Too positive to change? Examining optimism bias as a barrier to media effects on environmental activism

Too positive to change? Examining optimism bias as a barrier to media effects on environmental activism

Accepted Manuscript Too positive to change? Examining optimism bias as a barrier to media effects on environmental activism David Jiménez-Castillo, Jo...

588KB Sizes 0 Downloads 26 Views

Accepted Manuscript Too positive to change? Examining optimism bias as a barrier to media effects on environmental activism David Jiménez-Castillo, José Manuel Ortega-Egea PII:

S0272-4944(15)30026-8

DOI:

10.1016/j.jenvp.2015.07.004

Reference:

YJEVP 963

To appear in:

Journal of Environmental Psychology

Received Date: 13 June 2014 Revised Date:

17 July 2015

Accepted Date: 19 July 2015

Please cite this article as: Jiménez-Castillo, D., Ortega-Egea, J.M., Too positive to change? Examining optimism bias as a barrier to media effects on environmental activism, Journal of Environmental Psychology (2015), doi: 10.1016/j.jenvp.2015.07.004. This is a PDF file of an unedited manuscript that has been accepted for publication. As a service to our customers we are providing this early version of the manuscript. The manuscript will undergo copyediting, typesetting, and review of the resulting proof before it is published in its final form. Please note that during the production process errors may be discovered which could affect the content, and all legal disclaimers that apply to the journal pertain.

ACCEPTED MANUSCRIPT

Too positive to change? Examining optimism bias as a barrier

Authors:

RI PT

to media effects on environmental activism

M AN U

Affiliations:

SC

David Jiménez-Castillo a,* and José Manuel Ortega-Egea b

a

Department of Economics & Business, University of Almería (ceiA3), Almería, Spain. Email: [email protected]

b

TE D

Department of Economics & Business, University of Almería (ceiA3), Almería, Spain. Email: [email protected]

* Corresponding author at: Department of Economics & Business, University of

EP

Almería (ceiA3), Ctra. Sacramento, s/n, 04120 Almería, Spain. Tel.: +34 950214171;

AC C

fax: +34 950015178. Email: [email protected] (D. Jiménez-Castillo).

1

ACCEPTED MANUSCRIPT

Too positive to change? Examining optimism bias as a barrier to media effects on environmental activism Abstract:

RI PT

This paper adopts a novel optimism bias approach to improve the understanding of media effects on environmental activism. The main research question here concerns if and how optimism bias moderates media type associations with environmental activism. The data

SC

came from personal surveys conducted with a sample of 592 adults living in southeastern Spain. The results of hierarchical regression showed the differential perceived type of media

M AN U

influence on self for self-reported environmental activism; the Internet and print media were most associated with personal action, followed by televised documentaries and interviews. Optimism bias negatively moderated the relationships of TV documentaries/interviews and radio, but not those of the Internet and print media, with self-reported environmental

TE D

activism. Overall, the findings are indicative that print media, TV documentaries/interviews, and the Internet are helpful for environmental campaigning. Implications for theory, policy,

AC C

EP

and practice are discussed.

Key Words:

Media effects, optimism bias, environmental activism, hierarchical regression.

2

ACCEPTED MANUSCRIPT

Too positive to change? Examining optimism bias as a barrier to media effects on environmental activism 1. Introduction

RI PT

Increased media coverage has substantially raised public awareness and concern about environmental issues, but has typically failed to induce effective and persistent behavioral changes (Lorenzoni, Nicholson-Cole, & Whitmarsh, 2007; Whitmarsh, 2009). This

SC

environmental communication challenge raises the issue of if and how different types of media encourage people’s active involvement in environmental organizations and

M AN U

demonstrations (i.e., environmental activism).

Communication researchers have found that the perceived effect of media coverage is a reliable predictor of pro-environmental behavior changes (e.g., Jensen & Hurley, 2005). Environmental psychology scholars have argued that optimism bias (i.e., people’s tendency

TE D

to underestimate their likelihood of suffering from environmental risks, as they feel less vulnerable than others) (Hatfield & Job, 2001; Pahl, Harris, Todd, & Rutter, 2005) limits people’s perception about their susceptibility to environmental hazards, thus leading to

EP

human inaction (Gifford, 2011; Gifford et al., 2009). Yet, empirical evidence suggests that optimism bias does not negatively predict environmental activism (e.g., Pahl et al., 2005).

AC C

This backdrop raises the important research question about the role of optimism bias in the relationship between perceived type of media influence on self and environmental activism. A novel optimism bias approach is adopted in this study to explain how different

types of media are linked to environmental activism. Consistent with prior findings in communication research, perceived type of media influence on self is proposed as a direct predictor of self-reported environmental activist behavior; this relationship is posited to be contingent on people’s optimism bias. The view of optimism bias as a moderator delves into the idea that the optimism bias mechanism acts as a psychological barrier that may attenuate

3

ACCEPTED MANUSCRIPT or annul media effects on pro-environmental behavior change (see Gifford, 2011). The importance of understanding the interplay between different types of media, optimism bias, and environmental activism is valuable not only from a scientific, but also from a practical point of view (e.g., efficient resource allocation in environmental communication

RI PT

campaigns). Indeed, the fact that some types of media are more helpful than others in

developing pro-environmental perceptions and behaviors (e.g., Nixon & Saphores, 2009; Nonami et al., 1997) questions whether optimism bias is a factor that helps better understand

SC

these differential media effects.

M AN U

2. Background 2.1. Media effects on pro-environmental behavior

Mass media is considered a major source of social influence on citizens’ pro-environmental behavior (Chan, 1998). Evidence for this assumption comes from prior research based on the type of media. For instance, several scholars have shown that use of or exposure to print

TE D

media is positively associated to pro-environmental behavior (e.g., Besley & Shanahan, 2004; Chan, 1998; Nixon & Saphores, 2009). They also showed the superiority of print media over

EP

other media in predicting this behavior. Results for TV are mixed. Several studies found no significant relationship between TV viewing and pro-environmental behavior (e.g., Chan,

AC C

1998; Nixon & Saphores, 2009), whereas others revealed that this association is highly dependent on the type of TV genre. For instance, TV news or nature documentaries are shown to predict recycling practices (Corral-Verdugo, 2003; Holbert, Kwak, & Shah, 2003) but watching commercial TV is negatively correlated with these practices (e.g., CorralVerdugo, 2003). A more complex view of media effects suggests that pro-environmental beliefs and behaviors may be indirectly affected by people’s expectations of how others perceive and react to media influences (i.e., third- and first-person effects). However, these perceptions

4

ACCEPTED MANUSCRIPT seem to have limited or null effect on positive, desirable behaviors in response to health or environmental hazards (Jensen & Hurley, 2005; Lin, 2013; Wei, Lo, & Lu, 2008). Importantly, the recent empirical evidence from Lin’s (2013) environmental study has emphasized the positive behavioral consequences of perceptions of media influence on self

RI PT

when focusing on positive pro-social or pro-environmental content and media.

