News Coverage of Environmental Risks: Subjective Knowledge, Personal Efficacy and Perceived Usefulness of Different Media

By Brooke Weberling, Jennette Lovejoy and Daniel Riffe 

WJMCR 40 (February 2011)

Introduction | Previous Research | Theoretical Framework | Research Questions and Hypotheses | Method | Findings | Discussion and Conclusion


Telephone survey data (N=511) are used to explore relationships among exposure, attention, and usefulness of online, newspaper, and television environmental news; subjective knowledge; personal efficacy; and environmental risk. Correlations show exposure to online environmental news was related to personal efficacy while online attention to such news was related to subjective knowledge, but not personal efficacy or perceived risk.  Attention to television and newspaper environmental news were both related to subjective knowledge and perceived risk.  After controls (particularly education), many of these relationships disappeared, but both attention to and perceived usefulness of online environmental news were related to perception of being at environmental risk.


Environmental problems are sometimes considered “unobtrusive” or “impersonal risks,” because individuals may have little contact with the threats. In these instances, the media are often primary—or the only—sources of information.1 Indeed, studies have shown that individuals rely on local media as the principal sources of information about local environmental problems2 and that online sources are increasingly popular for seeking information about science,3 business,4 health issues and environmental risks,at least for those with Internet access (although the “digital divide”6has narrowed, a 2010 report showed that 50% of rural American adults lack home access).7  

Even when media increase public awareness of environmental problems, coverage does not always 
reflect who is responsible or what individuals can do to deal with them. For example, Kensicki’s study of pollution and other social problems8 found that news coverage rarely indicated a specific cause, effect, responsible party, or means to act to solve the problems.  This creates a “media-constructed disconnect”9 between people and problems, “an audience lost and then uncaring about an issue, since they do not understand where it came from, do not see any meaningful way to pitch in and help, and do not see their helping as useful or effective anyway because the problems lack solutions.”10 Kensicki’s11 conclusions echo Spencer and Triche,12 who noted that “newspaper reports defined relatively simple causal frameworks” that “tended to avoid defining social or technological events as causal factors.” These critiques reflect recent practices that are in implicit opposition to what Lasswell proposed to be the most basic societal functions of news media: surveillance of the larger environment, and interpretation or correlation of society’s response to events in the environment.13

Individuals seek information about environmental risks based on various factors, including the risk’s relevance and severity and the individual’s vulnerability, and belief in the efficacy of a preventive response.14 A risk that is familiar or understood, over which those at risk feel they have some control, is more acceptable than risks that are unknown, undetectable, or controlled by others.15   Those who have access to information about environmental risks arguably believe they are more knowledgeable and in control than those who may not have access to such information. Beyond access to information, factors such as attention and perceived knowledge may affect how people respond (if at all) to environmental news coverage. The purpose of this study is to explore the relationships between attention to such information and other factors, including subjective knowledge, a sense of personal efficacy, and perceptions of local environmental risks.

Local environmental problems tend to be more imminent and threatening, which may create a heightened sense of risk and lead to increased—and more “discriminating”—information seeking and use of local media.16 Indeed, attention to local newspaper and television coverage of environmental problems has been positively related to perceived severity of nearby risks, and greater perceived risk has been related to more critical evaluation of the coverage.17 Presumably, then, these perceptions about the usefulness of local sources of environmental news are likely linked to individuals’ levels of “subjective knowledge.” 

Unlike objective knowledge, subjective knowledge is “what we think we know” about a topic.18 Also called “the illusion of knowing” and “self-perceived knowledge,”19 subjective knowledge is widely used in studies of consumer behavior and health communication, and in studies of public understanding of environmental issues such as global warming.20 Subjective knowledge is similar to “information sufficiency,” which is a measure of whether an individual believes s/he has enough information to cope with a risk.21 However, while information (in)sufficiency has been proposed as a predictor of information seeking,22 subjective knowledge level may be a consequence of seeking, exposure, and attention.  Indeed, Salwen and Driscoll,23 building upon Graber’s “assurance” function of mass media,24 demonstrated that “assuredness”—feeling “as informed as needed”—was correlated with television news viewing.  Park later found a similar role for news viewing in respondents’ “illusion of knowing.”25

Channel beliefs26 about the usefulness of various media for environmental news may also be related to people’s subjective knowledge and to another subjective state, “personal efficacy,” the sense of being able to take action. Using a telephone survey of a Midwestern state (n = 511), this study explored these variables, with the purpose of answering the following question: How are individuals’ use and evaluations of online, newspaper, and television environmental news coverage related to subjective knowledge and personal efficacy?

Media and the Environment

Since the 1970s, studies of environmental news coverage have documented public use of the media for information about the environment27 and content analyses show that the amount of environmental issue news coverage corresponds closely with level of public issue salience.28 Other studies have shown that attention to environmental news coverage is positively associated with perceived environmental risk.29 More recently, researchers30demonstrated that exposure to a greater variety of televised environmental information content is associated with greater concern about environmental risk. Good31 found a similar relationship of viewing and environmental worldview among environmentalists.

