Negative Articles Predict Clinical Trial Reluctance

Negative Articles Predict Clinical Trial Reluctance

Len-Rios, Maria E

Findings in this study on clinical trials show that front page and section stories are overwhelmingly negative and that newspaper exposure is a negative predictor of the intention to participate.

Merriam-Webster’s Medline Plus defines a clinical trial as

a scientifically controlled study of the safety and effectiveness of a therapeutic agent (as a drug or vaccine) using consenting human subjects.1

From biomedical research to animal trials to human clinical trials, medical research is needed to discover new treatments and cures.

However, because there are risks, costs and uncertainties associated with participation, few eligible adult patients enter clinical trials. Even though 71 percent of child cancer patients participate in clinical trials, just 3 percent of adults do.2 Parents appear to go to great lengths to find cures for their children, yet the same adults are far less likely to participate themselves. Estimates show that 5 percent of those who initially respond to calls to join clinical trials actually complete them.3 Scientists believe clinical trials are essential to find medical treatments, increase the generalizability of findings and shorten the time from discovery to treatment availability.4 Industry research by Harris Interactive suggests that while the public widely supports the idea of clinical trials, public support does not necessarily translate to clinical trial participation.5

This study examines the unique role of newspapers as purveyors of health information about clinical trials. First, the study will examine use of national newspapers and attitudes toward clinical trial participation. Then the paper will analyze the content of newspaper stories about clinical trials to see whether newspaper reader attitudes toward clinical trials may be influenced by their presentation in national newspaper stories. From a theoretical perspective, the study has implications for second-level agenda setting theory, which predicts that the attributes of news coverage influence the public’s interpretation of issues.6 Previous research on reporting about the economy supports the notion that negative news coverage can negatively affect the public’s estimation of future events.7 From a practical perspective, the study informs health reporters, consumers of health news and health communicators about the possible effects news stories may have on public attitudes toward medical research.

Health News and Public Opinion

In the United States, many people gain health information from the news media. A recent study revealed that about 80 percent of U.S. Internet users have sought health information online. The two most popular health topic categories searched were “disease or medical problem” (66 percent) or “certain medical procedure” (51 percent). The study also reported that from 2002 to 2005, searches for information on experimental treatments and medicines significantly increased from 18 to 23 percent.8 Similarly, other research found that 42 percent of U.S. adults closely tracked health news stories and one-fifth to one-third pay close attention to medical research news.9 These data show the public is interested in health news and news related to clinical research.

Clinical Trials and Public Opinion

Little academic research in the communication literature examines news content and public attitudes toward clinical trials. However, an article in Columbia Journalism Review, a trade publication for journalists, suggests clinical trial news stories are either overly positive or negative.10 Three medical academic studies address public opinion of clinical trials, mostly cancer trials. Cassileth et al., interviewed members of the general public, cardiology patients and cancer patients and found that nearly three-quarters felt serving as a clinical trial participant was very important. The strongest reason for clinical trial participation was to “get the best medical care”.11 Harris Interactive’s May/ April 2000 nationwide survey revealed that those who chose not to participate in clinical trials worried about risks and side effects, receiving a placebo, getting poorer health care and paying more for fear insurance would not cover treatment.12 Other factors that can negatively affect participation in clinical trials are: negative news coverage about trials, quality of life issues, eligibility, informed consent procedures, lack of awareness, low rates of physician referrals, costs and family objections to the trial.13

Conversely, additional analysis of the Harris 2000 survey data by academic researchers revealed those who were younger and had greater knowledge of clinical trials had more favorable attitudes toward clinical trials.14

Pentz et al., studied the tie between news coverage and clinical trial participation. They surveyed 100 patients referred to a clinical trial for a cancerous tumor treatment. Forty-seven percent of patients heard about the clinical trial first from the mass media-18 percent from newspapers and 13 percent from TV news-and 65 percent said they felt the news coverage was positive. Pentz et al., analyzed newspaper stories of the treatment and found that 59 percent were positive, with 16 percent negative and 25 percent neutral. While the authors did not analyze the specific news content read by participants, the authors believed that the tone of coverage was correlated with the patient’s reasons for participating and with increased knowledge about the trial’s purpose.15 Mans and Stream correlated the tone of news coverage with the number of clinical trial volunteers at a medical research center and concluded that the positive news coverage was associated with volunteerism. However, their study did not report looking at those who learned about the clinical trials and chose not to volunteer, so it is difficult to draw definitive conclusions.16

In summary, research results suggest that negative news about clinical trial research may decrease the likelihood that people will participate in clinical trials. Conversely, it appears that positive news coverage about clinical trial studies may increase volunteerism for clinical trials.