In line with Lin (2013), for desirable (beneficial) environmental behaviors such as environmental activism, a key determinant is the perceived media influence on self. Why

SC

should that be so? People rely on the media for environmental information (Arlt, Hoppe, & Wolling, 2011; Eyal, Winter, & DeGeorge, 1981). Furthermore, media can increase the

M AN U

personal relevance of environmental risk information by helping people recognize that environmental hazards could affect them (e.g., Coleman, 1993). Information that is perceived as personally relevant is more likely to spur interest and influence people (e.g., Ajzen, Brown, & Rosenthal, 1996). In addition, according to self-serving bias, individuals see

TE D

themselves as more influenced by socially desirable messages (Gunther & Mundy, 1993; Gunther & Thorson, 1992). These conceptions support the receiver-centered approach to media effects on pro-environmental action (see Arlt et al., 2011).

EP

From an operationalization perspective, the dyads between media types and

AC C

communication genres are important in analyzing media influences, and accounted for in this study. In real-world scenarios, media influences (on a variety of outcomes) likely differ and are shaped by the media context (e.g., television, magazines, the Internet) and the specific genre or format being chosen (e.g., campaigns, news, documentaries). This is acknowledged in much of the communication literature (e.g., Gunther & Storey, 2003). Thus, in this study, perceived type of media influence on self is operationalized as the extent of influence of environmental information conveyed by a specific media (e.g., TV) in a specific genre (e.g., documentaries).

5

ACCEPTED MANUSCRIPT 2.2. The optimism bias Optimism bias refers to a psychological mechanism—a perceptual judgment—that has been proposed to articulate the self-other perceptual gap (Weinstein, 1980). Formally, it has been defined as a psychological tendency that leads people to “think they are less likely to

RI PT

experience negative future events but more likely to experience positive future events as compared to others” (Wei, Lo, & Lu, 2007, p. 667). Hence, optimism bias regarding

environmental hazards (i.e., environmental optimism bias) would predict that people think

SC

others are more vulnerable to harmful influences of environmental problems than they are (Gifford et al., 2009; Hatfield & Job, 2001). Compared to other personal risks, such as

M AN U

contracting AIDS, smoking risks, suffering accidents or the Y2K problem, the optimism bias has received less direct consideration in the environmental risk domain—for exceptions, see Hatfield and Job (2001) and Pahl et al. (2005).

A first critical issue coming into play in research on optimism bias is the dichotomy

TE D

between desirable versus undesirable outcomes (Li, 2008). Messages with desirable (beneficial) outcomes such as environmental protection messages (see Hoorens & Ruiter, 1996) might gear self-protective actions from environmental risks. However, the optimism

EP

bias would make people believe that the risk is higher for others, thus leading to human

AC C

inaction; that is, environmental optimism bias could be understood as a mechanism that attenuates or annuls the effect of perceived media influence (on self) on environmental activism. Several arguments support this assertion. First, Gifford (2011) stressed that optimism bias, as a psychological barrier, has become a contingency factor that may hinder pro-environmental behaviors. Accordingly, the cognitive underestimation of the likelihood of suffering from environmental risks can lead people to reject the behavioral changes that environmental media messages attempt to encourage. Second, and in the same line of reasoning, scholars have argued that optimism helps reinforce self-esteem in evaluating

6

ACCEPTED MANUSCRIPT media messages of negative events (e.g., Brosius & Engel, 1996; Culbertson & Stempel, 1985). This highlights the idea that there is “a human tendency to see the world through optimistic or self-serving lenses” (Gunther & Mundy, 1993, p. 58). Third, many scholars have implicitly assumed that media variables and optimism bias can interact to account for

RI PT

pro-environmental behavior. For example, Pahl et al. (2005) recommended avoiding

messages aimed at undermining optimism bias in media campaigns based on their finding that this psychological mechanism is not directly related to pro-environmental behavior.

SC

A second key concern in the environmental risk domain lies in understanding the

relative helpfulness of the different channels now available for message delivery. It is argued

M AN U

here that the effect of optimism bias on the relationship between perceived media influence on self and environmental activism will be more or less intense depending on the type of channel used to deliver environmental messages. This rationale is consistent with the assumption that the different nature of media (e.g., informational, communicative) can alter

(Li, 2008). 2.3. The current study

TE D

people’s information processing mode, and thus can make a difference in the optimism bias

EP

Based on the above discussion, an interactive fit argument is proposed arguing that

AC C

environmental activism is attributable to a match between perceived type of media influence on self and optimism bias. In particular, the authors posit that optimism bias moderates the relationship between perceived type of media influence on self and self-reported environmental activism. Given the focus on environmental hazards, the following general hypothesis (HG) is formally stated: HG: The association between perceived type of media influence on self and self-reported environmental activism will be less intense under higher levels of environmental optimism bias.

7

ACCEPTED MANUSCRIPT 3. Method 3.1. Participants

Participants were adult residents (aged 18 and older) of a medium-size city in southern Spain. In total, 650 people were approached for survey, of whom 592 (91%) agreed to participate in

RI PT

the study. This final sample consists of 256 males (43%) and 336 females (57%), with a mean age of 36.41 years (SD = 15.61). Comparison with census data of the Spanish population indicates that the sample is balanced in terms of gender and age. Practically all participants,

SC

except for 5 (1%), reported some form of formal education. Nearly half of the sample (45%) had completed post-secondary education degrees, 39% had completed secondary education,

M AN U

whereas 15% had only primary education. As for political ideology, half of the sample (50%) self-identified as moderates (center), 35% as liberals (left-wing), and the remaining 15% as conservatives (right-wing). 3.2. Procedure

TE D

A convenience sample was collected using a paper-based structured questionnaire that participants were asked to fill out. Trained students from four undergraduate business and marketing classes were enlisted to administer the questionnaires in designated public areas,

EP

based on gender and age quotas. The research assistants personally delivered and assisted

AC C

participants in their self-completion of the questionnaires (i.e., by clarifying issues or concerns about the questions), explained the purpose of the study, and assured the respondents’ confidentiality. The study was introduced to potential participants as “academic research investigating people’s environmentally responsible behavior and media influence”. Participation was entirely voluntary; no financial or additional incentives were offered. Each survey took an average of 15 minutes to complete.

