AUTUMN 1986 Quarterly Devoted to Research in Journalism and Mass Communicatiorz

~~

By Cede Gaziano and Kristin McGrath

Measuring the Concept of Credibility

Factor analysis produces of a national survey which treated credibil- 12-item credibility factor. ity as a multidimensional concept. Scores for Credibility research has two primary and on this ancestors, the work of Mitchell V. Charn- credibility measure correlated. ley2 on newspaper reporting accuracy and the work of the Yale group on believabil- + Five decades of research on media cred- ity of as a component of persuasive ibility have raised a number of questions communication impact.3 Interest in credi- about the extent of media credibility prob- bility increased after the Roper polling lems.1 Estimates of the magnitude of prob- organization began to ask a question in lems depend on how "credibility" is meas- surveys for the Television Information Office about the relative believability of ured. The purpose of this article is to media and got differing results between describe measurement of this concept in 1959 and 1961, which pointed to increased previous research and to report the results public in television, compared with

I For historical data. see: Cecilie Gaziano and Kristin .4 Since 196 I, television's lead McGrath. "The Media Credibility Problem: Putting the over other media has increased. However, Research into Perspective." paper presented to the American Association for Public Opinion Research. McAffcc. N.J. (1985): Hazel Caudet Erskine. 'The Polls: Opinion of the News Media." Public Opinion Quartrrl?.. 34:63043 (1970-71); D. Charles Whitney. 'The Media and the People: Americans' Experience with the News Media: A Fifty-Year Review." Gan- nett Center for Media Studies. Columbia University. New Kristin McGrath, president of MORl Research, York. NY 10027 (1985). headed the study and coordinated planning with the 2 'Preliminary Not- on a Study of Newspapcr Accuracy." ASNE Credibility Committee, of which David - Journalism Quarterlr. 13394-401 (1936). rence, Jr.. was chairman. Cccilie Gaziano, research 1 Carl 1. Hovland. Irving L. Janis. and Howard H. Kelley. analyst, was project director of the national survey. Communication and Persuasion. (New Haven, Conn.: Yale Funding was provided by: the Associated Press. the University Press. 1953): Carl 1. Hovland and Walter Weiss. ASNE Foundation. the American Newspaper Pub- 7heInfluence ofSource Credibility on Communication Ell=- lishers Association Foundation, Associated Press tivencls," 'Public Opinion Quarterly. 15635dSO. Managing Editors Association, Cox Enterprises. the 4 Burns W. Roper. Public. Attitudes Toward Television ad Gannett Foundation, the Scripps-Howard Founda- Othrr Media in a Time of ChanKe. (New York: Television tion and the Knight Foundation. Special acknowl- Information OW=. 198s). The question was: 'I( you 801 con- flicting or different reports of the same news story from radio. edgement is made of comments and insights offered by televirion. the mapzines and the newspapers. which of the four David Lawrence, Jr.. publisher, the Detroit Frce versions would you be most inclined to believe-the one on Press, and Virginia Dodge Fielder, director of news radio or television or or newspapers?" and circulation research, Knight-Ridder. 45 1