Agenda Setting and Second-Level Agenda Setting

The study by Mans and Stream suggested that the affective attributes, or tone, of newspaper stories affect whether people will take part in clinical trials.17 The study draws a link between news about clinical trials and how the public perceives clinical trials. Agenda-setting theory conceptualizes this relationship in a general manner. First introduced by McCombs and Shaw, the theory links the amount and prominence of coverage the media give an issue to the public’s perception of the issue’s importance.18 An extension of agenda-setting theory is second-level agenda setting, a secondary process whereby media highlight certain attributes of an issue that then influence not only how important the public deems an issue, but also the public’s interpretation of the issue.19

In this study, we are especially interested in the concepts of story prominence and story tone, the latter of which is associated with second-level agenda setting research. Both of these concepts have been conceptualized by Kiousis as elements of story salience.20 The idea is that stories with greater salience on the media’s agenda will have higher salience on the public agenda. In explicating the concept of salience, Kiousis suggests that visibility and valence comprise salience. Visibility is defined by the volume of stories about an issue compared with other issues and by the location of stories in the newspaper (i.e., front page, section front). Also, an article in an elite newspaper would be considered more visible than one in a newspaper with less prestige. Valence refers to whether the stories are positive or negative in tone. Kiousis argues that “affective elements in news can also enhance or reduce the overall salience of objects.”21 In summary, the underlying assumption is that stories that are more visible (i.e., on the front page or section front page) will have more effect on readers. In addition stories that have an affective impact on the reader (i.e., stories that are positively or negative valenced) will draw more reader attention.

Studies of second-level agenda setting are largely unexplored in the topic of health news.22 However, many studies have found support for second-level agenda setting in political news, including one conducted by, for example, McCombs and colleagues, which found that news coverage of both substantive (issue) and affective (tone) attributes during a Spanish election were correlated with public attitudes toward Spanish political candidates.23 The strongest effects were for affective attributes, or tone. Kleinnijenhuis, van Hoof and Oegema found that the negativity of news coverage of a Dutch presidential campaign was associated with increased voter distrust in politicians. This study also discovered that the effects of negative news persisted over time influencing voter behavior.24 Carroll and McCombs, in their proposal for research on corporate reputation, suggest that the more positive news coverage a company receives, the more favorably people will view the company. Likewise, they predict negative news will result in low opinions.25 Wanta, Golan and Lee studied international TV news and found that the more negative the news about a nation, the more the public disapproved of the nation when surveyed. However, unlike previous research, positive coverage of nations did not lead to positive evaluations of nations.26 These studies show that the tone of the news, and particularly negative news, appears to be associated with public opinion and, in some cases, behavior.

Hypotheses and Research Question

The literature strongly suggests that affective attributes, or the tone of news coverage, are associated with how the public will perceive the topic or object of study. As research has identified an association between newspaper coverage and attitudes toward participation in clinical trials,27 and second-level agenda setting theory suggests the tone of newspaper stories will be associated with public attitudes toward clinical trials, this study hypothesizes:

H1:

The greater the negative tone of national newspaper stories about clinical trials, the weaker the intention of individuals to participate in clinical trials. If the newspaper coverage is positive, there will be a greater intention of individuals to participate in clinical trials, or no effect.

Tone is an affective attribute in second-level agenda setting. It may also be that the tone of stories that are given more prominence may have a larger effect on public perception. In other words, a positive story on Page 1 may have more of an effect on perceptions than an inside story in section C of the paper. Prominence can be assessed by story location and length. Based on this, we hypothesize:

H2;

The tone (positive or negative) of the clinical trial stories on the front page and section front pages will be associated with intentions to participate in clinical trials more so than the tone of the clinical trial stories on the inside pages.

Lastly, different story themes may have different levels of salience with the audience and may also be presented in a particular predominant tone. To determine whether certain clinical trial story themes receive a certain tone of coverage that may influence reader attitudes, the following research question is presented:

RQl:

Is the theme of the clinical trial story associated with a specific story tone (positive or negative)?