8

ACCEPTED MANUSCRIPT 3.3. Measures The questionnaire comprised questions on personal self-reported environmental activism, perceived type of media influence on self, perceived environmental risks, as well as sociodemographic information. The main variables and constructs were measured as follows:

RI PT

Environmental activism. The scale of environmental activism comprised six self-

reported statements of personal activism action intended to motivate greater engagement with the environment in the public/political sphere (i.e., in the policy system, organizations, and of

SC

the broader population). The environmental activism scale includes both lower- and highercommitment measures of environmental activism behavior; two representative items are

M AN U

“Circulate petitions demanding an improvement of government policies regarding the environment” (lower-commitment activism) and “Participate in protests against current environmental conditions” (higher-commitment activism). Environmental activism items were sourced from two key environmental studies (Dono, Webb, & Richardson, 2010;

TE D

Seguin, Pelletier, & Hunsley, 1998) and rated on 7-point scales from 1 to 7—where 1 denotes never and 7denotes always (see the Appendix for the complete list of items and sources). A single measure of environmental activism was constructed by computing regression-based

EP

factor scores from a principal components analysis of the six items. Computing factor scores

AC C

offers a more refined approach to synthesizing the outcome (dependent) variable here, compared to alternative averaging methods (DiStefano, Zhu, & Mindrila, 2009), in light of the unidimensional nature and previously tested psychometric properties of the environmental activism scale (see Dono et al., 2010). Perceived type of media influence on self. The perceived influence of different media types and genres on self was measured with 25 self-report ratings, on a 1 to 7 scale, regarding the extent of influence of environmental media coverage in different media genres (e.g., campaigns, news, ads, series, documentaries, interviews)—conveyed by four types of media

9

ACCEPTED MANUSCRIPT and communication technology (i.e., television, radio, print media, and the Internet) (1 = no influence; 7= extreme influence).For instance, participants rated the extent to which “environmental campaigns broadcasted on TV” or “ads on social networks” influenced themselves. Separate composite indexes (mean scores) of perceived influence of media on

RI PT

self were created from exploratory factor analysis of the individual items and subsequent averaging of the resulting factors (as discussed in the Results section). The averaging

reduction method is appropriate to the exploratory characteristics of the media constructs

SC

under investigation (Tabachnick & Fidell, 2001), which are “untested and exploratory, with little or no evidence of reliability or validity” (Hair, Black, Babin, Anderson, & Tatham,

M AN U

2006, p. 140).

Optimism bias. The procedure performed by Pahl et al. (2005) to develop a measurement score for optimism bias was followed in this study. Participants were asked to rate 19 environmental risks (e.g., pollution, genetically modified food, pesticides and

TE D

herbicides, water shortage) to the self, followed by ratings of the same risks for a typical person of the same background as respondents. Respondents were asked “How much at risk do you think you [a typical person (of your age, sex and background)] are [is] of experiencing

EP

harmful effects of each of the following hazards sometime in the future?” Again, responses

AC C

were given on 7-pointscales ranging from not at risk at all (1) to strongly at risk (7). Optimism bias scores, calculated as the difference between other- and self-ratings, were computed for each of the 19 environmental risks. A composite index (mean score) of optimism bias was constructed by averaging across the 19 difference items. The averaging method is also preferred for synthesizing the main moderating variable of the study—that is, an additive operationalized index of optimism bias (see Pahl et al., 2005). Control variables. Gender, proxied as biological sex, was coded with 1 designating men and 2 designating women. Age was measured as a continuous variable in years.

10

ACCEPTED MANUSCRIPT Education was measured by a three-category variable (coded 1 to 3) reflecting the highest level of formal education completed, with1 for elementary education, 2 for secondary

education, and 3 for post-secondary education. Political ideology was assessed by having the respondent place his/her predominant political views on a three-point scale: 1 =

RI PT

conservative/right-wing; 2 = moderate/center; and 3 = liberal/left-wing. Following Pahl et al. (2005), the absolute level of perceived environmental risk was also controlled because, a priori, it could play a role in predicting pro-environmental behaviors. It was measured by

SC

calculating the averaged risk ratings for self and other.

M AN U

4. Results

To simplify the analysis of variables measuring perceived type of media influence on self, an exploratory factor analysis with maximum likelihood extraction was performed. An oblique rotation method (direct oblimin, delta = 0) was used to facilitate interpretation, because correlations among the different media factors could be expected. The factor analysis of the

TE D

25 items generated five factors with eigenvalues greater than one, accounting for 74% of the total variance. As the first factor did not explain more than half of the total variance (45%),

EP

common method variance does not seem to be problematic in this study (Podsakoff & Organ, 1986). Results are shown in Table 1.

AC C

Table 1

Factor loadings for exploratory factor analysis (EFA) with oblique rotation TV campaigns/news/ ads/series .78

TV documentaries/ interviews .03

TV news

.75

TV ads

Radio

Print media

Internet

-.04

.03

-.01

.14

-.08

-.04

.01

.89

-.14

-.01

.04

-.03

TV series

.35

.09

.01

.12

-.22

TV documentaries

.43

.47

.03

.05

-.05

TV interviews

.13

.56

-.23

.07

-.06

TV environmental campaigns

11

ACCEPTED MANUSCRIPT .06

-.05

-.84

.03

-.02

Radio news

.01

.05

-.90

-.01

-.00

Radio ads

.15

-.17

-.79

.08

-.04

Radio interviews

-.11

.26

-.71

.08

-.06

Press environmental campaigns

.08

.09

-.15

.64

.00

Magazine environmental campaigns

.02

.04

Press news

.03

.11

Magazine news

-.06

.05

Press ads

.11

-.10

Magazine ads

.04

-.14

Website/email environmental campaigns

.04

.02

Social network environmental campaigns

.01

Website/email news

-.07

Social network news

-.03

Website/email ads

.06

Social network ads

RI PT

Radio environmental campaigns

.89

-.00

-.28

.58

.01

.02

.87

-.04

-.27

.62

-.02

-.00

.86

-.05

.03

.04

-.81

-.07

.01

-.07

-.96

.09

.04

.13

-.80

-.03

-.02

-.06

-.96

-.07

.04

.14

-.76

.02

-.12

-.05

-.08

-.92

Websites specializing in environmental issues

-.06

.19

-.05

.19

-.48

Comments and opinions of close people through social networks

.05

.07

-.05

-.06

-.62

Comments and opinions of opinion leaders through social networks

.08

.12

-.08

.13

-.37

AC C

EP

TE D

M AN U

SC

.09

Note. Bold denotes primary factor loadings for each item.

