasing or omitting of words in the writing stage.”(8) Of course, only a check of the kind described above can offer real assurance that the degree of reportorial care has been sufficient to avoid error.
It will be immediately obvious that this accuracy-check policy is contrary to the general practice of American newspapers. What isn’t so obvious is why newspapers have refused to take such a step, in view of the continuing industry-wide concern about accuracy. Neither the professional nor the scholarly literature offers an answer. Many of the best-known magazines do follow fact-checking policies, which typically do not extend to reading material back to sources but do at least seek to verify objective accuracy. Perhaps the answer to newspaper reluctance lies in the converse of Cranberg’s speculation about the apparent lack of candor by respondents to newspaper post-publication accuracy checks. He suggested that sources asked to identify errors might hesitate to do so “either to not antagonize or cause trouble for reporters or to avoid appearing to the papers to be complainers.”(9)
One might speculate that journalists equally which to avoid pre-publication conflict. One reason for such a wish, frequently voiced to this writer in conversations with reporters, may be a concern that hard-hitting stories could be softened by pre-publication pressure from influential sources or targets. Or it might be that journalists have more confidence than the public does in their ability to get things right without assistance. Or, a cynic might suggest, editors count on the disinclination of sources to make public complaint even when they perceive inaccuracy, a disinclination substantiated by the work of Cranberg and Meyer.
Two questions remain about the Missourian’s accuracy-check policy: How effective is it in preventing errors, and how is it perceived by the newspaper’s most important sets of constituents, sources and readers? On the first question, there is ample anecdotal but no definitive evidence. The anecdotal evidence strongly suggests that many errors, both objective and subjective, are caught before publication. The newspaper still publishes mistakes, but examination reveals that virtually all result from either post-check alterations and additions (including headlines and photo captions, written at the copy desk and not accuracy-checked) or failures to follow the accuracy-check policy. Such failures chiefly arise from late-night deadline pressures or inability to re-contact sources. (The only exception allowed to the accuracy-check policy is that stories written or edited close to late-night deadlines –the Missourian is a morning newspaper–may not be checked, at the discretion of the editor.)
The paper until recently also employed the same kind of post-publication check used by many newspapers, in which sources were mailed articles and invited to indicate errors. This policy was abandoned for the reason that some editors like it–the return rate was low and the results consistently positive.
Research in 1992 permits more confident answers to the question of source and reader reaction to the policy. To assess source reaction, a questionnaire was mailed to 38 individuals chosen as representative of groups and institutions frequently called on by Missourian reporters.
Sixteen sources completed and returned the questionnaire, a number insufficient to permit any statistical analysis or projections, but sufficient to provide useful information. Confidence in the utility of the results is increased by the strong similarity in their views despite a diversity in their known attitudes toward the newspaper. Respondents included, for example, a chairman of the city’s school board, the senior judge of the circuit court, the presiding officer of county government and a spokeswoman for a local hospital.
Without exception, the sources said they liked the accuracy-check policy. Two-thirds said the policy made them more likely to agree to give information to a Missorrrian reporter. (None said it made them less likely to agree.) One important response was that every source agreed or strongly agreed with the statement “A newspaper that is accurate is more important to me than getting a newspaper that has stories as soon as they break.” Only one source disagreed with the statement “I think all newspapers should have an accuracy-check policy.”
Some individual responses show the value of the policy in the eyes of sources: “I definitely appreciate and respect your accuracy-check policy. Quite honestly, I trust your paper more than the other local in this regard.” “I have recently changed my subscription because I feel the Missourian presents a more accurate, less editorialized, view of news stories.” “Keep it.” “I appreciate your newspaper’s desire to print fresh and honest information. I also enjoy working with your staff. This media interaction is new to me, but I have come to realize the difference between human beings reporting news more through the Missourian than any other entity.”