When it comes to sources of environmental information, surveys show that most people identify mass media as their main sources.32 Of course, news coverage itself has always been influenced by many forces, including media professionals’ routines, their attitudes and ideological positions, outside institutions, economics, and trends.33 And while there are more choices of media now, television, newspapers and the Internet remain popular sources for health, science and environmental news. Local newspapers rank higher than local television in providing more complexinformation, such as costs and solutions for environmental problems,34 while recent research has shown increasing reliance on the Internet for environmental, health and science news.35 More than 50% of people reported going online at least once a week for health information,36 while 22% of Internet users search the Web for information on environmental health hazards.37

How does the Web complement or supplant other media use for environmental news?  Researchers in Ohio38 conducted a statewide survey about environmental news sources across global, national, state, city, and community levels.  A majority of participants cited television news most frequently, but only for problems at global (64%), national (61%), and state (58%) levels. Newspapers were used most for problems at city (61%) and community (57%) levels, while 19% of respondents identified the Internet as their environmental news source for national or international problems.   A national replication found similar results,39 but slightly fewer respondents named the Internet for national or international environmental news .

Evaluating Media Environmental Coverage: A Framework

Whatever the environmental problem or source of information, news coverage is often criticized for failing to provide causes, context and “mobilizing information” that empower citizens to act.40 Spencer and Triche41 observed that descriptions and frames used in media coverage are “highly selective.” Wiebe42 lamented nearly four decades ago that the public may develop knowledge of environmental problems but lack a vision of how to solve them, creating a sense of “well-informed futility” among the general public. In other words, some media coverage is simply less useful to people than other coverage because it does not discuss likelihood that problems can be solved and promotes apathy by failing to connect problems with nonprofit or government organizations, and potential individual action.43   Foust and Murphy44 analyzed press coverage of global warming, noting that its “apocalyptic framing” depicts a process “outside the purview of human agency.” A sense of subjective knowledge and personal efficacy, which could be strengthened by information provided by the media if it is useful45 and told in a “coherent” way,46 may be the missing link between news coverage of environmental problems and an individual’s perceived ability to act regarding environmental risks. 

One way of thinking about news media usefulness is through the lens of what Lemert called “mobilizing information.”47    In brief, mobilizing information promotes more than awareness by presenting the public with the means to act related to an issue or problem.  For example, Lemert et al.48 stated that news media may tell how, whether, and where to express attitudes to influence political processes. In news coverage, mobilizing information might consist of names, phone numbers, addresses, Web links, times and dates of meetings, titles of documents and more. Lemert et al.49found that 36% of newspaper items contained at least some type of mobilizing information, but it occurred most often in “positive” stories, rather than those deemed negative or controversial. Local stories were more likely (about half) to provide mobilizing information than non-local items (about a fourth).

Mobilizing information in media content continues to be analyzed.  For example, one study indicated that online newspapers did not offer significantly more mobilizing information than their print counterparts.50 A study of Ohio dailies and weeklies found that three-fifths of stories on cancer screening contained specific mobilizing information.51 A study of disaster coverage found that 24-hour news channels provided more mobilizing information on making donations to relief efforts than newspapers, with Web addresses being the most common form provided by all media.52 Nicodemus53 reported that one newspaper’s language on a controversial waste facility “fostered the formation of community by overcoming apathy and encouraging residents to act collectively,” while another’s suppressed it “by reinforcing the belief that residents were powerless against the entrenched economic and political power base.”  Murch54 found that local newspapers focused more on national pollution problems than local problems and excluded information that would enable the audience to act. Rubin and Sachs55 studied media environmental coverage in the 1970s, and while they did not use the term “mobilizing information,” they referred to “specific information, and the day-to-day alerting which is necessary for any citizen to participate effectively” in the public arena.56

Of course, mobilizing information provides only a normative framework for our thinking about how news media—performing Lasswell’s surveillance function—might serve the public, and many studies of mobilizing information have been via content analysis.  While we do not assess mobilizing information in media content here, we do focus on people’s beliefs or perceptions about how well different media provide information about environmental concerns.  Our argument is simply that if information from a particular source is judged to be useful among the general public, it may contribute to the study’s focal variables: subjective knowledge and sense of personal efficacy regarding environmental risks.

Subjective Knowledge and Personal Efficacy

People who are knowledgeable about an issue (or perceive themselves to be) may be more likely to attend to and act on news and information related to that issue.57 Subjective knowledge, which refers to people’s perceptions about their understanding of a topic, differs from objective knowledge, which reflects actual accurate information that people possess. While objective knowledge is generally dependent upon ability and expertise, subjective knowledge may be based on experience and factors such as self-confidence and other self-beliefs.  Some scholars have noted that objective and subjective knowledge levels may not match (more charitably, they may be described as “miscalibrated”).58   Because of the importance of what we “think we know,” subjective knowledge is used widely in research on consumer behavior, health communication and other areas that associate information or knowledge with subsequent attitudes, beliefs or behaviors. A recent meta-analysis59 reported that relationships between subjective and objective knowledge are stronger for assessing knowledge about products, such as cars, computers, cameras, etc.,60 than research assessing “nonproduct” issues, such as ecological behaviors.61