The first and second factor comprised TV-related items. Namely, the first one

included TV campaigns, news, advertising and series; the second factor contained TV documentaries1 and interviews. This division could be explained by the fact that TV documentaries and interviews with scientists or opinion leaders usually provide technical and 1

Although the factor loading of item ‘TV documentaries’ is higher on the second factor (0.47), it should be noted that there is a similar cross-loading on the first factor (0.43) due to the conceptual relationship between the item and this factor (Hair et al., 2006).

12

ACCEPTED MANUSCRIPT scientific detail that conveys higher source credibility (expertise and objectivity), which contrast sharply with the environmental content typically found in TV news, series, and

campaigns, with its over-dramatization of episodic environmental events (see, e.g., Holbert et al., 2003; Shanahan & McComas, 1997). Accordingly, TV documentaries and interviews

RI PT

constitute sources of environmental procedural information that may arouse particular interest in, and concern about environmental issues (Corral-Verdugo, 2003; Lyons & Breakwell, 1994). The third factor comprised items related to radio and the fourth included items linked

SC

to print media (i.e., newspapers and magazines). Finally, the fifth factor contained all items related to the Internet (i.e., websites, email, and social networks).

M AN U

Also, reliability of the scales representing latent variables was checked. Cronbach’s alpha ranged from 0.74 (TV documentaries and interviews) to 0.94 (Internet) (see Table 2), indicating that the level of reliability is satisfactory. The corrected item-total correlations within each scale were in general high (>0.35), and thus indicative of internal consistency.

TE D

The items comprising each of the media factors were averaged to compute five composite mean scores that were then used as inputs (predictors) to subsequent regression analyses. Table 2 shows the summary statistics and correlations among the study variables after

AC C

EP

combining items into single indicators.

ACCEPTED MANUSCRIPT

13 Table 2 Means, standard deviations, and correlations among study variables

M

SD

Cronbach’s alpha

n.a.

n.a.

n.a.



36.41

15.61

n.a.

.01



1

2

3

4

RI PT

Correlations

5

6

7

8

9

10

11

Gender

2

Age

3

Education

n.a.

n.a.

n.a.

.06

-.47***



4

Political ideology

n.a.

n.a.

n.a.

.11*

-.20***

.16***

5

Environmental activism

1.79

1.00

.83

.02

-.00

.12**

.19***



6

TV campaigns/news/ads/series

4.54

1.33

.83

.12**

-.01

-.03

.04

.09*



7

TV documentaries/interviews

3.56

1.59

.74

.13**

.00

.07

.16***

.23***

.57***



8

Radio

4.59

1.55

.92

.04

.15***

.02

.02

.26***

.51***

.51***



9

Print media

3.34

1.37

.92

.13**

-.04

.16***

.10*

.29***

.52***

.51***

.66***



10 Internet

3.65

1.55

.94

.10*

-.31***

.28***

.14**

.27***

.50***

.43***

.43***

.58***



11 Absolute perceived risk

5.22

1.07

.94

.16***

.05

-.06

.15**

.15***

.12**

.13**

.18***

.16***

.20***



12 Optimism bias

.09

.60

.86

-.11**

.07

.00

-.10*

-.04

.00

-.02

-.05

.01

-.02

-.18***



M AN U

TE D

EP

AC C

Note. ***p<.001; **p<.01; *p<.05; n.a.=not applicable (N=592)

SC

1

14

ACCEPTED MANUSCRIPT

To exclude the risk of multicollinearity, correlation values and the variance inflation factor (VIF) criterion were examined. Correlations with values above 0.80 would indicate multicollinearity. Here, all correlations are well below this threshold, confirming that each variable represents different concepts. Likewise, all VIF values were below 3, further

RI PT

indicating that multicollinearity is not a problem in this study (Kleinbaum, Kupper, & Muller, 1988). Also worth noting are the low mean value of reported environmental activism (similar to prior studies on environmental activism such as Dono et al., 2010) and the significant

SC

correlations between this outcome construct and media variables. This finding is consistent with prior research that observed that only certain groups in the population are more likely to

M AN U

participate in environmental organizations and demonstrations (see Tranter, 2010, for a review); it also confirms that as the perceived influence of media increases, so does selfreported environmental activism.

Next, a hierarchical regression analysis was conducted to test the general hypothesis

TE D

of this study. Following the recommendations of Cohen, Cohen, West, and Aiken (2003), the variables were entered in a stepwise procedure: first, four socio-demographic variables (gender, age, education, and political ideology) were entered as control variables (Model 1);

EP

second, the absolute perceived risk was also controlled for (Model 2); third, the composite

AC C

variables of perceived type of media influence on self were entered (Model 3); fourth, the optimism bias variable was entered (Model 4); and, finally, the variables representing the interaction between perceived type of media influence on self and optimism bias were entered (Model 5). The results of hierarchical regression are shown in Table 3. Table 3 Hierarchical regression analysis predicting environmental activism Model 1

Model 2

Model 3

Model 4

Model 5

Step 1: Control variables Gender

-.02

-.04

-.04

-.04

-.05

15

ACCEPTED MANUSCRIPT Age

.08

.07

.09

.09

.08

Education

.12**

.13**

.05

.05

.05

Political ideology

.15***

.13**

.11*

.11*

.10*

.13**

.05

.04

.04

TV campaigns/news/ads/series

-.21***

-.21***

-.18**

TV documentaries/interviews

.12*

Radio

.08

Print media

.13*

Internet

.20***

Absolute perceived risk

TV documentaries/interviews x Optimism bias Radio x Optimism bias Print media x Optimism bias R2 Adjusted R2 2

Change in R F(change)

.08

.05

.13*

.16**

.20***

.19**

-.03

-.06

.01 -.11* -.11* .08 -.00

.04

.05

.14

.15

.17

.03

.04

.13

.13

.14

.04

.02

.09

.00

.02

5.13***

8.44**

11.36***

.48

3.07**

TE D

Internet x Optimism bias

.11*

M AN U

Step 3: Interactions TV campaigns/news/ads/series x Optimism bias

.12*

SC

Optimism bias

RI PT

Step 2: Main effects

EP

Note. Standardized coefficients. ***p<0.001; **p<0.01; *p<0.05

AC C

The incremental F statistics and the t tests for the coefficients of the interaction terms provide indications of the strength of the interactions as correlates of self-reported environmental activism. In Model 2, absolute perceived risk significantly predicts selfreported activism, but it lost its significance after entering media-related variables. All perceived media influence variables significantly relate to environmental activism in Model 3, except for radio2. Unexpectedly, the association with environmental activism was negative 2

The measurement of perceived type of media influence on self focuses on how likely the receivers thought media were to influence themselves and, therefore, does not imply any action or behavioral change. It is not surprising that participants reported that some media have the greatest (lowest) influence on them according to