Method

The study is concerned with how national newspaper exposure and the tone of news content may be associated with intentions to enter clinical trials. Therefore, we used content analysis and a telephone survey. The phone survey gauged whether national newspaper exposure was associated with reader intentions to participate in clinical trials. The content analysis examined the overall tone and of clinical trial stories. Content analysis systematically analyzes the content of media messages and thus sheds light on how the media cover issues.28 This study focuses on newspaper coverage of clinical trials because of newspapers’ capacity to provide better context and perspective than a typical television news broadcast.29

Telephone Survey

Sample. A random-digit dialed telephone survey was administered by a professional research center at the University of Kansas. The survey was completed with 432 adults in the Kansas City Metropolitan area. The response rate was 41 percent. The center fielded the survey from June 11,2003 to July 3, 2003. The questions on clinical trials were part of a larger survey. The multiple regression analysis is based on N=381 cases.

Independent variables. Past research has shown that demographic variables, knowledge, prior experience and trust in the medical profession are often times predictors of participation in medical studies. Therefore, these variables are included as predictors to determine if newspaper exposure accounts for variance in predicting participation.

The demographic variables measured are age, gender (0=male; 1=female), ethnicity (0=white; 1=nonwhite) and education. Other predictors are knowledge of clinical trials (additive index of 5 questions),30 past participation in a clinical trial (1=yes; 0=no), participation of a relative or close friend in a clinical trial (1=yes; 0=no), perceived importance of clinical trials (1=not at all important; 4=extremely important), reliance on physicians for health information (1=very unlikely; 7=very likely), reliance on family for health information (1=very unlikely; 7=very likely) and perceptions of whether patients are informed of the risks associated with participation in clinical trials (1=not at all confident; 6=very confident). The main variable of interest as a predictor, or independent variable, is exposure to national newspapers. This was measured by the question, “About how many days a week do you read a national newspaper?” Response categories were from “0” to “7” days.

Dependent variable. Intention to participate in clinical trials was determined by the level of agreement to the question, “I intend to participate in a clinical trial sometime in the future.” Response categories were: l=strongly disagree to 5=strongly agree.

Content Analysis

Sample. Stories from two national newspapers were selected for the studyThe Washington Post and The New York Times. Both papers are among the nation’s elite papers and serve as large circulation national newspapers.31 In addition to ranking among the five top U.S. newspapers in circulation, McCombs points out that both newspaper Web sites are among the top five most visited U.S. newspaper Web sites.32 Specifically, The New York Times is the “principal newspaper of record” in the United States and The Washington Post is “one of the nation’s most highly regarded daily newspapers.”33 Although these papers may be atypical in size and breadth from other newspapers, researchers have argued that The New York Times, in particular, “is considered by many to be the national newspaper of record” and that it serves an important agenda-setting function for other media.34 Research also suggests news content is largely homogenous and that there is “a considerable inter-media agenda-setting role for newspapers.”35 In other words, other media look to these premiere newspapers when choosing story ideas. Therefore, their news content represents that which is considered national news.

Stories were obtained in full-text format from the Lexis-Nexis database. We coded news stories containing the term “clinical trial” from the two papers from May 1, 2002, to June 11, 2003, prior to the telephone survey. The time frame establishes a potential causal relationship between media content and public perception. Selecting a year’s worth of national news about clinical trials is based on the premise that exposure to clinical trial news through time leads to an overall “worldview.” Relevant news stories, briefs, editorial and special section articles were included for analysis while letters to the editors, book reviews, calendar and magazine stories were excluded. In total, 340 stories containing “clinical trial” messages were coded, 216 from The New York Times and 124 from the The Washington Post.

Coding categories. The variables analyzed were story prominence, story type, theme of story, tone of story headline and overall story tone:

* Story prominence was coded in two ways-by where the stories appeared-or location-and by story length. Story location was defined as whether the stories were on the front page (section “A” or section “I” stories), a section front page story (front page of section other than section “A” or section “1”) or as an inside story (on a section inside page). Story length was measured by word count. Story length was further categorized on a scale from 1 to 6, where: l=stories of 1 to 500 words; 2=stories of 501-1,000 words; 3=stories of 1,001 to 1,500 words; 4=stories of 1,501-3,000 words; and 5=stories of 3,001 words or more.