Nonetheless, measures of knowledge have been used in studies on health and environmental issues, including cancer and global warming. Kahlor and Rosenthal62 found that the number of media sources used, information seeking effort, and education levels were relatively strong predictors of knowledge about global warming. Nabi, Roskos-Ewoldsen and Carpentier63 experimented with messages about cancer and found that participants with high subjective knowledge were more persuaded by messages based on self-efficacy or empowerment than those with low subjective knowledge. These results seem to indicate a relationship between subjective knowledge and personal efficacy, such that “what you think you know” may be one of the most important factors when it comes to understanding perceived risk.64

Subjective knowledge is similar to the concept of “information sufficiency,” which has been used widely in environment and risk communication research. Griffin and colleagues65 used the term when they introduced the Risk Information Seeking and Processing (RISP) model and argued that the size of the “gap” between information known and information needed determines an individual’s information seeking and processing related to risk. They defined information sufficiency as “the amount of information people say they need to deal adequately with a given risk in their own lives.”66 While the current study did not measure information sufficiency explicitly, the variables involved in this study are loosely based on the RISP model and the work of Griffin and other colleagues.

Using path analysis, Griffin and colleagues67 examined predictors of information sufficiency, conceptualizing it as an antecedent to information seeking-behaviors. Kahlor et al.68 used hierarchical regression and found information insufficiency was a significant predictor of “active seeking” of information about an impersonal risk. Later research examined information sufficiency as a consequence of information seeking, media use and information acquisition.69 However, that research did not actually measure perception of a “gap” between information known and needed as conceptualized by Griffin et al.70 Instead, the researchers were likely measuring subjective knowledge, a reasonably clear consequence of media use and channel beliefs about usefulness of information in one’s “repertoire of sources.”71   One’s subjective knowledge alone—whatever the contribution of news coverage to it—does not mean an individual will take action regarding a risk. People also have to have a sense of empowerment or efficacy—a sense that their actions will be effective.

Internal or personal efficacy, “the conviction that one can successfully execute the behavior required to produce the outcomes,”72 has been explored in numerous health communication contexts and employed with many theoretical models (e.g., theory of planned behavior, health belief model, social cognitive theory, etc.) in research on personal choice and follow-through on behavior change, such as quitting smoking or exercising regularly.73   Griffin, et al.74 included “personal control” in their RISP model because they realized that, beyond information sufficiency, individuals’ sense of efficacy influences their response to risks.

In political communication, on the other hand, efficacy is “citizens’ perceptions of powerfulness (or powerlessness) in the political realm,” and may involve perceptions of access to, understanding of, and trust in public/civic arenas and processes.75   Morrell distinguished internal efficacy as a personal ability to understand an issue and participate effectively, and external efficacy as an individual’s perceptions of the responsiveness of entities to citizens’ demands.

Similarly, Newhagen76 examined socio-economic status and political efficacy, using self- and system-efficacy to distinguish between beliefs that one can protect oneself, and trust or confidence in existing political systems and institutions. He found that self-efficacy tended to increase with media use while system-efficacy decreased (with newspaper reading, in particular: the more people read newspapers the less they trust political systems and institutions). He also found differences based on race, media use frequency and type (newspaper, television, and talk radio). However, his study did not include Web or online news or perceptions of media usefulness.

Other studies have linked—either empirically or conceptually—efficacy, media use and news coverage.  One study using phone survey data found political involvement positively associated with perceived importance of newspapers and radio talk shows, which were positively associated with political efficacy and involvement. 77   On the other hand, after conducting a content analysis which found few “calls to action” in media reports, Kensicki78 argued that the public’s sense of non-efficacy could be due to a lack of useful media coverage.

The current study aligns itself with the political communication use of “efficacy” and the concept of “empowerment,” because responses to environmental risk so often involve collective public or political activity (as opposed to self-initiated behaviors like exercising or smoking cessation).  More important, this study embraces Newhagen and Morrell’s implicit conceptual linkage of efficacy and knowledge, whether objective or subjective.

Research Questions and Hypotheses

Because this study focuses on subjective knowledge and personal efficacy as consequences of exposure and attention to media coverage of environmental problems, and evaluation of how useful media coverage is, the research questions and hypotheses focus primarily on the relationships among those variables, and draw comparisons among Web, newspaper, and television environmental news.

First, based on the findings of Salwen and Driscoll79 and, later, Park,80 of a relationship between media (television news) exposure and subjective knowledge, we offer the first hypothesis:

H1: Media use will be positively related to subjective knowledge.

Moreover, based on the finding by Riffe and Hrach81 that related information sufficiency to attention to environmental news coverage, this study seeks a similar relationship with subjective knowledge:

H2a: Attention to environmental news coverage will be positively correlated with sense of subjective knowledge regarding environmental risks.