16

ACCEPTED MANUSCRIPT

for TV campaigns/news/ads/series, as compared to the positive correlation between these two variables shown in Table 2. The addition of the optimism bias in Model 4 did not result in a significant increase in the variance explained in self-reported environmental activism; given the non-significance of this term, no influence can be attributed to optimism bias on

RI PT

environmental activism. To test whether the interactions included in the regression

significantly contribute to explaining the variance of the model, the adjusted R2 values between models were compared using F-test (Atuahene-Gima & Li, 2006). Model 5

SC

explained a greater percentage of variance than Model 3, as is shown by the significant

increase in R2. The results obtained in Model 5 indicated the existence of a moderating effect

M AN U

of optimism bias exclusively on the relationships between radio and TV

documentaries/interviews with self-reported environmental activism. Thus, the general hypothesis is supported only for radio and TV documentaries/interviews. In both cases, the moderating effect is pure because the direct association of optimism bias with environmental

TE D

activism is not significant and the interaction terms are significant and negative, as predicted (Sharma, Durand, & Gur-Arie, 1981). Therefore, the relationship between TV documentaries/interviews and self-reported environmental activism is reduced as optimism

EP

bias increases. This psychological mechanism further attenuates the already non-significant

AC C

relationship between the perceived influence of radio (on self) with environmental activism. For illustration purposes, two-way interaction plots (see Figures 1 and 2) are

constructed for the two significant interaction terms in Model 5 (i.e., those involving the media variables of TV documentaries/interviews and radio) by plotting the changes in selfreported environmental activism for three levels of the moderator and predictor variable: low (1 SD below the mean), moderate (the mean), and high (1 SD above the mean). The rest of the mean values, but their correlations with environmental activism are smaller (higher) and their regression weights negative or non-significant (positive and significant) compared to other media. If an individual believes that a medium exerts high influence on him/her, this effect is not necessarily more likely to be influential in activist terms. This perceived influence could be (or not) related to other cognitive, affective, or behavioral variables.

17

ACCEPTED MANUSCRIPT media (i.e., TV campaigns/news/ads/series, print media and the Internet) are significantly associated with self-reported environmental activism and these relationships are not

moderated by optimism bias; surprisingly, the coefficient is negative for messages delivered via TV campaigns/news/ads/series.

M AN U

SC

RI PT

Figure 1. Two-way interaction plot for perceived influence of TV documentaries/interviews predicting environmental activism, contingent on optimism bias

AC C

EP

TE D

Figure 2. Two-way interaction plot for perceived influence of radio predicting environmental activism, contingent on optimism bias

5. Discussion

The basic tenet of this study is that optimism bias is an important contingency factor moderating the influence of media on people’s environmental activism. To test this general hypothesis, a novel approach is adopted that combines perceived type of media influence on self with optimism bias to explain public-sphere, self-reported environmental activism.

18

ACCEPTED MANUSCRIPT Past research on environmental communication suggests that it does not seem

reasonable to assume that media coverage translates directly into pro-environmental concerns or actions (Besley & Shanahan, 2004). This study joins to other recent works that contradict this assertion (e.g., Sampei & Aoyagi-Usui, 2009). Two main questions are addressed in this

RI PT

study: (1) whether optimism bias inhibits perceived media effects on self-reported

environmental activism and (2) whether this psychological mechanism holds this power regardless of the type of media used to deliver environmental messages.

SC

Regarding the first question, this study points to an important conceptual refinement concerning optimism bias. The present work provides initial evidence in support of the

M AN U

proposition that the optimism bias is a moderator rather than a predictor for media effects on pro-environmental behavior. The theoretical insight that arises from this conclusion and prior findings (e.g., Pahl et al., 2005) is that it is unproductive to conceive optimism bias as a predictor in communication-based models explaining pro-environmental behavior.

TE D

Contrarily, prior studies in other research contexts have found evidence for the role of optimism bias as a predictor of behavioral outcomes such as the adoption of privacyprotective behaviors and a higher level of support for government policies to restrict the use

EP

of online information (Baek, Kim, & Bae, 2014) or taking precautions against sun exposure

AC C

(Eiser & Arnold, 1999). This controversy could be related to the differences in the social desirability of messages—optimism bias helps reinforce self-esteem in evaluating undesirable messages leading people to act. Therefore, the view of optimism bias as a contingency factor (instead of as a predictor) sheds light on how optimism bias can act as a barrier to environmental activism behavior, when people are encouraged by media messages about environmental hazards. Importantly, absolute perceived risk lost its significance as a predictor when media-related variables were entered in the model. It is similar to what happened in Pahl’s et al. (2005) study after controlling for the group membership variable. In

19

ACCEPTED MANUSCRIPT

both cases, the entered variables accounted for the predictive value of absolute perceived risk because they rendered this latter effect non-significant. The study contributes to support the receiver-centered approach to the influence of media on pro-environmental behavior and provides a useful foundation on how the perceived

RI PT

influence of media on self varies from media to media. Previous environmental research has typically confined the perceived influence of media to a single media source or genre. Yet, the present study shows that people distinguish between different media sources of influence

SC

and that these perceptual differences appear to be meaningful in behavioral terms.

The results show the differential efficacy of media on environmental activism, with

M AN U

the Internet and print media being most associated with self-reported personal action. It seems that environmental messages delivered via print media and Internet are more likely to spur people’s interest and, therefore, more likely to be influential. Unexpectedly, perceived media influence on self is stronger, but negative, for TV campaigns/news/ads/series. This is

TE D

consistent with prior findings suggesting that segments such as environmental activists tend to reject or be sceptical of TV as a believable source, preferring print media and the Internet as believable sources of environmental information (Ostman & Parker, 1987; Postmes &

EP

Brunsting, 2002). It has been also shown that greater TV viewing makes people less likely to

AC C

make sacrifices for environmental reasons (Shanahan, Morgan, & Stenbjerre, 1997). The result for TV campaigns/news/ads/series is also consistent with past (cultivation, narrative, and framing) research warning that direct exposure to and use of media may even engender anti-environmental attitudes and action (Besley & Shanahan, 2004; Mikami, Takeshita, & Kawabata, 1999). However, a positive effect is obtained for televised documentaries and interviews; arguably, as the information in those media genres comes from more credible sources (e.g., experts, environmental researchers), it seems logical that such TV viewers will be more likely to positively respond in terms of activist actions. These findings thus confirm

20

ACCEPTED MANUSCRIPT that media do make a difference when informing people about environmental hazards; it is also crucial to highlight the differential media effects on pro-environmental behavior. Comparing the results with prior studies (that examine differential media effects on behavioral outcomes) in non-environmental research, exposure to print media and Internet

RI PT

stimuli have been also found to be associated with behavioral responses such as purchase or voting behaviors; in such studies, the results for TV were significant (e.g., Danaher &

Dagger, 2013; Kaid, 2003). Therefore, it appears that TV effects on behavioral outcomes

SC

vary from domain to domain.