* Story type was categorized as hard news, feature, brief, editorial or special section.

* Theme of article was coded as economic prospect, policy/regulations, ethical advocacy, safety and risks, development and health benefits, public opinion or other. Economic prospect stories were about the economic potential of a drug or treatment and the economic impact of its research and development. Policy/regulation stories dealt with government policies and regulations of drugs and therapies, their development and applications. This included stories about FDA approval. Ethical/advocacy articles focused on calls for ethical principles in medical research, public control and involvement in the clinical trial process and advocacy of patient protections. Articles about safety and risks addressed risks associated with participation in specific clinical trials. Development and health benefits articles centered on the medical progress achieved through a new therapy and its potential health benefits to patients. Stories on public opinion dealt with the attitudes and opinions of the general public about clinical trials through public polls, surveys, etc. Stories that did not fit the six categories were classified as “Other.”

* Headline tone was coded as positive, negative or neutral. Coders assumed the role of a newspaper reader and determined the headline tone before reading the actual story text. Positive headlines referred to a favorable outcome or desired condition. Positive headlines might indicate a drug or medicine received FDA approval or a treatment was proven effective. An example is “Genvec says Treatment Shrinks Tumors.” Negative headlines featured unfavorable outcomes or undesirable conditions. Negative headlines indicated problems with clinical trials, side effects, patient complaints or setbacks in clinical trial research. An example is “Battered Biotechs can’t get past paralysis of FDA.” Neutral headlines were either strictly factual in nature or contained both favorable and unfavorable statements. An example of a neutral headline is “Specter asks Bush to permit more embryonic cell lines.”

* Overall story tone was similarly coded as positive, negative or neutral. It was determined by classifying each sentence about clinical trials in the story as positive, negative or neutral. The sum for each of the three tone categories was calculated and the tone that predominated was selected as the overall tone. In rare cases when one type of tone did not prevail, the coders chose the tone associated with the headline tone.36

Unit of analysis and intercoder agreement. The story and the clinical trial messages were the units of analysis. Clinical trial messages were defined as sentences within the stories containing statements pertaining to clinical trial research and its processes or outcomes. Two coders coded the variables. Clinical trial messages were first determined in two sessions where the coders talked through the stories and agreed on which sentences fit the definition of a clinical trial message. A random sample of 10 percent of the stories was then selected for intercoder agreement prior to the study. Coders independently coded the variables of interest. There was an 89 percent overall agreement. Intercoder agreement using Cohen’s Kappa for nominal variables was: story prominence = 1.0; story type = .93; theme of article = .87; tone of story headline = .83; and overall tone of story = .84.

Hierarchical multiple regression analysis was used to determine whether national newspaper exposure was associated positively or negatively with reader intentions to participate in clinical trial research. Variables from the survey were entered in four blocks. The first block of variables included the demographics: age, gender, ethnicity and education. The second block comprised the variables of knowledge of clinical trials, past participation in clinical trials, participation of a close friend or relative in a clinical trial and perceived importance of clinical trials. The third block contained the health information reliance variables-reliance on physicians and nurses for health information and reliance of family and friends for health information. Also included was a variable that measures the perceptions of whether participants are informed about the risks of clinical trial participation. The last block held the main variable of interest-national newspaper exposure. Regression analysis findings were then compared with content analysis results. Simple descriptive statistics and chi-square analysis were used to examine news content variables. The level of significance for analyses was p=.05.

Results

Survey Analysis

To assess how national newspaper exposure was associated with clinical trial participation, a hierarchical multiple regression analysis was performed.37 Control variables positively associated with clinical trial participation were: past participation in clinical trials (≤ = .19, p

All the variables included in the analysis accounted for R^sup 2^=. 13, or explained 13 percent of the variance that accounts for why people choose to participate in clinical research. The results suggest that increased exposure to national newspapers is associated with lowered intentions to participate in clinical trials. In order to answer our hypotheses, we examined the content of national newspapers.

