summary articles

Global Environmental Change 26 (2014) 39–52

The ‘‘I’’ in climate: The role of individual responsibility in systematic processing of climate change information

Laura N. Rickard a, Z. Janet Yang b,*, Mihye Seo c, Teresa M. Harrison c

a Department of Environmental Studies, State University of New York College of Environmental Science & Forestry, 108B Marshall Hall, Syracuse, NY 13210,

USA b Department of Communication, State University of New York at Buffalo, 329 Baldy Hall, Buffalo, NY 14260, USA c Department of Communication, University at Albany, State University of New York, 331 Social Science Building, Albany, NY 12222, USA

A R T I C L E I N F O

Article history:

Received 13 August 2013

Received in revised form 6 February 2014

Accepted 17 March 2014

Available online 22 April 2014

Keywords:

Risk Information Seeking and Processing

model

Attribution of responsibility

Risk perception

Information processing

Climate change

A B S T R A C T

Past research suggests that how we perceive risk can be related to how we attribute responsibility for

risk-related issues, such as climate change; however, a gap in research lies in exploring possible

connections between attribution of responsibility, risk perception, and information processing. Using

the Risk Information Seeking and Processing model, this study fills this gap by examining how RISP-

based variables are related to information processing and whether attribution of responsibility for

mitigating climate change influences communication behaviors that are often predicted by elevated risk

perceptions. Undergraduates at two large research universities (N = 572) were randomly assigned to

read one of two newspaper articles that emphasized either individual responsibility (by highlighting

personal actions) or societal responsibility (by highlighting government policy) for climate change

mitigation. Results indicate that subjects in the individual responsibility condition were significantly

more likely to process the message in a systematic manner; however, attribution of responsibility did not

interact with risk perception to influence systematic processing. Moreover, attitudes toward climate

change information and negative affect mediated the relationship between other key variables and

systematic processing. These and other findings suggest that strategic communication about climate

change may benefit from emphasizing individual responsibility to attract more attention from diverse

audiences and to promote deeper thinking about the issue. Additional theoretical implications are

presented.

� 2014 Elsevier Ltd. All rights reserved.

Contents lists available at ScienceDirect

Global Environmental Change

j o u r n a l h o m e p a g e : w w w . e l s e v i e r . c o m / l o c a t e / g l o e n v c h a

1. Introduction

When the U.S. National Oceanic and Atmosphere Administra- tion (NOAA) documented the recent surpassing of 400 parts per million of carbon dioxide in the Earth’s atmosphere, there was no cause for celebration among scientists and policymakers (Gillis, 2013). Instead, the unsettling milestone has provided even further evidence that anthropogenic climate change is occurring, and that current mitigation efforts may be less than adequate. Robust evidence of the detrimental effects of climate change on human and environmental health has motivated scientists and policy- makers alike to inform public audiences about this risk and encourage support for policy and individual behavior change. Yet,

* Corresponding author. Tel.: +1 716 645 1169; fax: +1 716 645 2086.

E-mail addresses: [email protected] (L.N. Rickard), [email protected]

(Z.J. Yang), [email protected] (M. Seo), [email protected] (T.M. Harrison).

http://dx.doi.org/10.1016/j.gloenvcha.2014.03.010

0959-3780/� 2014 Elsevier Ltd. All rights reserved.

due to complex scientific concepts and contentious policy options, communicating with the public about climate change science and mitigation is no simple task (e.g., Pidgeon, 2012; Pidgeon & Fischhoff, 2011; Swim et al., 2011). Contributing to this challenge, culturally situated risk perceptions and political polarization have created segments of the public that are resistant to seeking or receiving this information (e.g., Kahan et al., 2012). The issue is further complicated by conflicting public rhetoric on who (or what) is responsible for mitigating the impacts of climate change on local and distant people and places (e.g., Olausson, 2011; Stoddart et al., 2012).

As decades of research in social psychology and communication have established, when we process information in a systematic manner (i.e., engaging in more cognitive resources, relying on rational judgment and analytical thinking), we are more likely to form stronger, more long-lasting attitudes about the topic, as well as intentions to engage in related behaviors (Griffin et al., 2002; Yang et al., 2012). Therefore, in the context of climate change

L.N. Rickard et al. / Global Environmental Change 26 (2014) 39–5240

communication, understanding how individuals come to process information in this way would be beneficial to encouraging certain normative (e.g., pro-environmental) attitudes and behaviors. The Risk Information Seeking and Processing (RISP) model (Griffin et al., 1999) illustrates how factors such as perception of risk, and the extent to which one feels pressure to stay informed about a topic, predict systematic information processing. Substantial gaps in research, however, lie in how risk perceptions, and other predictor variables included in the RISP model, might interact with attribution of responsibility for responding to an issue in ways that inform subsequent information processing. This study fills this gap by examining whether attribution of responsibility for alleviating a risk issue influences communication behaviors that are often predicted by elevated risk perceptions.

Findings suggest that attitudes toward climate change infor- mation and negative affect mediated the relationship between several of the RISP variables and systematic processing. Attribution of responsibility, however, exerted influence on systematic processing independent of the mechanisms proposed in the RISP model; those presented with a message framing climate change mitigation as an individual responsibility were more likely to process the information systematically. Following recent scholar- ship in science communication (Nisbet, 2009), practical implica- tions are drawn for designing message strategies to encourage systematic thinking about climate change and, in turn, to influence support for relevant policy and behavioral intentions. An agenda for future research is presented.

2. Background

2.1. RISP model

The RISP model was designed to examine factors that influence information seeking and processing with respect to health and environmental risk issues. Since the seminal work (Griffin et al., 1999), the RISP model has been used in a variety of contexts with respect to personal health risks related to drinking tap water and eating fish from the U.S. Great Lakes (Griffin et al., 2004), environmental risks related to the health of the Great Lakes ecosystem (Kahlor et al., 2002), economic risks such as damages from repeated flooding of an urban river (Griffin et al., 2008) and the energy crisis (Griffin et al., 2005), and clinical trial enrollment for both cancer patients (Yang et al., 2010a) and healthy volunteers (Yang et al., 2011). To date, the RISP model has shown its applicability in explicating Risk Information Seeking and Proces- sing in both personal and impersonal risk contexts, as in the case of climate change (Kahlor, 2007, 2010).

As a predictive, theoretical model, RISP is structured primarily on the heuristic-systematic model (HSM), which explicates two distinct types of information processing and the circumstances in which they are utilized. First, heuristic processing, as Eagly and Chaiken (1993) describe, is ‘‘a limited mode of information processing that requires less cognitive effort and fewer cognitive resources’’ (p. 327). For example, individuals who engage in heuristic processing may rely on general decision rules such as ‘‘experts are always to be trusted’’ when asked to evaluate new scientific evidence related to climate change and accept the new information as valid. Systematic processing, by comparison, is a ‘‘relatively analytic and comprehensive treatment of judgment- relevant information’’ (Chen & Chaiken, 1999, p. 74). For example, when asked to make the same judgment, individuals who engage in systematic processing will evaluate the validity of the information based on its merit, rather than accepting the evidence just because it is presented by experts.

Even though systematic processing is considered less superfi- cial, a heuristic strategy has the mental and economic advantage of

requiring minimal cognitive effort (Chaiken, 1980). Therefore, people tend to engage in heuristic processing unless motivated to adopt the more effortful strategy. However, as Chaiken (1980) points out, a heuristic strategy may be less reliable when used to judge message validity because an overreliance on simple decision rules may lead recipients to accept conclusions they might otherwise reject had they invested the time and cognitive resources to discover and scrutinize different arguments (p. 753).

To date, most of the RISP-based research has focused on information seeking (Griffin et al., 2008; Kahlor, 2007, 2010; Yang & Kahlor, 2013; Yang et al., 2013), with limited attention paid to social cognitive predictors of information processing. This study addresses this absence of research by focusing specifically on how various social cognitive variables, as included in the RISP model, motivate individuals to engage in the more rational, analytical, and deliberate form of information processing (systematic processing) when confronting information about climate change mitigation.

2.1.1. Information insufficiency

Consistent with other dual-process models, which delineate two parallel, but distinct, forms of information processing systems, the HSM argues that people tend to adopt a form of processing for a given message based on (1) their capacity to process information in each manner, and (2) their motivation to go beyond more superficial (heuristic) processing to engage in systematic proces- sing. Eagly and Chaiken observe that, ‘‘people will exert whatever effort is required to attain a ‘sufficient’ degree of confidence that they have accomplished their processing goals’’ (sufficiency principle) (Eagly & Chaiken, 1993, p. 330). The personal relevance of the topic, for instance, elevates the amount of judgmental confidence people desire. Communication researchers have elaborated on the sufficiency principle and proposed in the RISP model that the drive to overcome information insufficiency (i.e., the need to gain and hold enough information to feel confident in one’s existing attitude) motivates individuals to process information more systematically (Griffin et al., 1999). Thus, this study first examines this relationship, and hypothesizes that, controlling for current knowledge, information sufficiency threshold will be positively related to systematic processing.