Newhagen’s82 finding that increased newspaper use led to decreased system efficacy (confidence in political institutions).  However, he also found that media use predicted greater personal efficacy, or belief that one can help or protect oneself.  Thus:

H2b: Attention to environmental news coverage will be positively correlated with sense of personal efficacy regarding environmental risks.

Riffe and Hrach83 also found that their measure of information sufficiency was significantly and negatively correlated with perceived usefulness of television environmental coverage, but not with usefulness of newspaper environmental coverage.  Thus, this study seeks to explore similar relationships of usefulness to subjective knowledge and personal efficacy, posing a research question rather than a directional hypothesis because of their conflicting results:

RQ1:  How will the perceived usefulness of environmental coverage in different media be related to subjective knowledge and personal efficacy regarding environmental risks?

In addition, Riffe and Hrach84 found that perception of actual environmental risk where one lives (i.e., the likelihood of being affected adversely) was significantly but negatively correlated with perceived usefulness of both television and newspaper environmental coverage. Thus, this study predicts:

H3: Evaluation of the usefulness of environmental coverage will be negatively correlated with perceived local environmental risk.

Little evidence permits directional hypotheses about the relationship of attention and exposure measures with perceived utility or usefulness of news coverage, however intuitive such relationships seem.  Therefore:

RQ2: What is the relationship between perceived usefulness of environmental news coverage and exposure and attention to environmental news coverage?

Though Riffe and Hrach85 found no relationship between information sufficiency and perception of environmental conditions where one lives, we posed an additional research question using subjective knowledge:

RQ3: What is the relationship between subjective knowledge, personal efficacy and perception of actual local environmental risk?

Finally, this study also seeks to determine if perceived usefulness of environmental news coverage makes a unique contribution to sense of subjective knowledge, personal efficacy and perceived risk, above and beyond other individual and communication variables. Thus,

RQ4: After controlling for individual and communication variables, what is the relationship among perceived usefulness of environmental news coverage, perceived risk, subjective knowledge, and personal efficacy regarding environmental risks? 


Data are from a survey conducted by trained interviewers using computer-assisted equipment at the Scripps Survey Research Center at Ohio University to call randomly selected phone numbers in Ohio during a five-day period . Of 1,268 calls yielding viable connections with eligible adults (i.e., some non-working numbers, answering machines, businesses, etc., were also called, and final disposition of other numbers was impossible during the calling period), 511 or 40% were completed. Sampling error associated with a sample of 511 is +/- 4.3% at the 95% confidence level, although sub-analyses involve fewer respondents.

Measures: Measures adapted from Einsiedel and Thorne86 were employed. To measure subjective knowledge, respondents were asked how well the following statement described their view on a scale of 1-10 (where 1 means “totally wrong” and 10 means the statement “describes you perfectly”): “I consider myself knowledgeable about the environment.” Higher scores mean greater subjective knowledge (See Table 1 for descriptive statistics).

To measure personal efficacy, the 1-10 scale tapped respondents’ sense of being able to act or make a difference, “There’s not much we can do about the environment.”  This item was later reverse-scored so higher scores indicate greater sense of personal efficacy.

The usefulness of environmental media coverage was explored using three pairs of items—two per medium.  First, respondents used a 1-5 scale (1=”very poor” and 5=”very good”) for how well each news source (Web, television, and newspaper) provided information to “inform people about health risks related to the environment.”  Then, respondents used the same scale to rate each source on informing “you about things you can do personally to deal with health risks related to the environment” (emphasis in script). 
Perceived environmental risk was also measured.  Respondents were asked to rate “the risk of suffering health problems because of the environment for people where you live,” using a 1-10 scale, with 1 being “no risk” and 10 representing “extremely high risk,” an approach used by Riffe and Knight,87 Riffe,88 and Riffe and Hrach.89 This measure of perceived risk was used in testing hypotheses and answering research questions.

Both media exposure and environmental coverage attention were treated as distinct.90   Respondents were asked how many days per week they went online for news, and the number of minutes per day spent reading the newspaper. Television news exposure questions asked the number of days per week respondents watched national and local news programs.  These two television news measures were positively and significantly correlated (See footnote in Table 1 for the statistics) and were ultimately averaged for use in regression analyses for RQ4 .

Attention was measured by a series of questions about how often respondents paid attention to stories about environmental problems on Web sites, on television, and in the newspaper, using a 1-5 scale (e.g., 1=”never watch them” and 5=”always watch them”). Demographic variables were also recorded.