Downloaded from jmq.sagepub.com by FELICIA GREENLEE BROWN on April 12, 2012 452 JOURNALISM QUARTERLY the margin over newspapers has remained different approach to study and each of relatively stable.5 which varied in definitions of credibility.13 The Roper question has been the most A national survey by the Los Angeles frequently used operational definition of Times Poll concerned perceptions of polit- credibility in published research. At least ical bias in newspaper reporting and evalu- 14 reports have used some version of this ations of media in performance, accuracy, question.6 Some of these studies have found that whether the question specifies 5 In 1959.32% said they would believe newspapers. and 29% said they would believe television. In 1%1,24% said they would "local" news stories or "national" news believe newspapers. and 39% said television. In 1984.24%said can influence results on the relative credi- newspapers and 53% said tekvision. bility of television and newspapers. In 6 For example. rce: Lalie A. hxter and John R. Bittncr. 'High School and Colkgc Perccptiona of Media Credibility." some studies, newspapers have been per- Journalism Quarrrrly. 51:517dM (1974); Ronald Mulder, ceived as more believable than television 'Mdb Credibility: A Usc-Gratification Approach," Journal- ism Quarrrrly. 57474476 (1980); Jay Reagan and Jayne Zen- for local news? but this has not always aty. 'Local News Credibility: Newspapers vs. TV Revisited." been so.8 Although this question has been Journalism QuarIrr/v. %I68472 (1979); David Shaw.'Public Finds News Media More Fair than Accurate." Lor Angela pursued for about a quarter of a century, Timrs.pp. I,MZI (&I.II. I98l)ind"PublicRclinonTVfor apparently no researchers have asked New of World. Papers for Local Covcnge. Survey Finds." Lor people if they ever actually have received Angelcs Timrs. pp. 3, 22 (&I. 13. 1981. Ldte Final Edition); Eugene Shaw. 'Media Credibility: Taking the Measure of a conflicting news stories about different Measure." Journalism Quarrrrlv. 50306-31 I (1973). media. 7 James 8. Lemcrt. "News Media Competition Under Condi- When researchers' interest in credibility tions Favorable to Newspapers." Journalism Quarrrrly,47272- 280 (1970); Guido Stempel. 111. 'Effects on Rllormm of a began to mushroom in the 196Os, interest Cross-Media Monopoly." Journalism Monographs. No. 29 also grew in developing the concept more (June 1973). fully. Hovland and his colleagues identi- John D. Abel and Michael 0.Wirth. 'Newspaper vs. TV Credibility for Local News." Journalism Quarrrrly. W371-375 fied two main components of source cred- (1977); Walter Gantz. 'The lnflucna of Research Methods on ibility as "trustworthiness" and "exper- Television and Newspaper News Credibility Evaluations," tise.'- This led some scholars to be Journalo/I*Wdcasting. 25:155--169 (1981); Stempel, op. rir. 9 Hovland. rt al. op. rif.; Hovhnd and Weiss. op. rir. interested in identifying through factor 10 Harwy K. Jacobson. "Mass Media Believability: A Study analysis, underlying dimensions of credi- of Receiver Judgments." Journalism Quartrrly. 46:M28(1969); bility such as "safety," "qualification," "dy- Raymond S. H. Lee. "Credibility of Newspaper and TV News," Journalism Quorrrr/v, 55282-287 (1978); David K. Berlo. namism," "knowledgeability," "accuracy," James B. Lemert. and Robert J. Mertz. "Dimensions for Evalu- "fairness,'' and "completeness," as well as ating the Acceptability of Message Sources." hblir Opinion differences between media and people as Quartrr!i*, 33563-576 (1969-70). 11 Timothy J. Meyer. 'Media Credibility: The State of the communication sources.IO Research." Fublir Trlrrommunirarions Rrvirn: 2(4):48-52 The degree to which people differentiate (1974); Michael Burgoon. Juda K. Burgoon, and Miriam Wil- kinson, "Newspaper Image and Evaluation." Journalism Quar- among different criteria for media per- rrr/v, 5841 1419,433 (1981); Gantz. op. rir. formance, concepts implying credibility, 82 American Society of Newspaper Editors. Nrwspapr Crrd- different kinds of media and various ibiliry: Building Rradrr Trust. American Society of Newspaper Editors. conducted by MORl Research. Inc.. P.O. Box 17004, media functions has not been clear. Sev- Washington. D. C. 20041 (1985). Other ASNE publications eral researchers have concluded that cred- from this study include: 'Focus Groups on Newspaper Credibil- ity" and "Technical Appendix/ Newspaper Credibility: Building ibility is a multidimensional concept, Reader Trust" (both 1985). although the dimensions identified vary An annotated bibliography on media credibility is included in from study to study." the technical appendix. The data tapes for the study have been deposited with the Roper Center at the University of Connecti- Concern that the public distrusts the cut. Inquiries may be addressed In the Center at Box U164. news media and that this distrust may be Univerity of Connecticut. Storrs. CN 06268. Tekphone (203) increasing led the American Society of 4864441. 13 David Shew, 'The Times Poll: Media: High Ratings Arc Newspaper Editors (ASNE) to commis- Tempered," Los Angeles Timrs. pp. I. 8 (Aug. 12, 1985) (LOS sion a more comprehensive look at credi- Angelcs Timrs Poll No. 94); D. Charles Whitney, "The Media bility than any previous research, with a and the People: Soundings from Two Communities." Gannett Center for Media Studies. Columbia University. New York. NY special focus on newspapers.12 It was fol- 10027 (1985); The Proplr & rhr firsst A Timrs-Mirror Investi- lowed in 1985 by at least three other com- gation of Public Attitudes Toward the News Media, conducted by the Gallup Organization in collaboration with Michael J. 3 prehensive surveys of the public on credi- Robinson of George Washington University and the American bility, each of which chose a somewhat Enterprise Institute ( 1986).