2.1.2. Risk judgment and perceived salience

Within the RISP model, risk perception is a key factor that elevates information insufficiency. Largely informed by the psychometric model of risk perception (Slovic, 1987), the RISP model maintains that an individual’s cognitive evaluation of the potential risk (referenced in the model as perceived hazard characteristics) elevates his or her affective response to the risk, which will increase information insufficiency (Hovick et al., 2011; ter Huurne et al., 2009) and subsequently lead to more systematic processing. Past research has tested these indirect linkages through multiple regression analyses, showing a significant relationship between information insufficiency and systematic processing first, and then showing that perceived hazard characteristics and affective responses are significant predictors of information insufficiency (Griffin et al., 2008). Fig. 1 displays these relationships.

Other studies have found that instead of exerting an indirect effect through information insufficiency, affective responses to an issue can influence systematic processing directly (Hovick et al., 2011; Yang et al., 2010b). Thus, in addition to examining the indirect influence through information sufficiency threshold, this study will focus on the indirect effect that perceived hazard characteristics have on systematic processing through negative affect. Specifically, our second hypothesis states that, controlling for current knowledge, perceived hazard characteristics – per- ceived salience of climate change and risk judgment – will

Fig. 1. Risk Information Seeking and Processing model with hypotheses and research questions examined in this study. Source: Figure adapted from Griffin et al. (2012).

L.N. Rickard et al. / Global Environmental Change 26 (2014) 39–52 41

influence information sufficiency threshold through negative affect. Our third hypothesis is that, controlling for current knowledge, perceived salience and risk judgment will influence systematic processing indirectly through negative affect and information sufficiency threshold.

2.1.3. Informational subjective norms

The degree to which an individual feels social pressure to stay informed (informational subjective norms) also influences informa- tion insufficiency (Griffin et al., 2004) and at times, directly motivates information processing (Griffin et al., 2008; Yang et al., 2010b). Thus, our fourth hypothesis suggests that, controlling for current knowledge, informational subjective norms will be positively related to information sufficiency threshold and to systematic processing.

2.1.4. Perceived information gathering capacity and attitude toward

information

While information insufficiency primarily accounts for an individual’s motivation to process information, perceived behav- ioral control (perceived information gathering capacity) and attitudes toward information sources (relevant channel beliefs) moderate the relationship between information insufficiency and information processing. For instance, greater ability to understand important risk information can enhance systematic processing (Griffin et al., 2008), whereas trust in an information source, albeit often considered a heuristic cue, can also encourage systematic processing when the processing task is important to the individuals (Yang et al., 2010b). On this basis, we state our fifth hypothesis, that perceived information gathering capacity will be positively related to systematic processing, and our sixth hypothesis, that attitude toward learning about climate change (i.e., information seeking) will be positively related to systematic processing.

2.2. Attribution of responsibility

Attribution theory tells us how individuals come to attribute both the causes of and the responsibility for phenomena (Heider, 1982). The observations we make also influence how we determine the locus of causality of an event: internal (‘‘dispositional’’), brought about by actions or characteristics of the individual, or external (‘‘situational’’), by forces outside of the person (Shaver, 1985). Whereas causal attribution focuses on the occurrence of past events, treatment attribution (Brickman et al., 1982; Iyengar, 1989, 1990) looks to the future to assign responsibility for responding to such events; in essence, ‘‘questions of treatment

responsibility seek to establish what can be done to prevent recurrence of the outcome’’ (Iyengar, 1990, p. 23). In experimental studies in which subjects viewed news clips discussing societal problems such as urban poverty, Iyengar (1990) elicited treatment attributions by posing the question, ‘‘If you were asked to prescribe ways to reduce poverty, what would you suggest?’’ Unsurprisingly, subjects’ answers involved naming certain ‘‘agents of causation’’ (Iyengar, 1989, 1990), ranging from an individual him or herself to a government agency deemed responsible for, and in control of, the problem at hand. Whereas an internal attribution denotes personal responsibility to respond to a particular problem, an external attribution holds an external source, such as an institution, accountable for solving the problem. Using the issue of climate change mitigation, this study focuses exclusively on treatment attribution.

2.2.1. Message framing

How individuals come to attribute responsibility for the cause of or for responding to an issue can be influenced, in part, by the nature of the communication they encounter about it. Scholarship in communication and social psychology has established that framing of causal and treatment responsibility can influence support for relevant policy and behavioral intentions (Nisbet, 2009; Tversky & Kahneman, 1981; Weiner, 2006). For instance, experimentally manipulated episodic or thematic framing of the impacts of climate have been shown to influence support for policy; individuals exposed to the thematic framing were more likely to express support for government policies to reduce the effects of climate change (Hart, 2011). Other research has shown that messages highlighting in-group (i.e., U.S.) or out-group (i.e., China) responsibility for causing climate change can affect respondents’ support for domestic or global-level policy (Jang, 2013). Recent studies further suggest that partisanship may increase the degree of political polarization observed among the public in response to climate mitigation policies. In particular, among a U.S. sample, exposure to messages highlighting the health risks of climate change and distant victims increased Democrats’ support for climate change mitigation policy, while decreasing Republicans’ support (Hart & Nisbet, 2012).

2.2.2. Attribution of responsibility and perception of risk

In comparison to the amount of literature addressing how framing influences attribution, few studies have investigated the linkages between attribution of responsibility and perception of risk. Of those that have, the majority have been qualitative in nature, and, rather than focusing on cognitive processes, address the institutional-level implications of attributing responsibility for

L.N. Rickard et al. / Global Environmental Change 26 (2014) 39–5242

responding to the risks associated with an increasingly industrial- ized society (e.g., Bickerstaff et al., 2008; Bickerstaff & Walker, 2002; Petts, 2005). While critical to theoretical development, such studies are less helpful in suggesting the operationalization of and predictive relationships between variables, as is done in the current study.

Turning to the studies with an individual-level focus, the existing literature focuses on the relationship between risk perception and causal attribution. For instance, in a study of visitors to three U.S. national parks, Rickard (2013) demonstrated that, as visitors perceived park-related risk as more controllable, they were more likely to attribute the cause of a hypothetical visitor accident to characteristics of the victim (i.e., internal attribution), rather than characteristics of the park or park management (i.e., external attribution). A handful of studies consider perception of risk with respect to attributing the cause of AIDS (Mannetti & Pierro, 1991) or accidental injury (Kouabenan, 1998, 2002; Peltzer & Renner, 2003; Sellstrom et al., 2000). Surveying the public about a waterborne disease outbreak in the Milwaukee water supply, Kahlor et al. (2002) showed that the respondents’ personal risk estimates varied depending on their health outcomes, which, in turn, affected causal attributions. Specifically, elevated risk perceptions linked to becoming ill led individuals to attribute the risk to external sources (i.e., the city makes sure that the residents do not get sick), whereas lowered risk perceptions linked to avoiding illness led to internal causal attribution (i.e., I buy bottled water) (Kahlor et al., 2002). In another study, Griffin et al. (2008) found that attributing the cause of flooding losses to poor government management was signifi- cantly related to experiencing anger at the agencies, which was subsequently related to systematic processing of flood risk information. Although neither Kahlor et al. (2002) nor Griffin et al. (2008) established direct relationships between attribution and information processing, Griffin et al. (2008) posited that causal attribution might be viewed as part of perceived hazard character- istics since it influenced individuals’ general perception of risk.

While the studies reviewed above hinted at a relationship between risk perception, causal attribution and information processing, none of them has considered treatment attribution – i.e., the responsibility for responding to a given event or condition. In response to the impact of climate change, however, treatment attribution is a crucial factor because it affects the general public’s support for mitigation policies and intention to change individual behaviors. To begin to address these gaps in the literature, instead of stating hypotheses (as was done above, with respect to the RISP model), we pose two broad, exploratory research questions. First, we seek to ascertain whether subjects in the internal and external treatment attribution conditions will process climate change mitigation information differently. Second, we wish to determine how attribution of treatment responsibility interacts with perceived salience and risk judgment to affect information processing.

To review, the main objective of this study is two-fold. First, guided by the RISP model, we examine the relative impact of key RISP variables on systematic processing of information about climate change mitigation. Second, we explore attribution of treatment responsibility within the RISP framework, aiming to reveal whether treatment attribution interacts with perceived hazard characteristics to influence systematic processing, as suggested by Griffin et al. (2008). Fig. 1 illustrates the hypotheses and research questions examined in this study.