The sample’s mean age was 55 years; 27% were high school graduates, 24% reported “some college,” and 23% were college graduates; and 26% reported a family income of $65,000 or higher. As in many surveys, females were overrepresented (58% compared to 51% in Census reports), while the sample more closely matched the state’s racial demographics (89% white compared to 85% in Census reports). Table 1 provides descriptive statistics for key variables:

Table 1      
Descriptive Statistics for Key Variables      
Variable N M SD
 1.   Age 483 55.10 16.23
 2.   Incomea 365 3.96 1.94
 3.   Educationb 488 3.18 1.16
 4.   Media exposure      
             National TV newsc (days/week) 511 4.29 2.70
             Local TV newsc (days/week) 511 4.47 2.70
             Newspaper (minutes/day) 511 24.19 28.54
                  Online (days/week) 511 3.39 3.12
5. Attention to Environmental Newsd      
             Environmental news on television 472 3.51 .99
             Environmental news in newspapers 360 3.54 1.05
                  Environmental news online 297 2.54 1.16
 6.   Channel Usefulnesse      
Local television informing people 481 3.36 1.20
Local television informing you personally 462 3.37 1.17
                  Mean television usefulness 454 3.37 1.07
Local newspaper informing people 444 3.27 1.24
             Local newspaper informing you personally 427 3.22    1.20
             Mean newspaper usefulness 414 3.26 1.15
                  Web informing people 253 3.72 1.08
             Web informing you personally 243 3.56 1.13
                  Mean Web usefulness 220 3.65 1.03
 7.   Subjective Knowledge      
                  Consider myself knowledgeablef 521 6.58 2.31
     8.    Personal Efficacy (reverse coded)      
             There’s not much we can do about environmentg 519 7.97 2.73
9.   Perceived Environmental Health Risks      
             Risk of suffering health problemsh 522 3.49 2.25

Note: aIncome was measured as a continuous variable (1=$15,000-$25,000; 6 = more than $65,000.

bEducation was measured as a continuous variable (1=less than high school graduate 5=post college).

c Local and National news exposure measures were positively and significantly correlated (=.52, <.001) and were ultimately averaged for use in regression analyses for RQ4.

dAttention to environmental news scored 1-5 (1=never watch/read them and 5=always watch/read them).

eChannel usefulness scored 1-5 (1=very poor and 5= very good).

fSubjective knowledge scales scored 1-10 (1=totally wrong and 10=describes your belief perfectly).

gPersonal efficacy scales scored 1-10 (1=totally wrong and 10=describes your belief perfectly).

hRisk of suffering health problems scored 1-10 (1=no risk and 10=high risk).

Respondents reported going online 3.4 days per week, with 40% reporting they never go online.  They reported watching local television news 4.47 days per week and national television news 4.3 days per week (averaged exposure: = 4.37 days, SD = 2.35), and read newspapers an average of 24 minutes per day. Respondents paid similar attention to environmental news stories on television (3.51 on a 1-5 scale) and the local newspaper (3.54), but the subset of respondents who go online reported less attention to Web-based environmental news (2.54). 

Respondents rated their subjective knowledge level (using the 1-10 agreement scale) above the mid-point (M = 6.58, SD = 2.31). Similarly, personal efficacy was measured by reverse scoring (using the 1-10 agreement scale) so higher scores meant more personal efficacy (M= 7.97, SD = 2.73).

Local television (3.36) and newspapers (3.27) received similar usefulness ratings for “informing people about health risks related to the environment,” but the subset that rated Web usefulness rated such sites higher (3.72).  For the items measuring how well sources have done “in informing you about things you can do personally to deal with health risks related to the environment,” ratings followed a similar pattern, with local television (3.37) faring better than local newspapers (3.22), but the Web rated highest (3.56).   Correlation coefficients indicated that the two evaluations of each source (for “people” and for “you…personally”) were similar (Pearson’s r =.71 for Web; .65 for television; and .72 for newspapers, all significant at <.001). The paired items were later averaged to create a “usefulness” rating for each medium.

Hypothesis Tests and Research Questions

Bivariate correlations (Pearson’s r) were used to test the relationships predicted in all three hypotheses and to answer three of four research questions. Multiple regression analysis was employed to answer RQ4.

Table 2 provides the Pearson coefficients to address H1, whichpredicted thatmedia use would be positively related to subjective knowledge.  The hypothesis was tested with measures of exposure to Web, newspaper, and local and national television environmental news, and received limited support, for one of four media.  As Table 2 shows, only exposure to newspaper (=.10) was significantly correlated with subjective knowledge, and the correlation was modest.  Neither Web news exposure nor national or local television news exposure was related to subjective knowledge, though Web exposure was related to personal efficacy.

The coefficients in Table 2 also provide support for H2a, whichpredicted that attention to environmental news coverage—online, in the newspaper, and on television—would be positively correlated with subjective knowledge, while H2b predicted similar correlations with personal efficacy.  All three measures of attention were significantly correlated with subjective knowledge:  the bivariate r for subjective knowledge with online attention was .21, for newspaper it was .14, and for television news it was .16.  H2a was supported.

Table 2 is also relevant to H2b, which predicted a relationship between attention to environmental news coverage and personal efficacy, gained support with correlations being observed between personal efficacy and attention to television coverage (= .14) as well as attention to newspaper coverage (= .13).  However, attention to online news, while the strongest correlate of subjective knowledge, was not related to personal efficacy. Many of these correlations are modest, of course, but collectively they point to a positive relationship between media attention and subjective knowledge as well as between media attention and personal efficacy.