Downloaded from jmq.sagepub.com by FELICIA GREENLEE BROWN on April 12, 2012 Measuring the Concept of Credibility 453 fairness, responsibility, and role in criti- unlisted numbers to be included. Up to cism of government. The Gannett Center four attempts were made to reach respon- conducted studies in two communities, dents determined to be eligible for the sur- focusing on the role of public knowledge, vey. Respondents were selected within media use, and personal experience with households by a random method with a media and news in perceptions of credibil- quota for sex. Interviews, lasting 20 min- ity, including variation in opinions by utes on the average, were completed with community. The Times Mirror study used 1,600 adults between December 7, 1984 both cross-sectional and panel national and January 19, 1985. The statistical mar- surveys, scrutinizing public knowledge of gin of error for a random sample of this media, ratings of accuracy and believabil- size is plus or minus 3%. Error margins are ity, press criticism with regard to press larger for sub-groups. behavior and practice, political bias, inde- During the initial telephone interviews, pendence, freedoms, libel law and govern- respondents were asked to give their ment criticism. names and addresses so a second question- The national study sponsored by the naire, could be mailed to them. Those who ASNE and conducted by MORI Research, agreed received a 12-page questionnaire, a Inc., took the broadest approach to defini- cover letter and a $1 incentive. After about tions of credibility among these four sur- a week, these people were re-contacted by veys, highlighting areas in which credibil- telephone so they could read back numbers ity concerns seemed to be greatest.l4 corresponding to their answers on the mail Primary research questions included the questionnaire. Final telephone interviews following: were completed with 1,002 respondents I) What are the dimensions of credibility during this second phase (December 14 to and how are they related to one another? January 30). The completion rate was 58% of contacts with eligible respondents, and 2) How well do people differentiate those taking part in the second phase were between news in newspapers and news in 63% of those completing the first part. The other media? mail sub-sample somewhat underrepre- sented people aged 18 to 24 and 65 or 3) Are credibility problems concentrated in particular coverage areas? Do credibility older, people with lower education and problems increase or decrease as the loca- income, blacks, men, and widowed peo- tion of the news event gets closer to home? ple. It slightly overrepresented people aged 25 to 44, those with higher education 4) How do people feel about First and income, women, married people, and Amendment issues, such as those dealing those who read a newspaper the day with revealing sources and gaining access before. to news evente For the analysis, results were weighted Method by sex, education, household size and race The study had two phases, a series of to match national figures from the US. focus group discussions and a national Bureau of the Census. The weighted total survey of adults aged 18 and older living in of respondents participating in Part 1 was the contiguous . The quan- 1,469. The data reported in this article are titative phase of the project had two parts. based on responses from the mail sub- An initial telephone sample was selected sample, whose weighted total was 875 through random digit dialing to allow persons. '4 The ASNE and Lor Anpln ?imrs Poll surveys also had companion surveys of ncwpapcrjoumalirts on credibility.SCC: The Concept of Credibility Associated Prns MaMging Editors. lournrrlists and Rraders: Bridging the Credibility Cap, cmducled by MORI Research. The ASNE survey used a variety of Inc.. c/o Rokn W. Ritter. Executive Editor, llw Sun, 399 operational definitions of credibility, in- North D Street. San Bcnurdino. CA 92401; David Sbaw. me Times Poll: Public and Pmr-Two Viewpoints." Lo6 Angles cluding broad and narrow measures. In 77mes. pp. IA, 12-13 (Aug. 11. 1985). order to provide a standard against which