As regards the findings related to the optimism bias, results for TV

M AN U

documentaries/interviews and radio support the assertion that optimism bias moderates media effects on environmental activism. The larger the size of the optimism bias, the less likely people would adopt environmentally responsible actions encouraged by TV documentaries/interviews or radio influence, as hypothesized. However, such a bias may not

TE D

hinder the effect of print media and Internet on the adoption of pro-environmental actions. It seems that in assessing the impact of media on themselves, respondents took the environmental hazards presented in print media and Internet more realistically than if

EP

presented in TV documentaries/interviews or radio. This conclusion suggests that, even

AC C

though TV documentaries/interviews and radio could convey a certain degree of trust and credibility to the audience, biased optimism makes people question how realistic the communicated threat is to everyone (including themselves), which leads to reduce media effects.

Another likely explanation to these results can be drawn upon Wei’s et al. (2008) conclusions in the domain of health news. Although the environmental media coverage via TV documentaries/interviews or radio is socially desirable and beneficial as regards to promoting more environmentally sustainable behavior among the population, the unpleasant

21

ACCEPTED MANUSCRIPT

perception of suffering the consequences of an environmental hazard could activate optimism bias and thus inhibit pro-environmental behavior. Yet, the exposure to pro-environmental messages presented in print media and Internet is likely to reduce the above-mentioned ambiguity and to facilitate people’s appreciation of the communication content. Thus, a

RI PT

contribution to the optimism bias research is that this mechanism operates differently depending on the type of media. This conclusion could be interesting to refine the

inconclusive understanding of the relationship between particular media and perceptual or

SC

behavioral outcomes.

It is worth noting that, despite the existence of a moderating effect of optimism bias

M AN U

on the relationship between the perceived influence of radio on self and environmental activism, such a direct main effect of radio was not significant. This result is coherent with and confirms prior findings also showing the lack of effectiveness of radio in encouraging expected behaviors. For instance, Valenzuela (2013) found that listening to radio news does

TE D

not have a significant effect on certain forms of activism that reflect protest behavior. Overall, the findings suggest that people perceive a persuasive intention and a low-tomedium credibility from radio that can cause the above results.

EP

This study also presents some advances in the measurement of the influence of media

AC C

on people in environmental research. Most research on environmental communication has seen perceived media influence through the lens of measures of media coverage, exposure, and usage (see, e.g., Besley& Shanahan, 2004; Chan, 1999; Corral-Verdugo, 2003). The measures of perceived influence of media (on self) used in this investigation account to a greater extent for the cognitive influence of media environmental messages on people’s proenvironmental behavior change. Thus, this study overcomes the drawback that the influence of media on self at a purely perceptual or cognitive level is not captured by the media coverage, exposure, or usage constructs.

22

ACCEPTED MANUSCRIPT

The differences in the size and significance of the effects of different media types on self-reported environmental activism have clear implications for public or private interventions that promote pro-social, desirable attitudes and behaviors. As regards the benefits for environmental campaigning, this study offers guidance on how to promote more

RI PT

environmentally significant behavior, such as environmental activism, among general

audiences. Overall, the findings indicate that print media, TV documentaries and interviews with experts, and the Internet may be most helpful for environmental campaigning. The fact

SC

that optimism bias was not a significant contingent variable for print media or the Internet further reinforces the importance of such media for environmental activism.

M AN U

This study suffers from three main limitations. The first concerns the use of a convenience sample from a specific geographic area. This issue may limit the generalizability of this study’s findings; yet, compared to other environmental studies, a reasonably large sample size that satisfied age and gender quota requirements was used for the analyses. The

TE D

second is the possibility of social desirability biases. By presenting the study as an investigation of “people’s environmentally responsible behavior and media influence”, some respondents may have overstated their engagement in environmental activist behavior and

EP

perceptions of media influence on themselves—given that environmental activism and,

AC C

overall, environmental messages can be considered socially desirable. A look at the mean values of environmental activism, media influence, absolute perceived risk, and optimism bias suggests that social desirability bias is not likely to distort the findings of the study. The third important limitation arises from the use of cross-sectional data, which does not allow for excluding the possibility of reverse relationships in the models. Arguably, a person might infer his/her perceptions of media influences (on self) from the level of environmental activism behavior. For instance, the stronger association observed here of the Internet with environmental activism could be attributed to the possibilities and increasing use of the

23

ACCEPTED MANUSCRIPT Internet and social networks for activism. Another promising explanation for the greater effectiveness of the Internet (and even print media), is rooted in information-seeking biases such as selective information exposure and confirmation bias. Further research using longitudinal data and cross-lagged analysis will be necessary to exclude the potential for

RI PT

reverse relationships. Experiments are especially recommended in clarifying causal effects of media exposure and use on environmental activism. A correlational study can neither rule out third variable problems. In an attempt to minimize such confounding risks, the authors

SC

control for a set of important factors.

Future research should more closely examine the specific characteristics of media that

M AN U

generate differential media effects on ambitious pro-environmental behavior such as environmental activism. Also, the synergistic effects between media and type of environmental message should be examined. Finally, investigating the interplay between optimism bias and constructs such as environmental beliefs, values, concerns, or perceived

TE D

social influences, is of interest to gain deeper understanding of the psychological processes involved in media effects on environmental activism.

EP

6. Conclusions

The current study extends existing environmental activism literature by considering

AC C

the interplay between perceived type of media influence on self, optimism bias, and selfreported environmental activism. The results suggest that people who perceive that TV campaigns/news/ads/series influence them more report less activism; conversely, people who perceive that the Internet, TV documentaries/interviews and print media affect them more report more activism. Optimism bias attenuates the relationships between TV documentaries/interviews and radio with self-reported environmental activism. These findings add to the body of research regarding optimism bias, by showing that this mechanism can act as a contingent factor to media-environmental activism links. Implications

24

ACCEPTED MANUSCRIPT for environmental campaigning decisions are inherent to these findings. References

Ajzen, I., Brown, T. C., & Rosenthal, L. H. (1996). Information bias in contingent valuation:

RI PT

effects of personal relevance, quality of information, and motivational orientation. Journal of Environmental Economics and Management, 30, 43–57.

Arlt, D., Hoppe, I., & Wolling, J. (2011). Climate change and media usage: Effects on

SC

problem awareness and behavioural intentions. International Communication Gazette, 73, 45–63.