Table 2      
Pearson’s Correlations among Key Predictors and Criterion Variables
Variable Subjective Knowledge Personal Efficacy Perceived Risk
1. Age .09* -.12** -.08
2. Income .03 .12** -.06
3. Education .14** .20** .03
4. Gender dummya .08 -.12** -.04
5. Race dummyb .03 .13** -.06
6. Exposure national TV .08 -.03 .06
7. Exposure local TV .04 -.03 .06
8. Attention to TV .16** .14** .05
9. Exposure to newspaper .10* -.07 -.02
10. Attention to newspaper .14** .13* .05
11. Exposure to Web .01 .10* .03
12. Attention to Web .21** .07 .05
13. TV usefulness .01 -.04 -.12**
14. Newspaper usefulness .05 -.03 -.08
15. Web usefulness -.03 .01 .11

Note:   *p <.05, **<.01 (two-tailed tests). [1] Gender dummy coded 1=male, 0=female; b race dummy coded 1=white, 0=nonwhite.

The data in Table 2 also help answer RQ1, whichasked how the usefulness of media environmental coverage would correlate with subjective knowledge and personal efficacy.  The answer, simply, is not at all. Table 2 correlations between the measures of perceived usefulness of Web, newspaper, and television with both subjective knowledge and personal efficacy were inconsequential.

Readers should refer to Table 2 again for results related to H3, whichpredicted that evaluation of the usefulness of media environmental coverage would be negatively correlated with perceived local environmental risk (1-10 scale).The results were mixed. As hypothesized, and found previously,91 television coverage usefulness was significantly, negatively correlated with risk perception (r = -.12), but neither Web nor newspaper environmental news usefulness was significantly correlated with perceived risk.  To answer H3, then, those who perceived higher risk where they live were only critical of the usefulness of television environmental coverage. 
Data related to RQ2, which asked about the relationship between usefulness of environmental news coverage and exposure and attention to environmental news coverage, are not shown in table form.  However, the results were somewhat surprising.  Online news exposure was not significantly correlated (r = .13, NS ) with perceived Web usefulness, but attention to online environmental news garnered a strong positive and significant correlation with usefulness (= .23, p < .01). Television news exposure was significantly correlated with the usefulness of television news coverage (= .10, <  .05), but self-reported level of attention to television environmental news was not correlated significantly with usefulness (r = .06, NS).   Similarly, for newspapers, exposure (r = .10,  NS ) and attention (r = .01,   NS) were not correlated with usefulness of newspaper coverage. Thus, to answer RQ2, only exposure to television environmental news and attention to online environmental news were significantly correlated with usefulness of each respective news channel.

 Results related to RQ3, which asked about the relationship between subjective knowledge, personal efficacy, and perception of local environmental risk where one lives, are not shown in table form. There was no significant correlation between subjective knowledge and perceived environmental risk (r = -.06, NS) or between personal efficacy and risk (r = .01, NS.)
Finally, this study’s RQ4 sought to determine if perceived usefulness of environmental news coverage made an additional, or unique, contribution to subjective knowledge and personal efficacy and perceived risk, above and beyond other individual and communication variables .
First, however, note the Table 2 correlations of subjective knowledge, personal efficacy and perceived risk to the demographic variables, which were used to answer RQ4. Subjective knowledge was significantly correlated with two demographic indicators: age and education. Personal efficacy was positively and significantly correlated with all three socioeconomic indicators: education, income, and race (being white). In addition, personal efficacy was negatively, yet significantly, associated with age and gender (being female).  While Riffe and Hrach92 found that poorer and non-white respondents perceived the risk where they lived to be greater, a finding consistent with the environmental justice literature, the present data showed no significant relationship between perceived risk and race or income, though the negative signs are suggestive.

Tables 3, 4 and 5 show data related to RQ4, which asked about relationships among perceived usefulness of environmental news coverage, perceived risk, subjective knowledge, and personal efficacy regarding environmental risks. To answer RQ4, nine different hierarchical linear regression models compared how the perceived usefulness of each of the three environmental news channels predicted (in separate models) subjective knowledge, personal efficacy and perceived risk. In all models, however, all four entered blocks were conceptually similar: the first block contained individual demographic variables, the second block contained environmental news exposure via a particular channel, the third block contained environmental news attention to that channel and the final block contained perceived usefulness of the news channel. Standard diagnostic procedures were conducted. No violations of test assumptions were evident.

A hierarchical order of entry is reported here, based on the distinction between media exposure and attention,93 evidence in Table 2’s bivariate analyses, the empirical precedent in the Riffe and Hrach94 study, and the goal of distinguishing specific associations among three news channels for three types of criterion measures.  Variables in blocks one, two and three that were significant at time of entry remained significant at the p = .05 level in the final models that included all variables.

Predicting Subjective Knowledge

As shown in Table 3, across all three media channel models, education was a significant predictor of subjective knowledge, after controlling for other demographic variables. In other words, more education was associated with higher levels of subjective knowledge. Of the communication predictor variables, only attention to environmental news coverage on television significantly predicted subjective knowledge. Respondents with higher attention scores were associated with higher subjective knowledge. The perceived usefulness of environmental news coverage in newspapers, television, and on the Web did not add a unique contribution to prediction of subjective knowledge. 