Downloaded from jmq.sagepub.com by FELICIA GREENLEE BROWN on April 12, 2012 4 54 JOURNALISM QUARTERLY responses to different types of questions sationalizing, and being moral. Being about the media could be compared, cred- patriotic (rated highly for newspapers by ibility scores for newspapers and television 62% and highly for TV news by 54%) were constructed. People were then divided grouped only with concern for the com- into groups that were high, medium and munity as a third factor for newspapers, low on credibility. and it grouped with sensationalism and The first step in constructing an overall reporter training items as a TV news fac- credibility score was a factor analysis of 16 tor. The method was principal compo- items measuring people’s attitudes toward nents with a varimax rotation.” newspapers on a number of dimensions. People were asked to rate these character- Television and newspaper credibility istics from 1 to 5 on a 5-point scale. Items scores were correlated moderately with were coded or recoded so that they were each other (r = .40, p = .0001), suggesting scaled in the same direction with 5 as the that people’s attitudes toward these media highest possible rating and I the lowest are similar to a certain degree. About 61% possible rating. The question for news- of respondents scored either medium or papers asked people to rate the daily high on the two media, and about 13% newspaper with which they were most were low on both. Another 26% scored familiar.15 The television question ap- low on one but higher on the other. peared in a later section of the question- Respondent ratings were translated into naire and concerned television news. a score by adding up their ratings on each The factor analysis showed that 12 of of the 12 credibility items. The lowest pos- sible credibility score was 12 (less if re- these items grouped together. The items spondents did not give an answer for each concerned whether newspapers and televi- sion news are fair, are unbiased, tell the item), and the highest possible score was 60. Respondents were divided into high, whole story, are accurate, respect people’s medium and low credibility groups. Score privacy, watch out after people’s interests, are concerned about the community’s distributions were skewed toward higher ratings, with about one-fourth of respon- well-being, separate fact and opinion, can be trusted, are concerned about the public dents giving low ratings to each medium interest, are factual, and have well-trained (scores of 20 or less), and about one-third giving high ratings (scores higher than 30). reporters. Because most of these concepts have been treated as indicators of credibil- Most respondents (43% in each case) bunched into a middle group with a nar- ity in previous research, this factor was row range of scores from 21 to 30. Cut-off termed “credibility.” Some previous re- points were selected at 10-point intervals search suggested that some of these items because these breaks were a logical cutting would separate into several factors such as fairness, trustworthiness and accuracy, I’ The question wording was: ”Now. we’d like you to think but that did not happen in this case.16 about the daily newspaper you are the most familiar with . . . Please circle the number in between each pair (of words and Table 1 shows the factor loadings (the phrases with opposite meanings) that best represents how you higher the loading, the greater the rela- feel about the daily newspaper you have in mind.” This wording tionship) of the aspects of the credibility was repeated for ‘television news.” I* Television news items separated into two factors. contain- problem that were combined to create the ing the same items as the newspaper credibility factor. No new overall credibility score. The table also items grouped with these. There appeared to be no reason not shows the proportion of respondents rat- to use the same I2 items in a television credibility score. I’ The Eigenvalue of the newspaper credibility factor (first ing newspapers or television highly on factor) was 5.97. and the final communality estimate was 8.46. each characteristic (4 or 5 on the 5-point Eigenvalues for the televisioncredibility factor (first and second scales). factors) were 6.12 and 1.38, and the final communalityestimate was 9.73. The proportion of total variarrc explained by the newspaper factor was .37. and the proportion of interpreted Items which grouped into a second fac- variance was .7 I. The proportion of total variance explained by tor, termed “social concerns,” included the first and second TV credibility factors was .38 and .09. respectively:the proportion of interpreted variance was .38 and caring what readers or viewers think, sen- .32. respectively.

Downloaded from jmq.sagepub.com by FELICIA GREENLEE BROWN on April 12, 2012 Measuring the Concept of Credibility 455 TABLE 1 Items Chosen for Newspaper and Television Credibility Scoresa

Daily Newspapers Television News % Rating % Rating Newspapers TV News Factor Highly Factor Highly Credibility Factor Loading on Item Loading on Item Is fair or unfair .66 52% .77 52% Is biased or unbiased .66 36% .82 39% Tells the whole story or doesn't tell the whole story .71 43% .7 1 40% Is accurate or inaccurate .69 49% .68 50% Invades or respects people's privacy .66 30% .66 24% Does or does not watch after readers'/ viewers' interests .55 36% .76 31% Is or is not concerned about the community's well-being .47 61% .72 51% Does or does not separate fact and opinion .?I 38% .57 38% Can or cannot be trusted .72 51% .57 48% Is concerned about the public interest or is concerned about making profits. .57 42% .70 3 8% Is factual or opinionated .7 I 50% .63 48% Has well-trained or poorly trained reporters .57 48% .36 57% Social Concerns Factor Cares or does not care what audience thinks .7 I 44% .78 44% Sensationalizes or does not sensationalize .55 30% .38 2% Is moral or immoral .77 53% .74 47% aSubsample: Those who participated in Part 2 interview (N = 875). point and they provided large enough concern for the community's well-being groups for analysis. and patriotism. About 5 in 10 rated televi- sion news highly on these characteristics When all of the credibility and non- (t-tests significant at p < .0001). About 5 credibility items are taken into account, in 10 evaluated newspapers and TV news the percentages of people rating newspa- this favorably for fairness, trustworthi- pers and television news highly (4 or 5 on a ness, morality, accuracy, and being fac- 5-point scale) on these characteristics were tual. Very few people gave either television relatively similar, although t-tests on dif- or newspapers low ratings on these char- ferences in mean scores were statistically acteristics. T-tests were significant only for significant for 10 of the 16 items. Newspa- morality, with newspapers receiving slight- pers were rated highly by 6 people in 10 on ly higher scores (p <_ .001).