3. Method

Data were collected in online survey format through Qualtrics at two large research universities in the Northeast U.S. Research

subjects were recruited from large entry-level communication classes with students pursuing different subjects of study who received either research credit or extra credit for their participa- tion. Institutional Review Board approval was obtained at each university. To minimize burden on the research subjects and reduce sensitization, baseline risk perceptions and demographic information were measured in a prescreening survey from April 1 to 8, 2013. A week later (April 15–26), the experimental manipulation and the assessment of other key variables were carried out in the main study. In the main study, research subjects first read the simulated news story, next answered questions measuring systematic processing, and lastly, answered questions assessing other variables. A total of 618 respondents filled out the prescreening survey, and 92.6% of them completed the main study, resulting in a final sample size of 572.

3.1. Experimental manipulation

To manipulate internal vs. external attribution of treatment responsibility for climate change mitigation, research subjects were randomly assigned to read a news story that attributed the responsibility either to individual behavior – using less air conditioning in the summer, or to government policy – enacting a carbon tax. Both stories were formatted as a U.S.-focused, headline story published in the New York Times on Thursday, September 20, 2012, with a similar lead that described the heat wave experienced by most Americans in the summer of 2012. The two stories were similar in length and layout (see Appendices A and B). Although subjects were not told explicitly whether the story had appeared in the newspaper, the font and formatting used were intended to mimic those employed by the New York Times. Subjects rated the two stories similarly as to whether the story was credible, interesting, informative, realistic, and relevant to their life. A manipulation check indicated that the two stimuli successfully manipulated research participants’ attribution of treatment responsibility to either individual responsibility or government policy.

3.2. Measures

Table 1 shows item wording and descriptive data for all items and indices.

Systematic processing was measured based on ten 7-point scales ranging from 1 (strongly disagree) to 7 (strongly agree), in which respondents were asked to indicate how they processed the information presented in the news article (Kahlor et al., 2003). A factor analysis with principal axis factoring extraction and varimax rotation (KMO = .85, Bartlett’s X2 = 1886.77, p < .001) indicated that these items loaded onto two factors – systematic and heuristic processing. Based on a satisfactory reliability check (a = .84), the five items assessing systematic processing were averaged into an index. We also measured the amount of time each participant spent reading the news article. Among all the valid measures of reading time (between 60 s and 300 s, based on the word count of the article and average reading speed observed from pilot testing of the stimuli), participants in the internal attribution (M = 150.16, SD = 97.71, n = 119) condition spent significantly longer time reading the article than participants in the external attribution (M = 131.95, SD = 87.32, n = 131) condition (90% confidence inter- val: [1.38, 24.22]).

Information insufficiency was measured using two items on 0– 100 scales, in which respondents were asked to estimate their current knowledge of climate change from 0 (knowing nothing) to 100 (knowing everything) and their need for information from 0 (need to know nothing) to 100 (need to know everything). Information insufficiency was assessed by controlling the impact

Table 1 Descriptive data for key variables.

Concept Measures Internal External

M SD M SD

Systematic processing (1–7 scale) I thought about what actions I myself might take based on what I

read.

4.98 1.22 4.63 1.33

I found myself making connections between the story and what

I’ve read or heard about elsewhere.

5.00 1.30 4.89 1.26

I tried to think about the importance of the information for my

daily life.

5.10 1.23 4.89 1.35

I thought about how the story related to other things I know. 4.98 1.06 4.83 1.20

I tried to relate the ideas in the story to my own personal

experiences.

4.94 1.18 4.40 1.32

Averaged index 5.03 .95 4.40 1.32

Information (in)sufficiency (0–100 scales) Current knowledge:

Please estimate your knowledge of climate change on a 0–100

scale, where 0 means knowing nothing and 100 means knowing

everything you could possibly know about the topic. How much do

you think you currently know?

49.24 20.38 48.17 20.42

Sufficiency threshold:

This time, using that same scale, please estimate how much you

think you NEED to know about climate change.

74.15 22.51 72.31 22.54

Perceived hazard characteristics (varied scales) Salience (1–7 scale):

To me, the topic of climate change is important.

5.34 1.35 5.23 1.37

To me, the topic of climate change is of interest. 4.92 1.41 4.74 1.40

To me, the topic of climate change is relevant. 5.32 1.30 5.29 1.35

Averaged index 5.20 1.22 5.08 1.21

Risk judgment:

How much do you think climate change will harm (1–4 scale):

You and your family.

2.66 .86 2.62 .92

Your local community. 2.74 .85 2.71 .90

The U.S. as a whole. 3.17 .80 3.16 .83

People all over the world. 3.36 .70 3.33 .78

Nature (not including human). 3.52 .70 3.50 .75

How serious of a threat is climate change to (1–6 scale):

You and your family.

3.60 1.33 3.45 1.40

Your local community. 3.50 1.35 3.46 1.40

The U.S. as a whole. 4.13 1.26 4.10 1.39

People all over the world. 4.38 1.25 4.33 1.35

Nature (not including human). 4.77 1.20 4.71 1.36

Averaged index 13.38 5.35 13.21 5.87

Negative affect (1–7 scale) Information about climate change worries me. 4.86 1.42 4.80 1.52

Information about climate change makes me feel sad. 4.46 1.48 4.24 1.56

Information about climate change makes me feel anxious. 4.38 1.47 4.16 1.55

Information about climate change makes me feel guilty. 4.32 1.52 3.99 1.49

Averaged index 4.50 1.25 4.30 1.33

Informational subjective norms (1–7 scale) My friends expect me to know something about climate change. 3.44 1.58 3.39 1.54

Most people who are important to me think I should know

something about climate change.

3.84 1.51 3.72 1.51

My family expects me to know something about climate change. 3.67 1.53 3.53 1.53

Averaged index 3.67 1.39 3.54 1.35

Information gathering capacity (1–7 scale) I can’t make sense of the information about climate change

(reverse coded).

5.06 1.37 5.00 1.43

When it comes to information about climate change, I don’t know

how to separate facts from fiction (reverse coded).

4.29 1.47 4.21 1.49

Most information about climate change is too technical for me to

understand (reverse coded).

4.75 1.37 4.65 1.51

I can’t understand information about climate change even if I

make an effort (reverse coded).

5.23 1.39 5.20 1.44

Averaged index 4.84 1.16 4.77 1.22

Attitude (1–7 scale) Learning about climate change is useful. 5.60 1.24 5.62 1.16

Learning about climate change is beneficial. 5.61 1.20 5.60 1.21

Learning about climate change is wise. 5.62 1.20 5.60 1.16

Learning about climate change is valuable. 5.67 1.13 5.59 1.17

Averaged index 5.62 1.11 5.60 1.10

L.N. Rickard et al. / Global Environmental Change 26 (2014) 39–52 43

of current knowledge on information sufficiency threshold (Griffin et al., 2008).

Perceived hazard characteristics were assessed with two groups of measures – perceived salience (Yang et al., 2013) and risk judgment (Zhao et al., 2011). Salience was assessed with three items on 7-point scales, which were averaged into an index (a = .87). Risk judgment was assessed with both the perceived likelihood that climate change will harm various groups ranging from ‘‘you and your family’’ to ‘‘nature’’ (4-point scales ranging

from 1 = not at all to 4 = a great deal) and the perceived severity of the threat to these groups (6-point scales ranging from 1 = not at all serious to 6 = very serious). Product terms were created based on these two dimensions and averaged into an index to assess risk judgment (a = .92). Together, perceived salience and risk judgment represented the catastrophic potential dimension of the ‘‘dread risk’’ factor in the psychometric paradigm (Slovic, 1987) and a subset of the variables used to measure cognitive risk perceptions related to climate change (Leiserowitz, 2006).

L.N. Rickard et al. / Global Environmental Change 26 (2014) 39–5244

Negative affect was measured using four items on 7-point scales from 1 (strongly disagree) to 7 (strongly agree), assessing the degree to which information about climate change made the respondents feel worried, sad, anxious, and/or guilty about climate change. In addition to the specific negative emotions assessed in previous research on climate change (Yang & Kahlor, 2013), guilt was assessed as representing a specific moral emotion (Tangney et al., 2007). Upon reliability check (a = .88), all four emotions were averaged into an index.

Informational subjective norms were measured with three items on 7-point scales from 1 (strongly disagree) to 7 (strongly agree), in which respondents were asked to indicate the extent to which they agreed with three statements that depicted others’ expectations about their knowledge of climate change (Yang et al., 2013). These items were checked for reliability (a = .88) and averaged into an index.

Perceived information gathering capacity was measured with four items on 7-point scales from 1 (strongly disagree) to 7 (strongly agree), in which respondents were asked to indicate the extent to which they found it difficult to understand information about climate change (Yang et al., 2013). These items were reverse coded and averaged into an index upon satisfactory reliability check (a = .85).