Predicting Personal Efficacy  

As shown in the Table 4, several demographic variables were significant predictors of personal efficacy. Younger, female, white and more educated respondents reported more personal efficacy. In the Web model, gender and race—but not age or education—were significant predictors of personal efficacy. Among communication variables, only greater attention to television environmental news coverage significantly predicted higher personal efficacy. As with the models predicting subjective knowledge, perceived usefulness of environmental news coverage in newspapers, on television, and on the Web was not associated with personal efficacy.

Predicting Perceived Risk.

Table 5 demonstrates a different pattern. Across all three models, the only significant demographic predictor of perceived risk was education in the Web model. Attention to Web environmental news coverage was a significant predictor of perceived risk, after controlling for demographic variables and Web exposure. The more participants paid attention to Web environmental news coverage, the higher they rated the risk of health problems because of the environment in their area.

Finally, perceived usefulness of environmental news coverage in television and on the Web did significantly predict perceived risk, after controlling for demographic, exposure, and attention variables. However, the television and Web models differed. Lower usefulness was ascribed to environmental news coverage on television when there was higher perceived environmental risk, but higher usefulness of Web environmental news coverage was associated with higher perceived environmental risks.

Thus, to answer RQ4, perceived usefulness of environmental news coverage made a unique contribution only in predicting perceived risk (and only for the television and Web models), but it did not contribute to subjective knowledge or personal efficacy. Of course, these results should be understood as they are described; they suggest relationships and associations among variables, but they do not provide evidence of causation.

Table 3
Summary of Hierarchical Regression Analysis for Predicting Subjective Knowledge
Variable Newspaper(N=246) Television (N=310) Web(N=154)
  Betaa R2 b Beta R2 Beta R2
Block 1   .03   .04   .06
Age .05   .12   .09  
Genderc .02   .07   .14  
Raced .00   -.00   .03  
Education     .18**   .15*      .20*  
Income -.11   -.04   -.12  
Block 2   .03   .04   .06
Exposure -.02   .09   .02  
Block 3   .04   .07   .08
Attention toEnvironmental coverage .05   .16**   .14  
Block 4   .04   .07   .08
Perceived usefulness of environmental coverage .03   .05   .01  

Note: * p <.05; ** p <.01; ***p <.001; astandardized regression coefficients are reported at time of entry into the model, after controlling for other model variables (when applicable); bcumulative R2cgender was coded male=1, female=0; drace was coded white=1, other=0.

Table 4
Summary of Hierarchical Regression Analysis for Predicting Perceived Efficacy
Variable Newspaper (N=246) Television (N=310) Web(N=154)
  Betaa R2 b Beta R2 Beta R2
Block 1   .14   .13   .13
Age -.15*   -.14*   -.14  
Genderc   -.24**   -.24***   -.24**  
Raced   .20***   .21***   .23***  
Education .19**   .14*   .14  
Income .08   .06   .05  
Block 2   .14   .13   .14
Exposure .05   -.02   .06  
Block 3   .15   .14   .15
Attention toEnvironmental coverage .08   .12*   .12  
Block 4   .15   .14   .15
Perceived usefulness of environmental coverage -.01   .02   .01  

Note: * p <.05; ** p <.01; ***p <.001; astandardized regression coefficients are reported at time of entry into the model, after controlling for other model variables (when applicable); bcumulative R2cgender was coded male=1, female=0; drace was coded white=1, other=0.

 Table 5
Summary of Hierarchical Regression Analysis for Predicting Perceived Risk
Variable Newspaper (N=246) Television (N=310) Web(N=154)
  Betaa R2 b Beta R2 Beta R2
Block 1   .02   .02   .07
Age -.07   -.08   .03  
Genderc -.01   -.00   .01  
Raced -.03   .01   -.09  
Education .13   .10     .25**  
Income -.07   -.12   -.13  
Block 2   .02   .02   .08
Exposure -.06   .06   -.09  
Block 3   .03   .02   .13
Attention toEnvironmental coverage .10   .05   .23**  
Block 4   .03   .06   .15
Perceived usefulness of environmental coverage -.06   -.19***   .17*  

Note: * p <.05; ** p <.01; ***p <.001; astandardized regression coefficients are reported at time of entry into the model, after controlling for other model variables (when applicable); bcumulative R2cgender was coded male=1, female=0; drace was coded white=1, other=0.

Discussion and Conclusion

This study examined the relationships among exposure/attention to environmental news coverage, perceived usefulness of environmental news coverage, subjective knowledge, personal efficacy, and perceived local environmental risk. 

As hypothesized, attention to Web, television and newspaper environmental news w as positively related to subjective knowledge.  Only television and newspaper attention were related to personal efficacy.  Although causal direction is problematic with bivariate correlations and therefore not assumed in this study, this finding is consistent with the Riffe and Hrach95 suggestion that subjective knowledge (or “information sufficiency”) can be viewed as a criterion variable. The performance of television attention in predicting subjective knowledge in the regression model, controlling virtually all the other variables including education, is particularly strong evidence for this argument, and may indicate the need to further explore differences among sources of environmental information.