Downloaded from jmq.sagepub.com by FELICIA GREENLEE BROWN on April 12, 2012 456 JOURNALISM QUARTERLY TV news and newspapers did relatively ity scores tended to select television. How- less well (high ratings by about 4 in 10) on ever, a large proportion of those scoring being unbiased, caring what the audience high on newspaper credibility would be- thinks, separating fact and opinion, watch- lieve television. ing out for their audiences’ interests, tell- A follow-up question, “Why do you feel ing the whole story, and being concerned that way?,” was then asked. The most fre- for the public interest. T-tests were signifi- quent reason people gave for selecting tele- cant for higher newspaper scores in the vision was that “seeing is believing.” One last three areas mentioned, as well as for a respondent said, “You can see their eyes. higher television score on lack of bias. You can tell if they’re lying.” Other rea- Areas in which TV news and newspapers sons for believing television were that it were especially likely to receive low ratings delivers news fast, easy and first, has some were respecting people’s privacy and sen- advantage related to credibility, or that sationalism, with newspapers receiving other media are unsatisfactory. People slightly higher scores (p < .0001). who said they would believe newspapers Credibility scores were then used as most often cited the ability of newspapers “yardsticks” with which to compare results to provide detail, to take time to prepare on other credibility measures, including reports, or to have some advantage related believability when news reports conflict; to credibility. media preferred for local, state, national or international news; reliability of cover- Which medium people would believe is age of specific topics; and responsible an important consideration. Half of those exercise of First Amendment rights. Re- surveyed said they had heard conflicting sults given for credibility scores in the fol- reports of a news story from different lowing tables compare the extreme groups news sources. When asked to provide (high and low). Four other sets of ques- details about this, half of these respon- tions also were chosen as indicators of dents said that the conflict involved two or credibility, since most credibility studies more different media, such as television rely on only one or two definitions of cred- and newspapers.20 About one-third said ibility. These questions were: a) the Roper that two or more of the same kind of poll question about media believability medium were involved, such as several TV when news reports conflict, b) preferences stations, or they pointed out differences for media to aid understanding of com- within sections of the same newspaper. plex or controversial issues.18 c) prefer- When asked what conflicting reports were ences for reliable news coverage,l9 and d) about, people most often mentioned break- news source preferences. ing news stones about crimes, disasters, assassinations and attempted assassina- Comparisons of Dimensions tions involving public figures. Participants were asked the Roper Exposure to conflicting reports dimin- believability question, and their answers ishes media credibility. Almost two-thirds were quite similar to Roper’s results. They of those with low newspaper or TV credi- would be most inclined to believe televi- bility scores were aware of conflicting sion as to believe newspapers by a 2-to-I margin (52% versus 25%). Nine percent “Sn:Barry Sussman. ‘Public Has Sharp Complaints About said radio, and 14% said magazines. Cred- News Media. Poll Says,” Washington Posr, pp. Al-2 (Aug. 16. 1981); Washington Posr. ‘Public Attitudes Toward the Media ibility scores were related to the choice of Survey.” Washington Post Company (1981). Their question newspapers or television as the medium to was: ‘If there is a situation in thc news that is hard to under- stand or controversial. which pan of the major news medm believe in the face of conflicting reports would you trust the most to help you understand it?” (Table 2). For instance, those who gave I* See: Shaw (1981). op. cil.; 1. A. Lewis. ‘The Media,“ Los newspapers high credibility scores were Angela rims Poll No. 46. Times-Mirror Company (1981). more likely to believe newspapers than Their question was: ”Doer the new seem more reliable to you when you yc it on or when you read it in a newspapcfT those with low newspaper credibility TV. The next question was: ‘Please tell mc what the circum- scores. Those with high television credibil- stances were and whert you heard thc conflicting reports.”

Downloaded from jmq.sagepub.com by FELICIA GREENLEE BROWN on April 12, 2012 Measuring the Concept of Credibility 457 TABLE 2

Media Comparisonsa

I. BELIEVABILITY Newspaper Television Credibility Credibility a. If you got conflicting or score score different reports of the same news stow from LOWb HIGH LOW HIGH radio, television, maga- zines, and newspapers, Television 49% 51% 46% 61% which of the four versions Newspapers 15 34 23 23 would you be the most Magazines 22 10 19 I1 inclined to believe - the Radio 14 5 12 6 one on radio or television or magazines or newspapers? b. As far as you can recall offhand, have you ever heard conflicting reports of the same story from differ- Yes 62% 46% 67% 42% ent news sources. No 38 54 33 58

2. UNDERSTANDING THE NEWS a. If there is a situation in your local area that is hard to understand or is contro- versial, which one of your Television 33% 30% 29% 41% local news media would Newspaper 48 61 57 48 you trust the most to to Magazines 3 2 4 1 help you understand it? Radio 16 7 10 10 b. If there is a situation in the national news that is hard to understand or is controversial, which one of Television 55% 54% 48% 64% the news media would you Newspapers 24 35 33 28 trust the most to help you Magazines 12 9 13 5 understand it? Radio 10 4 9 5

3. RELIABILITY OF NEWS COVERAGE a. Overall, does local and state news seem more reliable when you see it on television, read it in a Television 58% 52% 53% 66% newspaper, hear it over the Newspapers 24 42 33 21 radio, or read it in a Magazines 8 2 6 2 ? Radio 10 4 9 5 b. Overall, does national and international news seem more reliable to you when you see it on Television 67% 68% 61% 80% television, read it in a Newspapers 15 22 19 14 newspaper, hear it over the Magazines 10 9 14 5 radio, or read it in a Radio 8 I 5 2 magazine?