Content Analysis

Before analyses of the hypotheses are presented, a general description of the clinical trial stories is provided for context. The average story length was 928 words. The first clinical trial message appeared on average in the sixth paragraph (median was third paragraph). Of the total of 340 stories, 72 percent were inside page stories, 20 percent were section front-page stories and 8 percent were front-page stories. For story type, 48 percent were hard news, 34 percent features, 7 percent special section stories, 6 percent news briefs and 5 percent editorials. When examined by story theme, 32 percent were on the development and health benefits, 21 percent on policy/regulations, 16 percent on economic prospects, 14 percent on safety and risks, 7 percent on ethical and advocacy issues, 1 percent on public opinion and 9 percent on other issues.

Hypothesis 1

The first hypothesis predicted that the greater the negativity of national news stories, the lowered intention respondents would have for participating in clinical trials. Conversely, if newspaper stories were positive, we expected an increased willingness to participate or no effect. Survey results showed that national newspaper exposure was negatively associated with clinical trial participation. Analysis of the story tone of news content revealed that overall 124 stories were neutral in tone (37 percent), 117 were negative (34 percent) and 99 were positive (29 percent). For story headlines, 127 were neutral (37 percent), 116 were negative (34 percent) and 97 were positive (29 percent). If we just examine messages that are valenced positively or negatively, as those stories are more likely to have increased salience to readers,38 there are more negative stories than positive stories. These data appear to support H1.

Hypothesis 2

Story prominence may be a variable that has an additional effect on the perception of clinical trial news. Agenda setting theory asserts that the more prominent news stories are on the news agenda, the more importance the public will attach to the issue. In terms of second-level agenda setting, the tone of prominent stories may impact reader attitudes more than stories with less prominence. Prominence was measured by location and story length. Again, we are concerned with stories that have a valenced tone. First, we ran a cross-tabulation of location (Page 1, section front, inside page) by tone (positive, negative). There was a significant association between tone and location (N=216; X^sup 2^=6.46, d.f.=2, p=.05). Of newspaper Page 1 stories, 10 (77 percent) were negative and three were positive. Looking at section front page stories, 31 (65 percent) were negative and 17 (35 percent) were positive. Inside page stories were more balanced in tone, with 76 (49 percent) negative and 79 (51 percent) positive in tone. These data clearly illustrate that the more prominent the story location, the more likely it was to be negative.

Next, cross-tabulation of story length by location was run (Page 1, section front, inside page). There was a significant association (N=340; X^sup 2^=134.4, d.f .=8, p=.001). Of inside page stories, 72 percent were 1,000 words or less, whereas 89 percent of section front page stories were between 501 and 3,000 words, and 87 percent of front page stories were between 1,001 and 7,000 words. Our previous analysis indicates that stories with a negative tone were more often on the front page. Front page stories, in turn, were longer, thus reflecting an even higher level of prominence.

These data show that negative stories were given more prominence and therefore may be one factor that explains the negative association between national newspaper exposure and decreased intention to participate in clinical trial research.

Research Question 1

Lastly, the research question-are the themes of the news stories associated with a predominant story tone? A cross-tabulation of the two variables shows that there is a significant association between story theme and tone (N=340; X^sup 2^=67.30, d.f.=6, p

Discussion

The findings from our survey and content analysis suggest that there is second-level agenda setting in the case of clinical trial news stories. It is clear from the descriptive analysis that clinical trial news stories are not at the top of the national newspaper agenda, with only a small proportion of clinical trial stories gaining front-page placement (8 percent). So, from the perspective of first-level agenda setting, clinical trial stories are also probably not high on the public agenda. However, those stories that do make it to the front page, or section page fronts, are much more likely to be negative than positive. These data show that the prominence of clinical trial stories, indicated by story location and length, emphasizes negative stories. Several topics that gained prominent coverage were the cessation of 27 gene therapy studies after the death of a French child, the discovery of genetic links to breast cancer and the dangers of hormone replacement therapy. If negative clinical trial stories are given more highly visible locations in the paper and longer treatment, then readers are more likely to view these stories than inside page stories. In addition, readers are probably more likely to remember the negative stories because of their salience.