M AN U

Atuahene‐Gima, K., & Li, H. (2006). The effects of formal controls on supervisee trust in the manager in new product selling: evidence from young and inexperienced salespeople in China. Journal of Product Innovation Management, 23, 342–358. Baek, Y. M., Kim, E. M., & Bae, Y. (2014). My privacy is okay, but theirs is endangered:

Behavior, 31, 48–56.

TE D

Why comparative optimism matters in online privacy concerns. Computers in Human

Besley, J. C., & Shanahan, J. (2004). Skepticism about media effects concerning the

880.

EP

environment: examining Lomborg’s hypotheses. Society & Natural Resources, 17, 861–

AC C

Brosius, H. B., & Engel, D. (1996). The causes of third-person effects: Unrealistic optimism, impersonal impact, or generalized negative attitudes towards media influence? International Journal of Public Opinion Research, 8, 142–162. Chan, K. (1998). Mass communication and pro-environmental behaviour: waste recycling in Hong Kong. Journal of Environmental Management, 52, 317–325. Chan, K. K. W. (1999). Mass media and environmental knowledge of secondary school students in Hong Kong. Environmentalist, 19, 85–97. Cohen J., Cohen, P., West, S. G., & Aiken, L. S. (2003). Applied multiple

25

ACCEPTED MANUSCRIPT

regression/correlation analysis for the behavioral sciences. London: Lawrence Erlbaum Associates. Coleman, C. L. (1993). The influence of mass media and interpersonal communication on societal and personal risk judgments. Communication Research, 20, 611–628.

RI PT

Corral-Verdugo, V. (2003). Situational and personal determinants of waste control practices in Northern Mexico: A study of reuse and recycling behaviors. Resources, Conservation and Recycling, 39, 265–281.

SC

Culbertson, H. M., & Stempel, G. H. (1985). “Media malaise”: Explaining personal optimism and societal pessimism about health care. Journal of Communication, 35, 180–190.

M AN U

Danaher, P. J., & Dagger, T. S. (2013). Comparing the relative effectiveness of advertising channels: A case study of a multimedia blitz campaign. Journal of Marketing Research, 50, 517–534.

DiStefano, C., Zhu, M., & Mindrila, D. (2009). Understanding and using factor scores:

14, 1–11.

TE D

Considerations for the applied researcher. Practical Assessment, Research & Evaluation,

Dono, J., Webb, J., & Richardson, B. (2010). The relationship between environmental

EP

activism, pro-environmental behaviour and social identity. Journal of Environmental

AC C

Psychology, 30, 178–186.

Eiser, J. R., & Arnold, B. W. (1999).Out in the midday sun: Risk behaviour and optimistic beliefs among residents and visitors on Tenerife. Psychology and Health, 14, 529–544. Eyal, C. H., Winter, J. P., & DeGeorge, W. F. (1981). The concept of time frame in agendasetting. Mass communication review yearbook, 2, 212–218. Gifford, R. (2011). The dragons of inaction – Psychological barriers that limit climate change mitigation and adaptation. American Psychologist, 66, 290–302. Gifford, R., Scannell, L., Kormos, C., Smolova, L., Biel, A., Boncu, S., et al. (2009).

26

ACCEPTED MANUSCRIPT

Temporal pessimism and spatial optimism in environmental assessments: An 18-nation study. Journal of Environmental Psychology, 29, 1–12. Gunther, A. C., & Mundy, P. (1993). Biased optimism and the third-person effect. Journalism & Mass Communication Quarterly, 70, 58–67.

RI PT

Gunther, A. C., & Storey, J. D. (2003). The influence of presumed influence. Journal of Communication, 53, 199–215.

Gunther, A. C., & Thorson, E. (1992). Perceived persuasive effects of product commercials

Communication Research, 19, 574–596.

SC

and public service announcements: Third-person effects in new domains.

M AN U

Hair, J. F., Black, W. C., Babin, B. J., Anderson, R. E., & Tatham, R. L. (2006). Multivariate Data Analysis (6th ed.). Upper Saddle River, N.J.: Pearson Education Inc. Hatfield, J., & Job, R. F. S. (2001). Optimism bias about environmental degradation: The role of the range of impact of precautions. Journal of Environmental Psychology, 21, 17–30.

TE D

Holbert, R. L., Kwak, N., & Shah, D. V. (2003). Environmental concern, patterns of television viewing, and pro-environmental behaviors: Integrating models of media consumption and effects. Journal of Broadcasting & Electronic Media, 47, 177–196.

EP

Hoorens, V., & Ruiter, S. (1996). The optimal impact phenomenon: Beyond the third person

AC C

effect. European Journal of Social Psychology, 26, 599–610. Jensen, J. D., & Hurley, R. J. (2005). Third-person effects and the environment: Social distance, social desirability, and presumed behavior. Journal of Communication, 55, 242–256.

Kaid, L. L. (2003). Effects of political information in the 2000 presidential campaign comparing traditional television and Internet exposure. American Behavioral Scientist, 46, 677–691. Kleinbaum, D. G., Kupper, L. L., & Muller, K. E. E. (1988). Applied Regression Analysis

27

ACCEPTED MANUSCRIPT and Other Multivariable Methods. Boston, MA: PWS-Kent Publishing. Li, X. (2008). Third-Person Effect, Optimistic Bias, and Sufficiency Resource in Internet Use. Journal of Communication, 58, 568–587. Lin, S.-J. (2013). Perceived impact of a documentary film – An investigation of the first-

RI PT

person effect and its implications for environmental issues. Science Communication, 35, 708–733.

Lorenzoni, I., Nicholson-Cole, S., & Whitmarsh, L. (2007). Barriers perceived to engaging

Environmental Change, 17, 445–459.

SC

with climate change among the UK public and their policy implications. Global

M AN U

Lyons, E., & Breakwell, G. L. (1994). Factors predicting environmental concern and indifference in 13- to -16 year-olds. Environment and Behavior, 26, 223–238. Mikami, S., Takeshita, T., & Kawabata, M. (1999). Influence of the mass media on the public awareness of global environmental issues in Japan. Asian Geographer, 18, 87–97.

TE D

Nixon, H., & Saphores, J.M. (2009). Information and the decision to recycle: Results from a survey of US households. Journal of Environmental Planning and Management, 52, 257–277.

EP

Nonami, H., Sugiura, J., Ohnuma, S., Yamakawa, H., & Hirose, Y. (1997). The roles of

AC C

various media in the decision making processes for recycling behavior: a path analysis model. ShinrigakuKenkyu: The Japanese Journal of Psychology, 68, 264–271. Ostman, R. E., & Parker, J. L. (1987). A public’s environmental information sources and evaluations of mass media. The Journal of Environmental Education, 18, 9–17. Pahl, S., Harris, P.R., Todd, H.A., & Rutter, D.R. (2005). Comparative optimism for environmental risks. Journal of Environmental Psychology, 25, 1–11. Podsakoff, P. M., & Organ, D. W. (1986). Self-reports in organizational research: Problems and prospects. Journal of Management, 12, 531–544.