Downloaded from jmq.sagepub.com by FELICIA GREENLEE BROWN on April 12, 2012 458 JOURNALISM QUARTERLY TABLE 2 continued 4. NEWS SOURCE PREFERENCES a. If you had to choose one Television 45% 46% 47% 54% source for local news, Newspapers 33 43 36 33 which source would you Magazines 2 I 2 1 choose - newspapers, Radio 20 10 15 12 television, radio, or magazines? b. If you had to choose one Television 58% 53% 51% 62% source for state news, Newspapers 29 41 38 33 which source would you Magazines 3 I 2 I choose - newspapers, Radio 10 5 9 5 television, radio, or magazines? c. If you had to choose one Television 69% 69% 64% 78% source for national and Newspapers 14 25 21 17 international news, which Magazines 8 5 8 4 source would you choose Radio 9 1 6 I - newspapers, television, radio, or magazines? aSubsample: Those who participated in Part 2 interview -newspaper credibility, N 862 with 13 observations missing; TV credibility, N 850.25 observations missing). bOnly the extreme groups are compared here (newspapers: low, N = 213; high, N = 275; TV: low, N 232; high, N 256). reports; less than half of those with high local news, about half the respondents scores knew of such reports. chose television. Only one-third chose Some previous credibility research has newspapers. Preferences for a single source suggested that geographic scope of news is of state news and as the most reliable an important variable. For example, some source for local and state news were sim- research has shown that people in some ilar. Even though credibility scores were communities would believe newspapers related to preferences, they did not make a more often than television if topics in- great deal of difference. About 5 in 10 volved are local, but that they would be people with high newspaper credibility more likely to believe television if topics scores chose television over newspapers in are national. these comparisons. Local News. Newspapers did better National News. When respondents were than television in only one of the following asked which one of the news media they comparisons. About half chose newspa- would trust to help them understand pers when asked which one of their local national news, 57% chose television com- news media they would trust the most to pared to 29% for newspapers (Table 2). help them understand “a situation in your About 7 in 10 chose television as the most local area that is hard to understand or is reliable medium for national and interna- controversial” (Table 2). Just under 4 in 10 tional news and as the one source they chose television. People who gave high would choose for national and interna- ratings to newspaper credibility were par- tional news. Credibility ratings were re- ticularly likely to trust newspapers for lated to choices, but did not make much helping them understand complex local difference. For example, 68% of respon- issues. dents with high newspaper credibility However, when the comparison was the scores said that television is more reliable single source people would choose for for national and international reporting.

Downloaded from jmq.sagepub.com by FELICIA GREENLEE BROWN on April 12, 2012 Measuring the Concept of Credibility 459 It is noteworthy that the credibility of TABLE 3 television increases as the focus moves Reliability of News Coveragea from local to state to national and interna- tional for each of these measures. For each topic please rate the reliabiliry of Overall, if respondents had to choose coverage by the daily newspaper you are the just one way of getting news and informa- most familiar with (on a scale of “1’ to “5,” tion, 64% would rather get it on TV or where “I” means “not at all reliable” and “5” radio and 36% would rather read it in means (“very reliable”). (Question repeated print. Even a majority of frequent news- for television news.) paper readers preferred broadcast. Giving Pearson correlations with:b high credibility scores to newspapers did Newspaper Television not have much influence on preferences Credibility Credibility for newspapers or magazines, but giving Score score high scores to television did affect prefer- ences for television or radio. Even among Local news .34 .31 people who rated newspaper credibility Election coverage .36 .30 high, 59% preferred broadcast media. This Natural disasters .3 I .33 is almost as high a proportion as those rating newspaper credibility low who pre- Crime .36 .46 ferred broadcast (64%). Entertainment news .30 .26 Coverage of Specific Topics Survey respondents gave daily newspa- President Reagan .36 .38 pers and television high marks for their Health and reliability overall, the quality of their medical news .35 .34 reporting and their coverage of controver- Business news .39 .4 I sial issues. However, evaluations differed among specific topics. The economy .40 .43 Respondents rated the reliability of The government newspaper and television coverage of 17 in Washington .45 .44 different topics on a scale of I-to-5 where Religious news .32 .38 I meant “not at all reliable” and 5 meant “very reliable.” Both media were rated The arms race .40 .46 highly (4 or 5) by a large majority (6 to 7 in The abortion 10) for coverage of local news, elections, issue .37 .42 natural disasters, crime and entertainment The Soviet Union .37 .39 news. Roughly 5 people in 10 evaluated televi- Israel .38 .35 sion and newspapers highly for reporting The Arab on the economy, President Reagan, the countries .33 .38 government in Washington, business news, religion news, and health and medical Latin America .39 .39 news. Both media were given relatively aSubsarnple: Those who participated in Part 2 lower marks for coverage of the Arab interview - newspaper credibility, N = 862 countries, Latin America, Israel, the Soviet with 13 observations missing; TV credibility, N Union, and the abortion issue. Roughly 4 = 850, 25 observations missing). in 10 rated reporting of these topics b~~~ correlations are significant at p < .OOOI. highly. Although respondents tended to evalu- ate newspapers and television similarly, advantage in coverage of natural disasters they gave newspapers an edge in local (P <_ .01). news (t-test on means significant at p <_ Respondents’ credibility scores were .OOOl), and they gave television a slight highly related to their ratings of television