Attitude toward climate change information was measured with four items on 7-point scales from 1 (strongly disagree) to 7 (strongly agree), assessing the degree to which respondents believed that climate change information was useful/beneficial/ wise/valuable (Kahlor, 2007). These items were checked for reliability (a = .95) and averaged into an index.

Individual characteristics, including demographic variables, religion, political philosophy, were used as control variables. Past research has shown that gender (McCright & Dunlap, 2011a), ethnicity (Leiserowitz, 2006), political philosophy (McCright & Dunlap, 2011b), and religion (Posas, 2007) often influence individuals’ risk perceptions related to climate change. Thus, we measured gender (1 = male, 2 = female), ethnicity (0 = non-White/ minority, 1 = White), and annual household income (1 = under $20,000, 15 = above $150,000, M = $81,100, SD = $44,200). Religion was measured as a binary variable (0 = non-Christian, 1 = Chris- tianity). Political party was measured as a nominal variable (15.1% Republican, 45.1% Democrat, 17.5% Independent, and 21.5% not interested in politics, 0.8% other such as libertarian). For the regression analysis, this variable was dummy-coded into three categories: Republican Party, Democratic Party, and Interested/Not interested in politics, with Independents as the reference group. Political philosophy was measured on a continuous scale from 1 (very liberal) to 5 (very conservative) (M = 2.61, SD = .96).

3.3. Analysis

We used hierarchical ordinary least squares (OLS) regression to test our hypotheses. Hierarchical OLS allows the researchers to enter variables in a series of blocks with the results at each step indicating the relative influence of the variables on the dependent variable while controlling for variables entered in previous steps (Cohen et al., 2003). For the models with information sufficiency threshold as the dependent variable, demographic variables were entered in the first block, followed by current knowledge in block 2, perceived hazard characteristics in block 3, negative affect in block 4, and informational subjective norms in block 5. For models with systematic processing as the dependent variable, information sufficiency threshold was added in block 6, perceived information gathering capacity was entered in block 7, and attitude toward climate change information was entered in the final block. To ensure that the assumptions for the hierarchical OLS analysis were not violated, we performed multi-collinearity tests, and found all

resulting tolerance values well above zero and VIF values (all lower than 2) well below the conventional cut-off value of 10 (Cohen et al., 2003). Unstandardized regression coefficients for the pooled sample and each experimental condition are presented in Table 2. Using the PROCESS macro (Hayes, 2013) for SPSS, mediation analysis with a bootstrapping approach was conducted to test H2a/ b and H3a/b; moderation analysis was conducted to examine RQ2.

4. Results

4.1. Relationship between information sufficiency threshold and

systematic processing

The first hypothesis stated that, controlling for current knowl- edge, information sufficiency threshold would be positively related to systematic processing. Regression results showed that neither current knowledge nor information sufficiency threshold was significantly related to systematic processing in either condition.

4.2. Relationship between perceived salience, risk judgment, and

information sufficiency threshold

The second hypothesis stated that perceived salience and risk judgment would influence information sufficiency threshold indirectly through negative affect. Regression results showed that in the pooled sample, controlling for current knowledge, both perceived salience (B = 2.44, p < .05) and risk judgment (B = .40, p < .05) were significantly related to information sufficiency threshold. Negative affect was significantly related to information sufficiency threshold in both conditions and in the pooled sample (B = 3.63, p < .001). However, risk judgment was significantly related to information sufficiency threshold only in the internal attribution condition, while perceived salience was not significantly related to information sufficiency threshold in either condition. Mediation analysis suggested that negative affect mediated the relationship between salience and information sufficiency threshold (indirect effect: 1.88, 95% CI [1.09, 2.77], Fig. 2a) and between risk judgment and information sufficiency threshold (indirect effect: .41, 95% CI [.25, .62], Fig. 2b). These mediation effects, however, were not moderated by the experimental condition.

4.3. Relationship between perceived salience, risk judgment, and

systematic processing

The third hypothesis stated that perceived salience and risk judgment would influence systematic processing through negative affect and information sufficiency threshold. Mediation analysis showed that only negative affect mediated the relationship between salience and systematic processing (indirect effect: .08, 95% CI [.04, .13], Fig. 3a) and the relationship between risk judgment and systematic processing (indirect effect: .02, 95% CI [.01, .03], Fig. 3b). Again, these mediation effects were not moderated by the experimental condition. Negative affect was also positively related to systematic processing in both the pooled sample (B = .20, p < .001) and in both experimental conditions.

4.4. Relationship between informational subjective norms,

information sufficiency threshold, and systematic processing

The fourth hypothesis suggested that controlling for current knowledge, informational subjective norms would be positively related to information sufficiency threshold and systematic processing; however, regression results showed that informational subjective norms were not a significant predictor of information insufficiency or systematic processing.

Table 2 Predictors of systematic processing of climate change information.

Information sufficiency threshold Systematic processing

Unstandardized coefficients Unstandardized coefficients

Pooled sample Internal External Pooled sample Internal External

Block 1: Individual characteristics

Female 1.01 �2.58 4.96 .05 .03 .06 White 2.01 .39 4.37 .09 �.01 .16 Christian �1.16 �1.98 �1.07 .23 .19* .10 Republican �4.80 �5.65 �3.43 �.23* �.10 �.39**

Democrat .66 �2.05 4.21 �.08 .01 �.18 Political philosophy �.14 �.29 �.36 .06 .10 .02 Household income �.55* �.39 �.71* .00 �.00 �.00

DR2 .05 .05 .07 .05 .06 .08 Block 2: Current knowledge .29*** .24*** .37*** .00 .00 .00

DR2 .09 .06 .14 .03 .03 .03 Block 3: Perceived hazard characteristics

Perceived salience 2.44* 2.12 2.49 .00 �.06 .04 Risk judgment .40* .63* .21 �.01 .01 �.02*

DR2 .06 .09 .04 .07 .06 .07 Block 4: Negative affect 3.63*** 4.68** 2.73* .20*** .12* .24***

DR2 .02 .03 .01 .15 .12 .16 Block 5: Informational subjective norms �1.29 �1.41 �1.02 .00 .02 �.01

DR2 .01 .01 .00 .00 .00 .00 Block 6: Information sufficiency threshold .00 �.00 .00

DR2 .01 .00 .01 Block 7: Information gathering capacity .02 .02 .02

DR2 .01 .02 .00 Block 8: Attitudes toward climate change

information

.39*** .47*** .34***

DR2 .11 .16 .09 Adjusted R2 .22 .20 .24 .41 .42 .42

ANOVA F12,512 = 13.16 *** F12,247 = 6.28

*** F12,252 = 7.79 *** F15,487 = 23.97

*** F15,234 = 12.90 *** F15,237 = 12.94

***

Chow statistics F13,498 = 0.83 F16,470 = 2.46 **

Note. *p < .05, **p < .01, and ***p < .001.

Fig. 2. (a) Unstandardized regression coefficients for the relationship between perceived salience and information sufficiency threshold as mediated by negative affect. The coefficient between perceived salience and information sufficiency threshold controlling for negative affect is in parenthesis. Note. *p < .05, **p < .01, and ***p < .001. (b)

Unstandardized regression coefficients for the relationship between risk judgment and information sufficiency threshold as mediated by negative affect. The coefficient

between risk judgment and information sufficiency threshold controlling for negative affect is in parenthesis. Note. *p < .05, **p < .01, ***p < .001.

L.N. Rickard et al. / Global Environmental Change 26 (2014) 39–52 45

Fig. 3. (a) Unstandardized regression coefficients for the relationship between perceived salience and systematic processing as mediated by negative affect. The coefficient between perceived salience and systematic processing controlling for negative affect is in parenthesis. Note. *p < .05, **p < .01, and ***p < .001. (b) Unstandardized regression

coefficients for the relationship between risk judgment and systematic processing as mediated by negative affect. The coefficient between risk judgment and systematic

processing controlling for negative affect is in parenthesis. Note. *p < .05, **p < .01, and ***p < .001.

L.N. Rickard et al. / Global Environmental Change 26 (2014) 39–5246

4.5. Relationship between information gathering capacity and

systematic processing

The fifth hypothesis stated that perceived information gather- ing capacity would be positively related to systematic processing; regression results showed no support for this hypothesis.