Our analysis of tone and story themes shows that development and health benefits, the largest theme category comprising 32 percent of stories, was the most positive. Yet, three categories-economic prospects (16 percent of stories), safety and risks (14 percent) and ethical/advocacy issues (7 percent)-were more often negative. There is no doubt that reporting on the safety and risks of clinical trials and ethical/advocacy issues is an important journalistic function. Journalists play a vital watchdog role of government and industry. However, a consequence may be more negative coverage and decreased participation in clinical trials, an important step in medical discovery. In such “watchdog” articles, it may be beneficial for reporters to indicate how clinical trial research as a whole has led to many of today’s lifesaving therapies even if a particular trial or course of treatment does not succeed. Of course, in the case of economic prospect-themed stories, it may not be easy or even practical to provide such context. Along with reporters offering readers greater context in medical research, medical and industry professionals may need to ensure reporters are informed on the positive developments of their clinical research.

From a theoretical perspective, our research provides support for second-level agenda setting and the assertion that affective attributes, or story tone, may have an influence on public opinion and behavior. Because of our methods, we cannot identify a clear causal direction. This connection is difficult to demonstrate empirically. As McCombs has written, “the transfer of object and attribute salience from one agenda to another may not have an empirical answer per se.”39 In other words, measuring cognitive processing and associating it with the cumulative effects of exposure to news content is a weighty endeavor. Still, as McCombs notes, through replication in many contexts, we can achieve more confidence in the reliability of our conclusions. This study is a building block in that process.

Limitations and Future Directions

As with many studies, there are limitations to the geneneralizability of this study. It examined a regional survey sample of respondents and the content of two national newspapers, but it did not analyze the exact content read by these readers. However, if much of the media coverage of clinical trials is the same (e.g., intermedia agenda-setting), then it could be expected that much of the coverage would be similar. In fact, previous research suggests there is much homogeneity of content across print and electronic sources.40 The study’s sample is also skewed female (70 percent) and toward the more educated, which also hampers the generalizability of the findings. Even so, research does suggest that women are more likely to seek health information than men, making them a valuable audience for this study.41 In addition, this study looks at just one year’s worth of clinical trial stories, and examining other time periods may reveal different results.

This study also differs from studies done in the academic health literature cited in the review. Much of the research on accrual to clinical trials focuses solely on diagnosed patients and specific diseases. There is a difference between public attitudes toward clinical trial participation and intention to participate among those who are diagnosed with a disease and those who are healthy. Additionally, probing about participation for specific diseases may reveal different results as some diseases offer fewer treatment options than others. Some diagnosed individuals may have few choices but to participate in clinical research. For the healthy, the risks of participation may seem much greater and thus influence their intentions to participate.

There are many avenues for future study. One direction includes further exploration of the concept of story salience. Subsequent studies may diversify the measures of salience by including measures of visuals, such as photographs or infographies. Salience could also be measured by determining how personally relevant readers believe clinical trials stories are to them. In other words, personal salience may become an important variable.

Examining individual differences, such as the personal relevance of news stories might lead to comparing groups according to psychological constructs such as self-efficacy, need for cognition, level of health literacy, etc. Although this studydid not find significant effects for knowledge about clinical trial information on intention to participate in clinical trials (perhaps an artifact of how it was measured) the knowledge questions did reveal that a majority of respondents (53 percent) believed that all clinical trials involved placebos. In addition, the data show that those who believe clinical trial participants are truthfully informed about the risks of participation are more likely to intend to enroll in clinical trials. These pieces of data may indicate that personality variables, such as trust in medicine and risk tolerance, should be included. In sum, there are many news content variables and news reader characteristics that remain to be investigated.

Notes

1. Merriam-Webster Medline Plus Medical Dictionary, s.v. “clinical trial,” (27 August 2005).

2. Phyllis M. Stein, The Gale Encyclopedia of Cancer, s.v., “clinical trials,” ed. Ellen Thackery (Farmington Hills, Mich.: Gale Group, 2001).

3. Nancy S. Sung et al., “Central Challenges Facing the National Clinical Research Enterprise,” JAMA 289, no. 10 (2003): 1278-1287.

4. Peter M. Ellis et al., “Randomized Clinical Trials in Oncology: Understanding and Attitudes Predict Willingness to Participate,” Journal of Clinical Oncology 19, no. 15 (2001): 3554-3561.

5. “Harris survey highlights public’s ambivalence about clinical trials,” Health Care Strategic Management,

20 (May 2002): 11-12.

6. Salma Ghanem, “Filling in the Tapestry: The Second Level of Agenda Setting,” in Communication and Democracy: Exploring the Intellectual Frontiers in Agenda-Setting Theory, eds. Maxwell E. McCombs, Donald L. Shaw and David H. Weaver (Mahwah, N.J.: Lawrence Erlbaum, 1997), 3-14.