Downloaded from jmq.sagepub.com by FELICIA GREENLEE BROWN on April 12, 2012 460 JOURNALISM QUARTERLY and newspaper coverage of these topics. More than two-thirds of people inter- Correlations between television and news- viewed agreed that a ’s right to paper credibility scores and evaluations protect confidentiality of sources is like ranged between .26 and .46, all of which the right of confidentiality extended to were statistically significant at p 2 .0001 doctors or members of the clergy. News- (Table 3). paper and television credibility scores were related to holding this view. Responsible Exercise of Rights When the issue was the rights of ordi- Respondents appeared to be the most nary people versus the rights of the press, supportive of press freedoms when the people surveyed were much less likely to press was seen as a watchdog over favor the press. Seven in 10 said that a government. They took the strongest posi- person’s right to a fair trial outweighed the tions on limiting freedom of the press public’s right to be informed. Further, 7 in when the press was seen as an institution 10 agreed with a more strongly worded which tramples on other people’s rights statement: “The news media often make and takes advantage of the ordinary citi- people accused of crimes look guilty zen. A substantial minority, 4 in 10 before they are tried in court.” respondents, showed a desire for curbs on People giving high credibility scores to freedom of the press. This minority dis- television and newspapers were almost as agreed that the media exercise their right likely as those giving low scores to support to freedom of expression responsibly and the right to a fair trial over the public’s that it is important to maintain a free right to be informed and to agree that the press. A similar number agreed that the news media often make people accused of government has a right to demand access crimes look guilty before trial. to all of the files and notes of a reporter, whether the information has been made Summary and Conclusions public or not. Newspaper credibility ratings were neg- This article highlighted research ques- atively related to agreement with the tions in four areas: I) insights into dimen- statement that “sometimes, there’s too sions of credibility, 2) people’s ability to much freedom of the press,’’ but television differentiate well between news in news- credibility ratings were not. Both newspa- papers and news in other media, especially per and television credibility ratings were television, with regard to credibility, 3) negatively correlated with agreement with credibility problems in particular coverage the statement that “the media abuse their areas, and 4) the relationship between Constitutional guarantee of a free press” credibility and attitudes toward First (Table 4). However, there was no relation- Amendment issues. ship between credibility scores and atti- Dimensions of Credibility: First, a fac- tudes toward a reporter’s right to maintain tor analysis showed that 12 items grouped confidentiality of information not yet together in a credibility factor (for exam- made public. ple, being fair, unbiased, trustworthy, The minority who took a critical view of complete, factual, and accurate). News- press rights was smallest when the issue paper and television credibility scores was whether or not the press oversteps were created, based on scores on the 12 bounds in its role as a watchdog over pub- items. (The Gannett Center asked these lic officials. One-fourth said that the press same questions for newspapers, replicat- often gets in the way so that public offi- ing evaluations in all but one. Respon- cials cannot do the job they were elected to dents in the two communities surveyed do. All the others subscribed to the view rated their newspapers lower on being fac- that the press helps keep public officials tual. The Times Mirror also replicated the honest. Television and newspaper credibil- morality finding for ”news organizations.”) ity ratings were negatively related to press Since the two scores were correlated, criticism. this suggests that many people’s attitudes

Downloaded from jmq.sagepub.com by FELICIA GREENLEE BROWN on April 12, 2012 Measuring the Concept of Credibility 46 I TABLE 4 Press Rights Versus Other Rightsa

