How Sources, Reporters View Math Errors in News

Research Questions

The need for mathematical competency in the newsroom is well documented, but curiously lacking in the literature is quantitative research gauging the media's ability to report numbers accurately and fairly. Scholars tend to rely on handpicked examples of media misuse of numbers. While a long line of newspaper accuracy studies has identified incorrect numbers as a major type of error, there has been little systematic investigation of the kinds of mathematical errors that appear in newspaper stories. Also absent from the literature is examination of the gravity of mathematical errors in the news media and how these errors occurred. To help address these gaps, this researcher conducted a mathematics accuracy audit of The News & Observer, a seven-day morning newspaper based in Raleigh, N. C. Mathematical errors were examined from two viewpoints: the news source (Part I) and the reporter (Part II). The case study, providing baseline data on numeracy in the newsroom, addressed four research questions:

RQ1:

What is the frequency of mathematical errors identified by news sources cited in a daily newspaper?

RQ2:

What types of mathematical errors are identified by news sources?

RQ3:

Which types of math errors are perceived as most serious by news sources?

RQ4:

How do perceptions of math error differ between news sources and the "offending" journalists?

Part I:

Source Perspective Of Error Method

A five-page, self-administered questionnaire was developed to assess news source perceptions of newspaper accuracy. The accuracy questions closely followed the factual error classifications established by Charnley and the subjective error classifications developed by Berry and his successors.31 Of the 21 error categories, two dealt directly with math-type errors: "numbers wrong"and "numbers misleading or misrepresented." Adhering to the Charnley model permits review of error rates across newspapers and time, although comparisons should be made cautiously because of small differences in methods through the years.

However, the Charnley model does little to reveal the types of numerical error made or the severity of errors, mathematical or otherwise. Following Blankenburg's largely unheeded call for accuracy investigators to distinguish between small and large errors,32 this survey also asked news sources to describe each inaccuracy and to rate the seriousness of each distinct error on a sevenpoint Likert-like scale. In addition, a checklist of 16 types of mathematical errors, drawn primarily from accuracy surveys examining scientific errors in the press, was provided.33

Survey packages were mailed to primary news sources cited in locally produced, bylined news stories appearing in the front, metro and business sections of The News & Observer over a 31-day period. Following Blankenburg's operational definition of a "significantly mentioned" news source, surveys were sent to the first two people who, either as witnesses or participants, have first-hand knowledge of the event.34 Each survey included a copy of the story and a cover letter on university letterhead explaining the purpose of the research. News sources were promised that they would not be identified by name or organization in published results. However, complete confidentiality was not assured: News sources were told that their responses might be shared with the newspaper's editors and reporters in order to trace how errors were made. Non-responding news sources were sent a follow-up survey and letter urging their participation.