7. Joe Bob Hester and Rhonda Gibson, “The Economy and Second-level Agenda Setting: A Time Series Analysis of Economic News and Public Opinion about the Economy,” Journalism & Mass Communication Quarterly 80, no. 1 (2003): 73-90.

8. Susannah Fox, “Health Information Online: Eight in Ten Internet Users Have Looked for Health Information Online, with Increased Interest in Diet, Fitness, Drugs, Health Insurance, Experimental Treatments and Particular Doctors and Hospitals,” 17 May 2005, (27 August 2005).

9. Mollyann Brodie et al., “Health News and the American Public, 1996-2002” Journal of Health Politics, Policy and Law 28 (October 2003): 927-950.

10. Melinda Voss, “The Hazards of Covering Clinical Trials,” Columbia Journalism Review 40 (September/October 2001): 29.

11. Barrie R. Cassileth, Edward J. Lusk, David D. Miller and Shelley Hurwitz, “Attitudes Toward Clinical Trials Among Patients and the Public,” JAMA 248 (27 August 1982): 968-969.

12. Humphrey Taylor and Robert Leitman, “Misconceptions and Lack of Awareness Greatly Reduce Recruitment for Cancer Clinical Trials,” Harris Interactive Health Care News 1, no. 1 (2001): 1-3, accessed from (28 May 2003).

13. Carolyn Cook Gotay, “Accrual to Cancer Clinical Trials: Directions From the Research Literature, ” Social Science & Medicine, 33, no. 5 (1991 ): 569-577; Eva Grunfeld, Louise Zitzelsberger, Marjorie Corisrine and Faye Aspelund, “Barriers and Facilitators to Enrollment in Cancer Clinical Trials,” Cancer 95, no. 7 (1 October 2002): 1577-1583.

14. Robert L. Comis et al., “Public Attitudes Toward Participation in Cancer Clinical Trials,” Journal of Clinical Oncology 21, no. 5 (March 1, 2003): 830-835.

15. Rebecca D. Pentz et al., “Study of the Media’s Potential to Influence on Prospective Research Participants’ Understanding of and Motivations for Participation in a High-Profile Phase I Trial,” Journal of Clinical Oncology 20, no. 18 (15 September 2002): 3785-3791.

16. Gary Mans and Christopher Stream, “Relationship Between News Media Coverage of Medical Research and Academic Medical Centers and People Volunteering for Clinical Trials,” Public Relations Review 32, no. 2 (2006): 196-198.

17. Ibid.

18. Maxwell McCombs and Donald L. Shaw. “The agenda-setting function of the mass media,” Public Opinion Quarterly 36, no. 2 (1972): 176-187.

19. Salma Ghanem, “Filling in the Tapestry”; Maxwell McCombs, Juan Pablo Llamas, Esteban Lopez-Escobar and Federico Rey, “Candidate Images in Spanish Elections: Second-Level Agenda-Setting Effects,” Journalism & Mass Communication Quarterly 74, no. 4 (1997): 703-717.

20. Spiro Kiousis, “Explicating Media Salience: A Factor Analysis of New York Times Issue Coverage During the 2000 Presidential Election,” Journal of Communication 54, no. 1 (2004): 71-87.

21. Kiousis. “Explicating Media Salience,” 76.

22. Craig E. Carroll and Maxwell McCombs. “Agenda-setting Effects of Business News on the Public’s Images and Opinions about Major Corporations,” Corporate Reputation Review 6, no. 1 (2003): 36-46.

23. Spiro K. Kiousis, Philemon Bantimaroudis and Hyun Ban. “Candidate Image Attributes: Experiments on the Substantive Dimension of Second Level Agenda Setting,” Communication Research 26, no. 4 (1999): 414-428; McCombs et al., “Candidate Images”; June Wong Rhee, “Strategy and Issue Frames in Election Campaign Coverage: A Social Cognitive Account of Framing,” Journal of Communication 47, no. 3 (1997): 26-48.

24. Jan Kleinnijenhuis, Anita M.J. van Hoof and Dirk Oegema. “Negative News and the Sleeper Effect of Distrust,” The Harvard Journal of Press/Politics 11, no. 3 (2006): 86-104.

