Explore Harvard's Nieman network Nieman Fellowships Nieman Lab Nieman Reports Nieman Storyboard

A correction on the matter of trust in newspapers

COMMENTARY | May 11, 2005

Credibility is low and declining – but a mislabeled Pew chart, picked up by some news organizations, makes things look a lot worse than they are. [Editor's note: Pew has since changed the chart label. Following this piece is a brief exchange between Andrew Kohut of Pew and Philip Meyer.]

By Philip Meyer



A widely circulated statement about newspapers contains the jarring news that 45 percent of Americans believe almost nothing of what they read in the newspaper. It has appeared in the Philadelphia Inquirer, the Rocky Mountain News, a Ted Rall cartoon, and numerous blogs. You shouldn’t believe it.


Skepticism about newspapers exists, and it is increasing. But the true proportion that believes “almost nothing” is closer to 13 percent than 45 percent.


The problem stems from a misinterpretation of a periodic survey that the Pew Research Center for the People and the Press had been running since 1985. Field work for the most recent tally was in May 2004.


Part of the problem is a mislabeled bar chart on page 49 of Pew’s Trends 2005 report. The bar represents a combined total for two categories of disbelief, but the caption identifies it as showing only the extreme category.


Nicholas Kristof gave the figure a national boost in his New York Times column on April 12 and got a bit closer to the truth when he said, “45 percent of Americans believe little or nothing in their daily newspapers.”


The “little or nothing” is a direct quote from the Pew report, but it is not supported by the data.


The Pew Center has used the same question wording in its eight media credibility surveys since 1985, and the term “little” is not now and was not then in the question. Here is its exact wording:


“Now I am going to read a list. Please rate how much you can believe each organization I name on a scale of 4 to 1. On this four-point scale, '4' means you can  believe all or most of what the organization says. '1' means you believe almost nothing of what they say. How would you rate the believability of the daily newspaper you are most familiar with?”


The 45 percent figure is the result of combining the 1 and 2 responses. But since the 2 was not defined for the interviewees, there is no way to tell what they meant – except that their ratings are in the bottom half of the available choices.


When the Knight Foundation used the same basic question in a survey of the communities where it makes grants, it substituted a verbal scale for the numerical choices.


“Would you say,” it asked, “you believe almost all of what it says, most of what it says, only some, or almost nothing of what it says?”


That is a question whose answers can be interpreted without putting words in the respondent’s mouth.  The Pew report could just as easily have interpreted the 2 on its scale as meaning “some” instead of the more negative “little.”


The absolute meaning is, of course, not as important as the trend, which is indeed pointing downward. In 1985, evaluations on the upper half of the 4-point scale were given by 84 percent of the respondents who were willing to make a judgment.


Last year, the positive responses had fallen to 55 percent.        


The lesson for journalists is to always look at the poll question before interpreting its results. If the exact wording isn’t available, ask the pollster. If your request is refused, complain to the standards chair of the American Association for Public Opinion Research, whose identity can always be found at the AAPOR Website. [Editor’s note: Philip Meyer is a past president of AAPOR.]

The NiemanWatchdog.org website is no longer being updated. Watchdog stories have a new home in Nieman Reports.