4.6. Relationship between attitudes toward climate change

information and systematic processing

The sixth hypothesis stated that attitudes toward climate change information would be positively related to systematic processing, and was supported in both the pooled sample (B = .39, p < .001) and in each experimental condition. Attitudes toward climate change information exerted strong, direct influence on systematic processing. Finally, mediation analysis indicated that attitudes toward climate change information mediated the relationship between information sufficiency threshold and systematic processing (indirect effect: .009, 95% CI [.006, .012], Fig. 4a), capacity and systematic processing (indirect effect: .10, 95% CI [.06, .15], Fig. 4b), and informational subjective norms and systematic processing (indirect effect: .10, 95% CI [.07, .14], Fig. 4c).

4.7. Relationship between attribution of responsibility and systematic

processing

We also explored the relationship between attribution of treatment responsibility and systematic processing. We found that participants in the internal attribution condition were more likely to systematically process the information presented in the news article than those in the external attribution condition (t[557] = 3.66, p < .001). As mentioned above, attribution of treatment responsibility did not interact with perceived salience,

risk judgment, or any other individual characteristics or indepen- dent variable to influence systematic processing.

4.8. Effect of individual characteristics

Among the individual characteristics, respondents who identified themselves as Christian (including Protestants, Catholics, and other denominations) were more likely to process climate change information systematically in the internal attribution condition, whereas Republican respondents were less likely to process information systematically in the external attribution condition and in the pooled sample. Regression results also showed that, controlling for existing knowledge, respondents with lower house- hold income reported a higher need for climate change information in the external attribution condition and the pooled sample.

5. Discussion

Contributing to the emerging body of literature applying the RISP model within the context of climate change (e.g., Kahlor, 2007; Yang & Kahlor, 2013), this study examined predictors of systematic processing of information related to climate change mitigation. Findings suggest that attitudes toward climate change information and negative affect mediated the relationship between several other RISP variables and systematic processing. Attribution of treatment responsibility, however, exerted influence on systematic processing independent of the mechanisms proposed in the RISP model.

5.1. Theoretical implications

Many of our results support relationships already outlined by the RISP model; however, our findings also extend existing RISP research by revealing the mediating roles that negative affect and

Fig. 4. (a) Unstandardized regression coefficients for the relationship between information sufficiency threshold and systematic processing as mediated by attitude toward climate change information. The coefficient between information sufficiency threshold and systematic processing controlling for attitude toward climate change information

is in parenthesis. Note. *p < .05, **p < .01, ***p < .001. (b) Unstandardized regression coefficients for the relationship between perceived information gathering capacity and

systematic processing as mediated by attitude toward climate change information. The coefficient between perceived information gathering capacity and systematic

processing controlling for attitude toward climate change information is in parenthesis. Note. *p < .05, **p < .01, and ***p < .001. (c) Unstandardized regression coefficients

for the relationship between informational subjective norms and systematic processing as mediated by attitude toward climate change information. The coefficient between

informational subjective norms and systematic processing controlling for attitude toward climate change information is in parenthesis. Note. *p < .05, **p < .01, and

***p < .001.

L.N. Rickard et al. / Global Environmental Change 26 (2014) 39–52 47

attitudes toward climate change information play between other predictors and the dependent variable, systematic processing. While the RISP model posits that negative affect mediates the relationship between perceived hazard characteristics and infor- mation insufficiency, the mediating role that attitudes toward climate change information serve is unexpected. In particular, past research has often concluded that certain variables, such as current knowledge, information sufficiency threshold, and perceived information gathering capacity, were not significant predictors of information seeking or processing (Kahlor, 2007; Yang et al., 2011). Mediation analysis conducted in this study suggests that, on the contrary, these variables may exert indirect effects on the dependent variables.

These findings suggest that respondents’ general attitudes toward climate change, especially their emotional responses, play a central role in shaping their communication behaviors related to climate change. That is, even though risk perceptions elevate information insufficiency to some extent, it is the negative affect induced by these risk perceptions that make the respondents want to know more about climate change. This result complements research demonstrating the important, but complex role of negative affect in promoting support for climate change policy

(Leiserowitz, 2006), or, alternatively, in spurring fatalism or disengagement from the issue (O’Neill & Nicholson-Cole, 2009). Further, when both motivation (information insufficiency and informational subjective norms) and capacity are accounted for, it is the respondents’ general evaluation of climate change informa- tion that leads them to process information presented in the simulated news story in a systematic manner. Along with the finding that respondents who read the internal attribution article were more likely to process the message systematically, these results suggest that it is possible to identify an audience who is receptive to climate change mitigation information and design messages that are appealing to them, both of which are promising for the future of climate change communication. In this way, our study adds to recent calls for audience segmentation research with respect to climate change messaging (e.g., Maibach et al., 2008).

Nonetheless, it is possible that affect and attitudes mediated the relationship between other key variables and systematic proces- sing due to the specific research context – i.e., the topic of climate change. Given that climate change is a contentious topic and a partisan issue in the U.S., individuals’ decision to process information about climate change mitigation policy is likely influenced by their general attitudes and feelings toward the topic.

L.N. Rickard et al. / Global Environmental Change 26 (2014) 39–5248

In particular, one surprising finding is that informational subjec- tive norms, which have exerted direct influence on systematic processing in various other research contexts (Griffin et al., 2008; Yang et al., 2010b), was not a significant predictor in this study. A plausible explanation is that subjective norms regarding one’s information level about climate change constitute a general perception. In contrast, we measured systematic processing in association with the specific simulated news stories presented in this study. Thus, the more general perception of subjective norms may have influenced systematic processing by triggering other general evaluative processes such as attitudes and affect toward climate change. In support of this explanation, Kahlor et al. (2003) also found that informational subjective norms were not significantly related to systematic processing of a simulated magazine article. Thus, it appears that informational subjective norms are a significant predictor of information processing when the latter is measured as a general tendency, not as a specific activity. For instance, Griffin et al. (2008) measured systematic processing with statements such as ‘‘If I need to act on this matter, the more viewpoints I get the better.’’ This measurement strategy differs from the one used in the current study, which was related to the specific news story that the subjects read.

While attribution of treatment responsibility did not moderate the relationship between risk perception and systematic proces- sing, individuals in the internal treatment attribution condition were more likely to process the message systematically than those in the external attribution condition. That is, individuals who read the news story portraying climate change mitigation as an individual responsibility were more likely to process the message in an in-depth manner than those who read the story portraying the responsibility as societal and policy-based. This finding poses important theoretical implications. Distinct from what Griffin et al. (2008) proposed, attribution of treatment responsibility did not influence systematic processing through risk perceptions or negative affect. Instead, attribution of responsibility exerted independent influence on systematic processing. Thus, there is probably inadequate evidence to consider attribution of treatment responsibility as constituting one’s risk perceptions related to climate change.

In discussing the relationship between systematic processing and an individual responsibility frame, two caveats deserve mention. First, we based this study on the theoretical premise that systematic processing represents a desirable outcome, as it foregrounds stronger, and longer-lasting attitudes and behaviors with respect to a given topic; in so doing, we also assumed that the resulting attitudes and behaviors would align normatively with what the communicator (and the message) intended. Emerging social psychological research on the concept of motivated reasoning, however, challenges this rather simplistic assumption of the message frame producing, in all instances, an intended audience effect. When evaluating information on a salient topic, such as climate change, individuals with prior strongly-held beliefs (e.g., climate change skeptics) may exercise bias in processing information; that is, even when presented with balanced information, they may selectively attend more to evidence that supports their stance on an issue, and less to evidence that challenges it (Ditto et al., 1998; Druckman & Bolsen, 2011; Taber & Lodge, 2006). Motivated reasoning, in turn, may foster increased attitude polarization, as Hart and Nisbet (2012) found in their experimental study of support for climate change mitigation policy among American Republican and Democratic subjects. Given the possibility that motivated reasoning may occur, caution is needed when extrapolating from systematic processing to attitude and behavior change. For individuals who already oppose climate change mitigation policy, or for whom such policies are incongruent with their worldview, we cannot exclude the

possibility that thinking deeply about climate change may promote decreased support for policy, and future research will be necessary to test this possibility (see below).

A second caveat relates to the implications of labeling climate change mitigation an individual responsibility. Grounded in social psychological and communication theory, this study focuses more on predictive relationships between individual-level variables (the so-called ‘‘attitude-behavior-choice’’ or ABC model), and less on interrogating the social conditions and processes that re(create) and maintain the environmental challenges we seek to surmount (Shove, 2010). In the context of climate change, ABC studies might implicitly or explicitly (such as the present study) reinforce the idea that the responsibility to counter climate change rests solely with individuals, at the same time possibly ‘‘[obscuring] the extent to which governments sustain unsustainable economic institu- tions and ways of life’’ (Shove, 2010, p. 1274). Research in sociology and geography, for instance, suggests that the framing of societal risks and responses as individual (i.e., involving citizens) or collective (i.e., involving multiple outside actors, such as govern- ments or corporations) poses both ethical and political implica- tions (Bickerstaff et al., 2008; Bickerstaff & Walker, 2002). Future research should examine ways to convey climate change as a ‘‘collective risk,’’ involving private citizens, governments, and other institutions, and how attribution of responsibility for mitigation might be presented in a more complex manner (see below).