25. Carroll and McCombs, “Agenda-setting Effects of Business News,” 41.

26. Wayne Wanta, Guy Golan and Cheolhan Lee, “Agenda Setting and International News: Media Influence on Public Perceptions of Foreign Nations,” Journalism & Mass Communication Quarterly 81, no. 2 (2004): 364-377.

27. Pentz et al., “Study of the Media’s Potential to Influence.”

28. Paula M. Poindexter and Maxwell E. McCombs, Research in Mass Communication: A Practical Guide (New York: Bedford/St. Martin’s, 2000).

29. In fact, a recent study of five television news channels by Ganz and Wang found that 70 percent of television health news stories were shorter than 60 seconds in length. Walter Gantz and Zheng Wang, “Health Content in Local Television News” (paper presented at AEJMC, Toronto, Canada, August 2004).

30. Knowledge was a score ranging from 0 to 5, based on the correct responses to the following statements: 1) Clinical trials involve medical studies to test new vaccines, therapies and new known treatments. (T, 95% of respondents got it correct); 2) The federal government and pharmaceutical companies are the only organizations that can conduct clinical trials (F, 83% of respondents got it correct); 3) There are ethical review boards that oversee how clinical trials are conducted (T, 80% of respondents got it correct); 4) All clinical trials involve placebos and withholding treatment from one set of participants (F, 47% of respondents got it correct); 5) Most clinical trials are open to anyone who volunteers to participate (F, 71% of respondents got it correct).

31. John C. Merrill, “The Global Elites,” IPI Report 5, no. 4 (1999): 13-15.

32. Maxwell McCombs, “A Look at Agenda-setting: Past, present and future,” Journalism Studies 6, no. 4 (2005): 543-557.

33. John C. Merrill, Global Journalism: A Survey of the World’s Mass Media (New York: Longman, 1983), 310-311.

34. William L. Benoit, Kevin A. Stein and Glenn J. Hansen, “New York Times Coverage of Presidential Campaigns,” Journalism & Mass Communication Quarterly 82, no. 2 (2005): 356-376; Kiousis, “Explicating Media Salience,” 77; McCombs, “A Look at Agenda-setting,” 545.

35. Tony Atwater, Frederick Fico and Gary Pizante, “Reporting on the State Legislature: A case of Inter-media Agenda-Setting,” Newspaper Research Journal 8, no. 2 (1987): 53-61; McCombs, “A Look at Agenda-setting,” 545.

36. The Pearson correlation between story headline tone and story tone was (r = .46, p

37. The survey sample descriptives are as follows: gender (female, 70%); age (M = 49); ethnicity (80% Caucasian; 2.2% Latino; 11 % African-American; 1.5% Asian American; 0.7% American Indian; 2.2% multicultural; and 1.9% other); employment (47.3% full-time, 8.3% part-time, 8.5% homemakers or caregivers, 6.6% unemployed, 25% retired); and education (4.4% some high school; 23.3% high school graduate; 20% some college but no degree; 7.5% junior college or a two-year degree; 29.9% an undergraduate degree; and 14.6% a graduate degree). Our sample differs from the general population because it is slightly more female, more highly educated, less ethnically diverse and has more respondents who are retired. Twenty-three percent read national newspapers.

38. Kiousis argues that “to code the number of stories in print or broadcast media that have a positive or negative tone toward the object of a story, the higher the number the higher the valence score. In such situations, more positive stories would indicate higher salience in some instances whereas more negative stories would indicate higher salience in other cases;” Kiousis, “Explicating Media Salience,” 76.

39. McCombs, “A Look at Agenda-setting,” 542.

40. Atwater et al., “Reporting on the State Legislature” 54; McCombs, “A Look at Agenda-setting,” 545.

41. Susannah Fox and Deborah Fallows. Internet Health Resources, PewInternet.org, 16 July 2003, (20 July 2003).

Len-Rios is an assistant professor and Qiu is a doctoral candidate in the School of Journalism at the University of Missouri, Columbia. The survey in this study was made possible through a New Faculty Research Grant awarded to Len-Rios from the University of Kansas. The authors wish to thank Miranda Phillips for her aid in the content analysis.

Copyright Newspaper Research Journal, Department of Journalism, University of Memphis Winter 2007

Provided by ProQuest Information and Learning Company. All rights Reserved