Which of the following statements comes the closest to how you personally feel Newspaper Television Credibility Credibility score score 1. The media abuse their Lowb High Low High Constitutional guarantee of a free press. 57% 24% 53% 20% OR The media are usually careful to be responsible. 43% 76% 47% 80% 2. It’s important to have a free press even when the press acts irresponsibly. 54% 65% 56% 59% OR Sometimes, there’s too much freedom of the press. 46% 35% 44% 41% 3. The press helps keep public officials honest. 64% 86% 59% 85% OR The press often gets in the way so that public officials can’t do the job they were elected to do. 36% 14% 41% 15% 4. A reporter has the right to refuse to give the government information he or she has obtained but which has not been made public. 62% 65% 66% 64% OR If the government feels it needs such information, it has a right to demand access to all of the files and notes of a reporter, whether the information has been made public or not. 38% 35% 34% 36% 5. A journalist’s right to protect confidentiality of sources is like the right of confidentiality extended to doctors or members of the clergy. 60% 75% 61% 71% OR Confidentiality is more important for doctors or members of the clergy than it is forjournalists. 40% 25% 39% 29% 6. A person’s right to afair trial is more important than the public’s right to be informed. 74% 66% 73% 66% OR The public’s right to know is just as important as the right to a fair trial. 26% 34% 27% 34%

aSubsample: Those who participated in Part 2 interview - newspaper credibility, N 862 with 13 observations missing; TV credibility, N 850, 25 observations missing). the extreme groups are compared here (newspapers: low, N = 213; high, N 275; TV: low, N 232; high, N = 256).

toward the two media are similar. Credi- These sets of questions involved: 1) believ- bility scores were compared with responses ability of media when news accounts con- to four other sets of questions used as flict, 2) media preferences for understand- measures of credibility in past research. ing news, 3) media preferences for reliable

Downloaded from jmq.sagepub.com by FELICIA GREENLEE BROWN on April 12, 2012 462 JOURNALISM QUARTERLY coverage, and 4) news source preferences slightly better on local news and television for local, state, and national news. did slightly better on coverage of natural Differentiating Among Media: In every disasters. Credibility scores were highly comparison, high credibility scores were correlated with ratings of newspaper and related to choice of newspapers over other television coverage of all I7 topics. media. Similarly, high television credibil- First Amendment Issues: Respondents ity scores were related to selection of tele- were most likely to support press freedoms vision over other media. However, in when media serve as watchdogs over (every comparison except one, television government. They were least likely to sup- tdid better than other media, and prefer- port First Amendment rights when the lence for television increased as geographic rights of ordinary people are at stake. scope moved from local to state to nation- These results are supported by findings of al/ international news. Despite high news- the Gannett Center studies and the Times paper credibility scores, many respondents Mirror study. (The Los Angeles Times were more likely to choose television when Poll found lower that the press exer- forced to choose among media. The cises its privileges responsibly than the exceptional comparison was media prefer- ASNE study did.) ence for understanding a complex or con- troversial local topic. In that case, half the Although people’s stands on newspaper sample chose newspapers. and television credibility frequently were Particular Coverage Areas: Although related to their attitudes toward press respondents evaluated daily newspapers freedoms, this was not so with regard to and television highly for overall reliability, the rights of ordinary citizens to receive a reporting quality, and coverage of contro- fair trial. People’s strong support of the versial issues, evaluations varied with press when it needs to keep information regard to 17 topics. Both media received and sources confidential also did not similar marks except that newspapers did depend on credibility ratings.

Third Class Mail Does Well Low-cost third-class mail sorted by postal carrier route is delivered nearly as quickly as more expensive first-class mail nationwide, accord- ing to research conducted for the American Newspaper Publishers Association (ANPA). The results were obtained from a five-city test completed for the ANPA in Wilmington, Del.; Oakland, Calif.; Sherman, Texas; Law- rence, Kan., and Madison, Wis. by Ruder Finn & Rotman, an indepen- dent communications and research firm. Nationally, it took two days after mailing to achieve delivery of 97.5% of the first-class mail, compared to three days for delivery of 97% of third-class mail sorted by carrier route. The first-class envelopes required nearly seven times the postage as the identical materials in the third-class envelopes. “Third-class rates do not reflect fairly the cost of delivering third-class presorted mail. A first-class letter weighing three ounces sent by a pri- vate citizen or a business costs 56 cents for virtually the same service that costs a direct-mail firm just over eight cents to send third- class,” said Sandra C. Hardy, chairman of ANPA’s Postal Committee. Hardy is vice president, Calkins Newspapers, Levittown, Pa. “The Postal Service should be congratulated for its performance in these tests. It got the mail delivered on time in nearly every case. Our question is why it chooses to underprice its good work in third-class mail,” Hardy said.

Downloaded from jmq.sagepub.com by FELICIA GREENLEE BROWN on April 12, 2012