5.2. Practical implications

The practical implications from this study relate to designing message strategies to encourage systematic thinking about climate change. Framing the responsibility for climate change mitigation as based on individual actions may motivate individuals to think about the implications of climate change for their own lives. Drawing this connection can potentially guide them to consider what they can do personally to meet the challenge of climate change, leading, in turn, to behavioral intentions and eventual behavior change. While we must acknowledge the potential limitations of implicating individual responsibility (discussed above), this study nonetheless contributes to a growing literature establishing effective framing techniques to encourage attitudinal and behavioral responses to climate change (Nisbet, 2009), including portraying climate change as significant to human health (Maibach et al., 2010), employing iconic imagery (O’Neill et al., 2013), and stressing local, rather than global, impacts (Scannell & Gifford, 2013; Spence et al., 2012). Most importantly, our results echo recent messaging studies that have found that emphasizing the societal benefits or gains resulting from climate change mitigation (rather than the risks or costs of forgoing action) can promote engagement (Bain et al., 2012; Gifford & Comeau, 2011; Spence & Pidgeon, 2010).

Further, our results suggest that an ‘‘individual responsibility’’ frame may promote systematic processing across individual differences. Public opinion data suggest that, among the U.S. public, belief in the anthropogenic causes of climate change, as well as support for its mitigation, tend to run along ideological lines; those identifying as more conservative in political philoso- phy as well as affiliated with the Republican party, for instance, tend to be more skeptical of the evidence for human-caused climate shifts, as well as demonstrate less support for mitigation policies (e.g., McCright & Dunlap, 2011a, 2011b; Kahan et al., 2012). In response, current social scientific research has sought strategies to encourage thinking about such contentious issues that spans partisan divides, including varying the ‘‘psychological distance’’ (e.g., socially, geographically, or temporally close to or removed from an individual) with which an issue is framed (e.g., Hernandez and Preston, 2013; Luguri et al., 2012; Spence &

L.N. Rickard et al. / Global Environmental Change 26 (2014) 39–52 49

Pidgeon, 2010). Our study contributes to this conversation in that our experimental manipulation led to more systematic processing of an issue that may not garner immediate support among certain populations (e.g., identifying as having conservative political philosophy), and among populations that likely already support the topic (e.g., those with favorable attitudes toward climate change information). In particular, the internal treatment attribu- tion condition led to more systematic processing among both Republicans (t[83] = 2.33, p < .05) and non-Republicans (t[471] = 2.96, p < .01) alike. Hence, it is possible that framing climate change as an issue requiring individual attention may be an appropriate strategy to encourage those aligning with disparate political camps to think further about the issue; however, more research is needed to ensure that this strategy would not, at the same time, introduce unintended, ‘‘boomerang’’ effects among either of the groups (Hart & Nisbet, 2012).

Other findings on individual differences merit further attention for designing climate change communication messaging to target particular audiences. First, we found that household income was negatively related to information sufficiency threshold when participants were exposed to a message that emphasizes govern- ment policy as a solution. That is, when government policy was highlighted as a solution for climate change, those with higher SES reported a lower need for information related to climate change mitigation. In the long run, this limited desire for new information, or a sense of satisfaction with existing information about climate change, might contribute to a more restricted knowledge base. Second, we found that self-identified Christians were more likely to systematically process the information presented in the news stories in the internal attribution condition. This relationship suggests that information that stresses individual responsibility related to climate change mitigation was effective in eliciting Christian participants’ attention. Since we lack data on respon- dents’ overall level of religiosity, we refrain from drawing further conclusions based on this finding. Future research, however, should take into account how religion affects individuals’ reaction to climate change information (see Clements et al., 2013).

5.3. Study limitations

In considering these results, it is also important to point out study limitations; first among them is the limited external validity due to the experimental survey design that involved university students as research participants. To date, however, most research that experimentally examines information processing has relied on university students as research participants (Chaiken et al., 1989, 1996). Past research has also shown that the information tasks university students must perform in their daily lives are similar to those of adults (Rieh & Hilligoss, 2008). Nonetheless, it is important to note that student participants’ responses in social science research tend to be more homogeneous than those of nonstudent participants, which may reduce the magnitude of differences or minimize relationships that do exist among variables (Peterson, 2001). Second, readers should take caution not to over-generalize our findings, as data were collected through convenience samples. Future research should consider conducting experimental surveys based on national probability samples to draw more generalizable conclusions.

Other limitations are more specific to the operationalization of the key variables in this study. Most importantly, we averaged four specific negative emotions together to measure negative affect, even though some of these emotions may have a greater influence on information processing than others. For instance, sadness is often attributed to uncontrollable events, whereas anxiety is often attributed to events that are modifiable (Frijda et al., 1989). Thus, anxiety may trigger more systematic processing than sadness

because the amount of time and effort spent in processing information related to controllable events will be more worthwhile. Future research should further examine the unique impact that discrete negative emotions have on information processing (Lerner & Keltner, 2000, 2001). Even though all the measures were adopted from past research and achieved acceptable reliability, the scope of assessment varied. Most of the RISP variables were measured in more general terms, but systematic processing was assessed specifically in relation to the simulated news stories in the experiment. As explained above, this difference might have contributed to the null result regarding how attribution of responsibility moderates the other relationships among RISP variables and systematic processing.

Lastly, compared to its stylized representation in the newspaper article in this study, climate change mitigation outside of the laboratory is a nuanced and complex process. Mitigation may be most effective when both government policy and individual action work hand in hand. By virtue of experimental design, however, our manipulation artificially separated these two aspects to test their relative effectiveness. Using messaging strategies that combine discussions of individual- and government policy-level actions may achieve promising results, and should be pursued in future research (see below). Moreover, it is possible that the substantive issues examined in the articles (i.e., reduction in the use of air conditioning, and the deployment of a carbon tax) may have influenced our results. For instance, subjects may have misunder- stood the relationship between air conditioning use and climate change, or alternatively, those more informed about climate change causation may have been more likely to have understood the reasoning behind a carbon tax. Following some recent examples (e.g., Bostrom et al., 2012), future studies should attempt to identify how perceived causes of climate change influence attribution of treatment responsibility.

5.4. Future research

Rather than focusing on the direct effects that each component exerts on the dependent variable (i.e., information processing), future research should continue to test indirect linkages proposed in the RISP model in the context of impersonal risks such as climate change. In addition, future research should work toward further clarifying the relationship between treatment attribution and risk perception and how these variables may work together to influence how individuals process information about climate change and other risks. Finally, we suggest studies that explore ‘‘shared’’ or ‘‘complex’’ attribution of treatment responsibility, wherein responsibility for responding to an issue is attributed not solely to an internal source (actor) or an external source (actor), but rather to a combination of both. Just as emerging research in public health has begun to link ‘‘complex attributions’’ (e.g., Barry et al., 2012; Jeong, 2007; Lundell et al., 2013) with communication about the causes and treatments of obesity, messaging strategies that combine individual and external sources may provide a more ecologically valid representation of how individuals think about climate change, and may also better represent climate change as a collective risk to be managed by both individuals and institutions alike (Shove, 2010).

6. Conclusions

As levels of atmospheric greenhouse gases rise and evidence of anthropogenic climate change mounts, encouraging public audi- ences to think about climate change mitigation becomes a critical first step to encouraging shifts toward pro-environmental attitudes and behaviors. While previous social science research has established relationships between risk perceptions and causal attribution, this experiment-based study proceeded a step further

L.N. Rickard et al. / Global Environmental Change 26 (2014) 39–5250

to explore how risk perceptions and attribution of responsibility for responding to climate change mitigation influence information processing about the risk. In addition to clarifying existing relationships in the RISP model, our results open new doors for exploring mediating relationships between attitudinal, affective, and information processing variables. Significant differences

Appendix A. Experimental stimulus for internal attribution condi

Appendix B. Experimental stimulus for the external attribution co

observed in information processing between the individual responsibility and societal responsibility experimental conditions, as well as differences due to individual characteristics (e.g., SES, religion, political party affiliation), represent important avenues to designing strategic communication to encourage deeper thinking about climate change.

tion

ndition

L.N. Rickard et al. / Global Environmental Change 26 (2014) 39–52 51

References

Bain, P.G., Hornsey, M.J., Bongiorno, R., Jeffries, C., 2012. Promoting pro-environ- mental action in climate change deniers. Nat. Climate Change 2, 600–603, http://dx.doi.org/10.1038/nclimate1532.