Some responses also noted the limitations of even this policy: “At least it is an opportunity to re-emphasize the correct information.” “Only the story content is checked for accuracy, not the headline…. I truly detest ‘cute’ headlines on important stories.” “Sometimes reporters are so ignorant or arrogant it bothers me…. As long as other news media tolerate and encourage bad reporting we will continue with that problem. P.S. Please call back to check my tone before you quote this.”
Public attitudes toward the policy were measured with several questions in a broader survey designed to replicate on a local level the national study of attitudes toward newspapers that was conducted for the American Society of Newspaper Editors in 1991. A random sample of 502 residents of Columbia, Missouri and the surrounding county was interviewed by telephone during March and April 1992. The survey was conducted by the Media Research Bureau of the University of Missouri School of Journalism.
After the Missourian’s policy was summarized, respondents were asked, “Would you be more likely to read a newspaper if you knew it had an accuracy-check policy?” The responses: 89.6 percent said Yes; 6 percent said No. Respondents, however, were more forgiving than were sources of published errors. Responding to the statement “I can accept some errors in a newspaper as long as they are acknowledged by corrections”, 85 percent agreed or strongly agreed. Only 11 percent disagreed or strongly disagreed.
The survey also found some public support (again, more than among sources) for the concern among journalists that accuracy checks could soften coverage. Fifty-eight percent agreed or strongly agreed with the statement “I think the accuracy-check policy could make reporters less likely to write news that their sources didn’t like.” Thirty-two percent disagreed or strongly disagreed.
IMPLICATIONS FOR NEWSPAPERS
The Missorrrian, as a teaching newspaper for a school of journalism, is not a typical daily. Its experience with pre-publication accuracy checking, however, may be worth emulating. That experience over 17 years shows that such a method of catching mistakes is practical even under the pressure of daily deadlines. Studies of reaction by sources and readers show that “read-backs,” as the checks are sometimes called, are popular with readers and door-opening with sources. Finally, these pre-publication checks offer an important opportunity not obtainable with post-publication checks or even the type of fact checking employed by some magazines. That is the opportunity to identify, discuss and often resolve–in advance–the subjective errors or perceived errors that more than 50 years of research have shown to be both the most common and the most disputed.
A primary goal of this report is to encourage continued attention to the central problem of newspaper accuracy. Another goal is to encourage other newspapers to experiment with pre-publication accuracy checks in hopes of learning whether this methodology may be at least a preliminary step toward improving newspaper dependability.
1. Mitchell Charnley, Preliminary Notes on a Study of Newspaper Accuracy. Journalism Quarterly, December 1936, p. 394.
2. Kristin McGrath and Cecilie Gaziano, Dimensions of Media Credibility. Newspaper Research Journal, Winter 1986, p. 55.
3. Charnley, op. cit. Typical of later studies substantiating this problem of definition is Michael Ryan and Dorothea Owen, An Accuracy Survey of Metropolitan Newspaper Coverage of Social Issues. Journalism Quarterly, Spring 1977, p. 27.
4. William Tillinghast, Source Control and Evaluation of Newspaper Inaccuracies. Newspaper Research Journal, Fall 1983, p. 13. See also Tillinghast, Newspaper Errors: Source Perception, Reporter Response and Some Causes. ANPA News Research Report 35 (1982).
5. Mark Fitzgerald, Accuracy Checks Grow in Popularity at Newspapers. Editor & Publisher, May 24, 1986, p. 17.
6. Gilbert Cranberg, Do Accuracy Checks Really Measure What Respondents Think about News Stories? ASNE Bulletin, July/August 1987, p. 14.
7. Philip Meyer, A Workable Measure of Auditing Accuracy in Newspapers. Newspaper Research Journal, Fall 1988, p. 39.
8. Tillinghast, op. cit., 1983.
9. Cranberg, op. cit., p. 15.
Copyright Ohio University Winter 1994
Provided by ProQuest Information and Learning Company. All rights Reserved