Somewhat surprisingly, though, none of the exposure or attention measures were significantly correlated (bivariate r) with participants’ perception of the actual risk for people where they live of developing health problems because of local environmental conditions.  Attention to Web coverage, however, was a strong predictor of perceived risk in the regression analysis when other variables were controlled, as was education.

In addition to examining the role of subjective knowledge as a consequence of media consumption, this study also argued for an important and independent role for perceived media usefulness in explaining people’s subjective knowledge, personal efficacy and risk perception.  Beyond exposure and attention to coverage, the perceived usefulness or “value” of that coverage should influence the three criterion measures.  In short, if a person believed the information obtained from individual media is valuable, accurate, or, generally, “useful,” etc., that person is more likely to self-appraise as being knowledgeable and empowered 
However, perceived usefulness of any of the three sources was not related to either subjective knowledge or personal efficacy as operationalized for this study.  Whether evaluating Web, newspaper or television coverage; whether the relationship was tested with bivariate correlations or regression analyses, the usefulness (or uselessness) of those sources, as measured here, had virtually nothing to do with people’s subjective knowledge or personal efficacy, given other factors controlled statistically. Even though self-report survey data like these are fraught with potential distortion, this lackof any relationship is striking. Perceived environmental risk where one lives, on the other hand, was related (negatively) to the usefulness of television coverage and (positively) to the usefulness of Web coverage. 
Among many things, the results suggest that those who pay attention specifically to media coverage of the environment generally consider themselves knowledgeable and confident that citizens can have an effect on environmental problems (personal efficacy), a far cry from the “media-constructed disconnect” described by Kensicki.96 Because media exposure was not related to these constructs (subjective knowledge and personal efficacy), however, the Chaffee and Schleuder97 distinction between exposure and attention comes to mind. On the other hand, exposure to Web news wasmodestly correlated with personal efficacy, but Web attention was not, suggesting that for some kinds of news and information, the Web may represent a new twist on the exposure-attention distinction.

In fact, perceived usefulness of Web environmental coverage was the only predictor that elicited higher perceptions of environmental risk in one’s area. Lower perceived usefulness of television environmental news, on the other hand, predicted higher perceived local environmental risk.  The case of the Web may suggest an information-seeking scenario: more risk leads to greater reliance on a medium that requires active seeking but ultimately enables the user to reach a certain level of subjective knowledge.98 The case of television coverage, however, may suggest a scenario of dissatisfaction: aware of local environmental risks, viewers find television coverage limited, wanting, and not useful. 
In general, though, perceived usefulness could not compete with attention variables in predicting subjective knowledge and personal efficacy.

Of course, this study has several limitations. As a telephone survey with a less-than-ideal completion rate, it suffers unknown sample biases. While it employed indicators (some adapted and some newly developed) of subjective knowledge, personal efficacy, and perceived environmental risk that seem to have face validity, multiple indicators for each would have been preferred. Moreover, the study’s measures of “usefulness,” while they were conceived within a “mobilizing information” framework, fell short of specifying “task-related” usefulness beyond “informing you about things you can do personally to deal with health risks related to the environment.”   Lemert’s99 levels of identificational, locational, and tactical mobilizing information are far more specific than the measures used here.

Finally, the study tested its hypotheses and sought answers to research questions for only three media sources—Web, newspapers, and television.   While newspapers and television remain dominant news and information sources for the general public, that duo constitutes only a subset of the available sources people use to develop their environmental knowledge base, particularly those people highly attentive to such issues.
Inclusion of the Web suffered from a problem common to many surveys: the measure asked about environmental news “on the Web or Internet,” without differentiating the New York Times’ online site from an environmental interest group’s site or the Environmental Protection Agency’s (EPA) site, etc.    

Still, the study has shown that there is a solid relationship of attention to environmental news and sense of subjective knowledge and personal efficacy.  However, perceived usefulness of that coverage, at least as measured here, seems strikingly but paradoxically to have little role or influence in subjective knowledge or personal efficacy, which one would assume are important if citizens are to live with, understand, and solve environmental problems in their communities.

Future research should continue to explore the relationships among these variables, focusing in particular on specific environmental threats or problems.  Contrasts between subjective knowledge and actual knowledge might be revealing, as would enhanced measurement of personal efficacy.  Citizens’ sense of empowerment in dealing with environmental problems is far more complex than measured here, and may include awareness of organized activity, knowledge about how to engage with environmental agencies or other actors, and even a sense of being able to act to protect oneself and one’s family.  Such future research would help scholars and practitioners better understand the complex relationships between exposure and attention to various types of environmental news, perceived knowledge and efficacy, and the likelihood that citizens will act to deal with environmental risks.

Brooke Weberling is an assistant professor in the Schools of Journalism and Mass Communication at the University of South Carolina.  Jennette Lovejoy is an assistant professor in the Department of Communication Studies at the University of Portland.  Daniel Riffe is the Richard Cole Eminent Professor at the School of Journalism and Mass Communication at the University of North Carolina at Chapel Hill.