Barry, C.L., Gollust, S.E., Niederdeppe, J., 2012. Are Americans ready to solve the weight of the nation? N. Engl. J. Med. 367 (5), 389–391, http://dx.doi.org/ 10.1056/NEJMp1206519.

Bickerstaff, K., Simmons, P., Pidgeon, N., 2008. Constructing responsibilities for risk: negotiating citizen-state relationships. Environ. Plan. A 40 (6), 1312, http:// dx.doi.org/10.1068/a39150.

Bickerstaff, K., Walker, G., 2002. Risk, responsibility, and blame: an analysis of vocabularies of motive in air-pollution (ing) discourses. Environ. Plan. A 34 (12), 2175–2192, http://dx.doi.org/10.1068/a3521.

Bostrom, A., O’Connor, R.E., Bohm, G., Hanss, D., Bodi, O., Ekstrom, F., Saelensminde, I., 2012. Causal thinking and support for climate change policies: international survey findings. Global Environ. Change 22, 210–222, http://dx.doi.org/ 10.1016/j.gloenvcha.2011.09.012.

Brickman, P., Rabinowitz, V.C., Karuza, J., Coates, D., Cohn, E., Kidder, L., 1982. Models of helping and coping. Am. Psychol. 37, 368–384, http://dx.doi.org/ 10.1037/0003-066X.37.4.368.

Chaiken, S., 1980. Heuristic versus systematic information processing and the use of source versus message cues in persuasion. J. Pers. Soc. Psychol. 39, 752–766, http://dx.doi.org/10.1037/0022-3514.39.5.752.

Chaiken, S., Giner-Sorolla, R., Chen, S., 1996. Defense and impression motives in heuristic and systematic information processing. In: Gollwitzer, P.M., Bargh, J.A. (Eds.), The Psychology of Action: Linking Motivation and Cognition to Behavior. Guilford Press, New York, pp. 553–578.

Chaiken, S., Liberman, A., Eagly, A.H., 1989. Heuristic and systematic information processing within and beyond the persuasion context. In: Uleman, J.S., Bargh, J.A. (Eds.), Unintended Thought. Guilford Press, New York, pp. 212–252.

Chen, S., Chaiken, S., 1999. The heuristic–systematic model in its broader context. In: Chaiken, S., Trope, Y. (Eds.), Dual-Process Theories in Social Psychology. Guilford Press, New York, NY, pp. 73–96.

Clements, J.M., McCright, A.M., Xiao, C., 2013. Green Christians? An empirical examination of environmental concern within the U.S. general public. Org. Environ., http://dx.doi.org/10.1177/1086026613495475.

Cohen, J., Cohen, P., West, S.G., Aiken, L.S., 2003. Applied Multiple Regression/ Correlation Analysis for the Behavioral Sciences, 3rd ed. Erlbaum Associates, Mahwah, NJ.

Ditto, P.H., Scepansky, J.A., Munro, G.D., Apanovitch, A.M., Lockhart, L.K., 1998. Motivated sensitivity to preference-inconsistent information. J. Pers. Soc. Psy- chol. 75, 53–69.

Druckman, J.N., Bolsen, T., 2011. Framing, motivated reasoning, and opinions about emergent technologies. J. Commun. 61, 659–688.

Eagly, A., Chaiken, S., 1993. The Psychology of Attitudes. Harcourt Brace and Jovanovich, Fort Worth, TX.

Frijda, N.H., Kuipers, P., ter Schure, E., 1989. Relations among emotion, appraisal, and emotional action readiness. J. Pers. Soc. Psychol. 57, 212–228.

Gifford, R., Comeau, L.A., 2011. Message framing influences perceived climate change competence, engagement and behavioral intentions. Global Environ. Change 21, 1301–1307.

Gillis, J., 2013, May. Heat-trapping gas passes milestone, raising fears. The New York Times. Retrieved from http://www.nytimes.com/2013/05/11/science/earth/ carbon-dioxide-level-passes-long-feared-milestone.html?ref=earth.

Griffin, R.J., Dunwoody, S., Neuwirth, K., 1999. Proposed model of the relationship of risk information seeking and processing to the development of preventive behaviors. Environ. Res. 80 (2), S230.

Griffin, R.J., Neuwirth, K., Dunwoody, S., Giese, J., 2004. Information sufficiency and risk communication. Media Psychol. 6, 23–61.

Griffin, R.J., Neuwirth, K., Giese, J., Dunwoody, S., 2002. Linking the heuristic- systematic model and depth of processing. Commun. Res. 29, 705–732, http://dx.doi.org/10.1177/009365002237833.

Griffin, R.J., Yang, Z.J., Boerner, F., Bourassa, S., Darrah, T., Knurek, S., Dunwoody, S., 2005. Applying an information seeking and processing model to a study of communica- tion about energy. In: Paper Presented at the Annual Convention of the Association for Education in Journalism and Mass Communication, San Antonio, TX.

Griffin, R.J., Yang, Z.J., ter Huurne, E., Boerner, F., Ortiz, S., Dunwoody, S., 2008. After the flood: anger, attribution, and the seeking of information. Sci. Commun. 29, 285–315, http://dx.doi.org/10.1177/1075547007312309.

Hart, P.S., 2011. One or many? The influence of episodic and thematic climate change frames on policy preferences and individual behavior change. Sci. Commun. 33, 28–51.

Hart, P.S., Nisbet, E.C., 2012. Boomerang effects in science communication: how motivated reasoning and identity cues amplify opinion polarization about climate mitigation policies. Commun. Res. 39, 701–723.

Hayes, A.F., 2013. Introduction to Mediation, Moderation, and Conditional Process Analysis: A Regression-Based Approach. Guilford Press, New York, NY.

Heider, F., 1982. The Psychology of Interpersonal Relations. Lawrence Erlbaum, Hillsdale, NJ.

Hernandez, I., Preston, J.L., 2013. Disfluency disrupts the confirmation bias. J. Exp. Soc. Psychol. 49, 178–182, http://dx.doi.org/10.1016/j.jesp.2012.08.010.

Hovick, S.R., Freimuth, V.S., Johnson-Turbes, A., Chervin, D.D., 2011. Multiple health risk perception and information processing among African Americans and Whites living in poverty. Risk Anal. 31, 1789–1799, http://dx.doi.org/ 10.1111/j.1539-6924.2011.01621.x.

Iyengar, S., 1989. How citizens think about national issues: a matter of responsibili- ty. Am. J. Polit. Sci. 33 (4), 878–900.

Iyengar, S., 1990. Framing responsibility for political issues: the case of poverty. Polit. Behav. 12, 19–40.

Jang, S.M., 2013. Framing responsibility in climate change discourse: ethnocentric attribution bias, perceived causes, and policy attitudes. J. Environ. Psychol. 36, 27–36, http://dx.doi.org/10.1016/j.jenvp.2013.07.003.

Jeong, S., 2007. Effects of news about genetics and obesity on controllability attribution and helping behavior. Health Commun. 22, 221–228, http:// dx.doi.org/10.1080/10410230701626877.

Kahan, D.M., Peters, E., Wittlin, M., Slovic, P., Ouellette, L.L., Braman, D., Mandel, G., 2012. The polarizing impact of science literacy and numeracy on perceived climate change risks. Nat. Climate Change 2, 732–735, http://dx.doi.org/ 10.1038/nclimate1547.

Kahlor, L., 2007. An augmented risk information seeking model: the case of global warming. Media Psychol. 10, 414–435, http://dx.doi.org/10.1080/ 15213260701532971.

Kahlor, L., 2010. PRISM: a planned risk information seeking model. Health Commun. 25, 345–356.

Kahlor, L., Dunwoody, S., Griffin, R.J., 2002. Attributions in explanations of risk estimates. Public Understand. Sci. 11, 243–257.

Kahlor, L., Dunwoody, S., Griffin, R.J., Neuwirth, K., Giese, J., 2003. Studying heuris- tic–systematic processing of risk communication. Risk Anal. 23, 355–368, http://dx.doi.org/10.1111/1539-6924.00314.

Kouabenan, D.R., 1998. Beliefs and the perception of risks and accidents. Risk Anal. 18, 243–252.

Kouabenan, D.R., 2002. Occupation, driving experience, and risk and accident perception. J. Risk Res. 5, 49–68.

Leiserowitz, A., 2006. Climate change risk perception and policy preferences: the role of affect, imagery, and values. Clim. Change 77, 45–72.

Lerner, J.S., Keltner, D., 2000. Beyond valence: toward a model of emotion-specific influences on judgement and choice. Cognit. Emot. 14, 473–493.

