Method: A Nationwide Poll, With Special Focus on 4 Major Newspapers

Telephone interviews for the Los Angeles Times Poll on the media began Feb. 23 and ended Aug. 2. Initially, 2,993 members of the general public were polled at random.

Each person was asked what newspaper, if any, he or she read most frequently. Based on those responses, a roughly proportional number of interviews were conducted with journalists on the 621 newspapers mentioned (i.e., if 14 members of the general public said they read the Miami Herald, then about 14 reporters and editors at the Miami Herald were interviewed). This was done so that larger (and, presumably, more influential) newspapers would be proportionately represented in the sample.

(Only news and editorial staff members were contacted; photographers, artists, sports, entertainment and business reporters and feature writers in such sections as food, real estate, travel and society were not interviewed. The primary purpose of the poll was to measure the difference between the views of the news staffs and the views of the public on a wide range of issues and then to examine public perceptions of bias, if any, resulting from these differences. It was generally agreed that the views of the non-news staffers did not generally have sufficient impact on the question of ideological bias in the newspaper to warrant inclusion in the survey.)

The Times poll also interviewed the top editor at 587 of the 621 papers mentioned by readers. (At 34 papers--including the New York Times, Los Angeles Times and Washington Post--the top editor declined to participate in the survey.)

All respondents--general public, editors and reporters alike--were asked the same 106 questions.

I.A. Lewis, director of the Times poll and coordinator of this survey, says that in 19 cases out of 20, the responses in a poll such as this will have a sampling error of no more than 2% in either direction from what would have been obtained had every adult in the country been interviewed.

On a few aspects of this poll, the margin of error is slightly larger because the sample is smaller--as with the poll of top editors, where the margin of error is plus or minus 4 percentage points.

The Times poll survey included an “oversample” of 424 readers of the Los Angeles Times to get a more definitive sense of the views of readers of this paper. In addition, the poll attempted to interview virtually every member of the news staffs of the New York Times, Washington Post, Los Angeles Times and Chicago Tribune.

These four newspapers were chosen to provide samples from three different geographic areas of the country (East, Midwest and West) and, more important, because these four (especially the New York Times) are generally regarded as among the most influential newspapers in the country. (In fact, in the Times poll, readers and journalists alike, by a wide margin, say they believe the New York Times is the best newspaper in America.)

If a “liberal media elite” or a “liberal media bias” exists, it seemed useful to compare the views of the staffs of the New York Times, Washington Post, Los Angeles Times and Chicago Tribune to those of the rest of the press and to those of the general public.

As it turned out, no definitive judgments can be made about the staffs of the New York Times and Washington Post because only 215 of 451 people contacted at the New York Times and only 60 of 286 people contacted at the Post agreed to participate in the poll. (At the Los Angeles Times, 224 of 263 people contacted agreed to participate, and at the Tribune, the figure was 164 of 199.)

Lewis said that with groups this small, chosen this way, he would want about 80% cooperation to feel confident of any conclusions he might draw from the results.

It would also be misleading to draw many conclusions from interviews with the readers of each of the four papers as a separate group because the poll interviewed fewer than 60 readers of each paper; each paper has a daily circulation of more than 700,000.

Nevertheless, even based on this incomplete data, there is a pattern to the responses that suggests (though by no means proves) a considerable similarity between the views of the staffs of these four papers (in the aggregate) and the views of the other journalists on most questions, just as there is considerable similarity between the readers of these four papers (in the aggregate) and the readers of the other papers.

The differences seem primarily to be differences in degrees of liberalism: The combined staffs of the two Timeses and the Post would seem, if anything, slightly more liberal than journalists on other papers; the staff of the Tribune is slightly less liberal than journalists on other papers.

The same pattern seems to hold for the readers of these four papers, although the small number of readers interviewed at each paper (except the Los Angeles Times) makes it impossible to make such determinations with any certainty.

Because of the oversample at the Los Angeles Times, one can analyze its readership, as well as its staff, and compare the two. Times reporters and editors are slightly more liberal than reporters and editors in the nationwide sample. However, Times readers are also slightly more liberal than readers nationally. Thus, the ideological gap between The Times staff and its readers is almost identical to that between newspaper journalists nationally and their readers.

Moreover, readers of the Los Angeles Times rate their paper almost exactly the same in terms of impartiality as readers nationwide rate their papers.

When asked, “How would you rate the job (the paper you most frequently read) is doing in terms of how fair and impartial it is to all sides?” 83% of the Los Angeles Times readers say “very good” (32%) or “fairly good” (51%); nationally, 84% of the readers say “very good” (30%) or “fairly good” (54%). Only 11% of The Times’ readers say “very bad” (1%) or “fairly bad” (10%); nationally, 13% of all readers say “very bad” (3%) or “fairly bad” (10%).