28

ACCEPTED MANUSCRIPT Postmes, T., & Brunsting, S. (2002). Collective action in the age of the Internet – Mass

communication and online mobilization. Social Science Computer Review, 20, 290–301. Sampei, Y., & Aoyagi-Usui, M. (2009). Mass-media coverage, its influence on public awareness of climate-change issues, and implications for Japan’s national campaign to

RI PT

reduce greenhouse gas emissions. Global Environmental Change, 19, 203–212.

Seguin, C., Pelletier, L. G., & Hunsley, J. (1998). Toward a model of environmental activism. Environment and Behavior, 30, 628–652.

SC

Shanahan, J., & McComas, K. (1997). Television’s portrayal of the environment, 1991-1995. Journalism & Mass Communication Quarterly, 74, 147–158.

M AN U

Shanahan, J., Morgan, M., & Stenbjerre, M. (1997). Green or brown? Television and the cultivation of environmental concern. Journal of Broadcasting & Electronic Media, 41, 305–323.

Sharma, S., Durand, R. M., & Gur-Arie, O. (1981). Identification and analysis of moderator

TE D

variables. Journal of Marketing Research, 18, 291–300.

Tabachnick, B. G., & Fidell, L. S. (2001). Using multivariate statistics (4th Ed.). Needham Heights, MA: Allyn&Bacon.

EP

Tranter, B. (2010). Environmental activists and non-active environmentalists in Australia.

AC C

Environmental Politics, 19, 413–429. Valenzuela, S. (2013). Unpacking the use of social media for protest behavior the roles of information, opinion expression, and activism. American Behavioral Scientist, 57, 920– 942.

Wei, R., Lo, V.-H., & Lu, H.-Y. (2007). Reconsidering the relationship between the thirdperson perception and optimistic bias. Communication Research, 34, 665–684. Wei, R., Lo, V.-H., & Lu, H.-Y. (2008). Third-person effects of health news: Exploring the relationships among media exposure, presumed media influence, and behavioral

29

ACCEPTED MANUSCRIPT intentions. American Behavioral Scientist, 52, 261–277.

Weinstein, N. D. (1980). Unrealistic optimism about future life events. Journal of Personality and Social Psychology, 39, 806–820. Whitmarsh, L. (2009). Behavioural responses to climate change: Asymmetry of intentions

AC C

EP

TE D

M AN U

SC

RI PT

and impacts. Journal of Environmental Psychology, 29, 13–23.

30

ACCEPTED MANUSCRIPT Appendix. Constructs and measurement items

RI PT

Environmental activism (Sources: Dono et al., 2010; Seguin et al., 1998) Please indicate how often you take each action a: ACT1 Participate in events organized by environmental groups. ACT2 Give money to an environmental group. ACT3 Circulate petitions demanding an improvement of government policies regarding the environment. ACT4 Participate in protests against current environmental conditions. ACT5

Vote for a government proposing environmentally conscious policies.

ACT6

Write letters to firms that manufacture environmentally harmful products.

SC

Perceived media influence on self (Source: Own elaboration) Please indicate the extent to which environmental media coverage in each of the following media genres has an influence on yourself b: GEN1 Environmental campaigns on TV. Environmental campaigns on the radio.

GEN3

Environmental campaigns in the press.

GEN4

Environmental campaigns in magazines.

GEN5

Environmental campaigns launched online through websites/email.

GEN6

Environmental campaigns launched online through social networks.

GEN7

News seen on TV.

GEN8

News heard on the radio.

GEN9

News read in the press.

GEN10

News read in magazines.

GEN11

News read on websites/emails.

GEN12

News read on social networks.

GEN13

TV ads.

GEN14

Radio ads.

GEN15

Ads in the press.

GEN16

Ads in magazines.

GEN17

Ads shown on websites/emails.

GEN18

Ads shown on social networks.

GEN19

TV documentaries.

GEN20

TV interviews (e.g., with environmental experts).

GEN21

Radio interviews.

GEN22

TV series.

GEN23

Websites specializing in environmental issues (e.g., websites of companies selling environmentally friendly products; webs or blogs promoting pro-environmental behavior) Comments and opinions of close people (e.g., relatives, friends) through social networks

AC C

EP

TE D

M AN U

GEN2

GEN24 GEN25

Comments and opinions of opinion leaders (e.g., singers, sportsmen, celebrities, engineers, doctors, experts) through social networks Optimism bias(Source: Pahl et al., 2005) How much at risk do you think you [a typical person (of your age, sex and background)] are [is] of experiencing harmful effects of each of the following hazards some time in the

31

ACCEPTED MANUSCRIPT future? c: Pollution from factories

OPT2

Pollution from cars

OPT3

Change to ozone caused by pollution

OPT4

Pollution from burning rubbish

OPT5

Fumes/fibers from synthetic materials

OPT6

Acid rain

OPT7

Chemical dumps

OPT8

Water pollution

OPT9

Food contamination*

OPT10

Pesticides and herbicides

OPT11

Genetically modified food*

OPT12

Soil erosion

OPT13

Noise pollution

OPT14

Visual pollution

OPT15

Light pollution*

OPT16

Water shortage

OPT17

Pollution from office equipment

OPT18

Radioactive fallout

OPT19

Germs/microorganisms

SC M AN U

a

RI PT

OPT1

Seven-point Likert-type scale ranging from [1] ‘never’ to [7] ‘always’. Seven-point Likert-type scale ranging from [1] ‘no influence’ to [7] ‘extreme influence’. c Seven-point Likert-type scale ranging from [1] ‘not at risk at all’ to [7] ‘strongly at risk’.

AC C

EP

* Newly developed items.

TE D

b

ACCEPTED MANUSCRIPT Acknowledgements

AC C

EP

TE D

M AN U

SC

RI PT

The authors acknowledge financial support from the Spanish Ministry of Economy and Science and the European Regional Development Fund-ERDF/FEDER (National R&D Project ECO2011-24921) and CySOC.

ACCEPTED MANUSCRIPT Highlights

-

AC C

EP

TE D

M AN U

SC

-

We examine if optimism bias moderates media effects on environmental activism. The results show the differential perceived efficacy of media on reported activism. Internet and print media are most associated with personal action. Optimism bias moderates the effects of TV documentaries/interviews and radio. Optimism bias does not change the effects of the Internet and print media.

RI PT

-