Lerner, J.S., Keltner, D., 2001. Fear, anger, and risk. J. Pers. Soc. Psychol. 81, 146–159. Luguri, J.B., Napier, J.L., Dovidio, J.F., 2012. Reconstruing intolerance: abstract

thinking reduces conservatives’ prejudice against nonnormative groups. Psy- chol. Sci. 23, 756–763, http://dx.doi.org/10.1177/0956797611433877.

Lundell, H., Niederdeppe, J., Clarke, C., 2013. Public views about health causation, attributions of responsibility, and inequality. J. Health Commun. 18, 1116– 1130, http://dx.doi.org/10.1080/10810730.2013.768724.

Maibach, E.W., Roser-Renouf, C., Leiserowitz, A., 2008. Communication and mar- keting as climate change intervention assets: a public health perspective. Am. J. Prev. Med. 35, 488–500.

Maibach, E.W., Nisbet, M., Baldwin, P., Akerlof, K., Diao, G., 2010. Reframing climate change as a public health issue: an exploratory study of public reactions. BMC Public Health 10, 299–310, http://dx.doi.org/10.1186/1471-2458-10-299.

Mannetti, L., Pierro, A., 1991. Health care workers’ reactions to AIDS victims: perception of risk and attribution of responsibility. J. Commun. Appl. Soc. Psychol. 1, 133–142.

McCright, A.M., Dunlap, R.E., 2011a. Cool dudes: the denial of climate change among conservative white males in the U.S. Global Environ. Change 21, 1163–1172, http://dx.doi.org/10.1016/j.gloenvcha.2011.06.003.

McCright, A.M., Dunlap, R.E., 2011b. The politicization of climate change and polarization in the American public’s views of global warming, 2001–2010. Sociol. Quart. 52, 155–194.

Nisbet, M.C., 2009. Communicating climate change: why frames matter for public engagement. Environ. Sci. Policy Sustain. Dev. 51 (2), 12–23.

Olausson, U., 2011. ‘‘We’re the ones to blame’’: citizens’ representations of climate change and the role of the media. Environ. Commun.: J. Nat. Culture 5, 281–299, http://dx.doi.org/10.1080/17524032.2011.585026.

O’Neill, S., Boykoff, M., Niemeyer, S., Day, S.A., 2013. On the use of imagery for climate change engagement. Global Environ. Change 23, 413–421.

O’Neill, S., Nicholson-Cole, S., 2009. ‘‘Fear won’t do it’’: promoting positive engage- ment with climate change through visual and iconic representations. Sci. Commun. 30, 355–379.

Peltzer, K., Renner, W., 2003. Superstition, risk-taking, and risk perception of accidents among South African taxi drivers. Accid. Anal. Prev. 35 (4), 619– 623, http://dx.doi.org/10.1016/S0001-4575(02)00035-0.

Peterson, R.A., 2001. On the use of college students in social science research: insights from a second-order meta-analysis. J. Consum. Res. 28, 450–461.

Petts, J., 2005. Health, responsibility, and choice: contrasting negotiations of air pollution and immunisation information. Environ. Plan. A 37 (5), 791–804, http://dx.doi.org/10.1068/a3779.

Pidgeon, N., 2012. Public understanding of, and attitudes to, climate change: UK and international perspectives and policy. Climate Policy 12, S85–S106, http:// dx.doi.org/10.1080/14693062.2012.702982.

Pidgeon, N., Fischhoff, B., 2011. The role of social and decision sciences in commu- nicating uncertain climate risks. Nat. Climate Change 1, 35–41, http:// dx.doi.org/10.1038/nclimate1080.

Posas, P.J., 2007. Roles of religion and ethics in addressing climate change. Ethics Sci. Environ. Polit. 2007, 31–49, http://dx.doi.org/10.3354/esep00080.

Rickard, L.N., 2013. Perception of risk and the attribution of responsibility for accidents. Risk Anal., http://dx.doi.org/10.1111/risa.12118.

Rieh, S.Y., Hilligoss, B., 2008. College students’ credibility judgments in the information-seeking process. In: Metzger, M.J., Flanagin, A.J. (Eds.), Digital Media Youth, and Credibility. The MIT Press, Cambridge, MA, pp. 49–72.

L.N. Rickard et al. / Global Environmental Change 26 (2014) 39–5252

Scannell, L., Gifford, R., 2013. Personally relevant climate change: the role of place attachment and local versus global message framing in engagement. Environ. Behav. 45 (1), 60–85, http://dx.doi.org/10.1177/0013916511421196.

Sellstrom, E., Bremberg, S., Garling, A., Hornquist, J.O., 2000. Risk of childhood injury: predictors of mothers’ perceptions. Scand. J. Public Health 28 (3), 188– 193, http://dx.doi.org/10.1177/14034948000280030701.

Shaver, K.G., 1985. The Attribution of Blame: Causality, Responsibility, and Blame- worthiness. Springer-Verlag, New York, NY.

Shove, E., 2010. Beyond the ABC: climate change policy and theories of social change. Environ. Plan. A 42, 1273–1285, http://dx.doi.org/10.1068/a42282.

Slovic, P., 1987. Perception of risk. Science 236 (April (17)), 280–285. Spence, A., Pidgeon, N., 2010. Framing and communicating climate change: the

effects of distance and outcome frame manipulations. Global Environ. Change 20, 656–667, http://dx.doi.org/10.1016/j.gloenvcha.2010.07.002.

Spence, A., Poortinga, W., Pidgeon, N., 2012. The psychological distance of climate change. Risk Anal. 32 (6), 957–972, http://dx.doi.org/10.1111/j.1539- 6924.2011.01695.x.

Stoddart, M.C.J., Tindall, D., Greenfield, K.L., 2012. ‘‘Governments have the power’’? Interpretations of climate change responsibility and solutions among Canadian environmentalists. Org. Environ. 25 (1), 39–58, http://dx.doi.org/10.1177/ 1086026612436979.

Swim, J.K., Stern, P.C., Doherty, T.J., Clayton, S., Reser, J.P., Weber, E.U., Gifford, R., Howard, G.S., 2011. Psychology’s contributions to understanding and addres- sing global climate change. Am. Psychol. 66 (4), 241–250, http://dx.doi.org/ 10.1037/a0023220.

Taber, C.S., Lodge, M., 2006. Motivated skepticism in the evaluation of political beliefs. Am. J. Polit. Sci. 50, 755–769.

Tangney, J.P., Stuewig, J., Mashek, D.J., 2007. Moral emotions and moral behavior. Annu. Rev. Psychol. 58, 345–372, http://dx.doi.org/10.1146/annurev.p- sych.56.091103.070145.

ter Huurne, E.F.J., Griffin, R.J., Gutteling, J.M., 2009. Risk information seeking among U.S. and Dutch residents. Sci. Commun. 31, 215–237, http://dx.doi.org/10.1177/ 1075547009332653.

Tversky, A., Kahneman, D., 1981. The framing of decisions and the psychology of choice. Science 211 (4481), 453–458.

Weiner, B., 2006. Social Motivation, Justice and the Moral Emotions: An Attribu- tional Approach. Lawrence Erlbaum, Mahwah, NJ.

Yang, Z.J., Kahlor, L., 2013. What, me worry? The role of affect in information seeking and avoidance. Sci. Commun. 35, 189–212, http://dx.doi.org/10.1177/ 1075547012441873.

Yang, Z.J., Kahlor, L., Li, H., 2013. A U.S.–China comparison of risk information-seeking intentions. Commun. Res., http://dx.doi.org/10.1177/0093650213479795.

Yang, Z.J., McComas, K.A., Gay, G., Leonard, J.P., Dannenberg, A.J., Dillon, H., 2010. From information processing to behavioral intentions: exploring cancer patients’ motivations for clinical trial enrollment. Patient Educ. Couns. 79, 231–238.

Yang, Z.J., McComas, K.A., Gay, G., Leonard, J.P., Dannenberg, A.J., Dillon, H., 2010. Motivation for health information seeking and processing about clinical trial enrollment. Health Commun. 25, 423–436.

Yang, Z.J., McComas, K.A., Gay, G., Leonard, J.P., Dannenberg, A.J., Dillon, H., 2011. Information seeking related to clinical trial enrollment. Commun. Res. 38, 856– 882.

Yang, Z.J., McComas, K.A., Gay, G., Leonard, J.P., Dannenberg, A.J., Dillon, H., 2012. Comparing decision making between cancer patients and the general popula- tion: thoughts, emotions, or social influence? J. Health Commun. 17, 477–494, http://dx.doi.org/10.1080/10810730.2011.635774.

Zhao, X., Leiserowitz, A., Maibach, E.W., Roser-Renouf, C., 2011. Attention to science/ environment news positively predicts and attention to political news negative- ly predicts global warming risk perceptions and policy support. J. Commun. 61, 713–731, http://dx.doi.org/10.1111/j.1460-2466.2011.01563.x.