Motivating the Skeptical and Unconcerned: Considering Values, Worldviews, and Norms When Planning Messages Encouraging Energy Conservation and Efficiency Behaviors

Abstract

The current study examined the association of personal values or worldviews with individual motivations for reduced energy use and energy-use-related risk perceptions. We also investigated how perceptions of others’ energy use influenced motivation to reduce use among those with high, versus low, risk perceptions. The perception that others were reducing use was associated with both increased feelings of personal responsibility and guilt-driven motivations to conserve energy among those who perceived greater energy-use-related-risk but was not associated with such motivations among those who perceived less risk. Implications for planning information campaigns to encourage environmental behaviors such as reduced energy use within a diverse public are discussed.

Notes

Risk has been conceptualized as “things, forces, or circumstances that pose danger to people or to what they value” (Stern & Fineberg, 1996, p. 215) and is a concept employed in studies associated with many of the theoretical frameworks relied upon in the current study. Problem awareness is a similar concept employed in many studies of prosocial behavior and some studies of environmental behavior. Both concepts (risk and problem awareness) reflect perceptions of aversive consequences to the self, others, and/or biosphere, so we use them interchangeably in this study.

For example, Wildavksy and Dake (1990) found consideration of oneself as a liberal to be moderately correlated (r = .50) with valuing egalitarianism and membership in the Democratic Party to be somewhat less strongly correlated (r = .30) with egalitarianism. Conversely, Republicans were inclined to embrace hierarchical values (r = .40) and reject egalitarian values (r = −.45).

To minimize costs and to promote a comparison of findings to those from a larger, national study, all items except those related to conservation attitudes and descriptive social norms replicated items used in a previously published study of energy-related behaviors (Leiserowitz, et al., 2008). Demographic data and political affiliation data were provided by Knowledge Networks.

Participants were also asked about their intention to take any of the six actions they had not already taken. These six items reflect Behavioral Intention. However, because participants were branched to these questions only if they had not already taken each action, Behavioral Intention is negatively correlated with Actual/Recent Behavior in this study. Because of this and for the sake of parsimony, we analyze only recent behavior in this study.

Respondents were first asked about recent home energy efficiency and conservation behaviors. Respondents were then branched so that those who had not already taken a specific action were presented questions regarding behavioral intentions, perceived behavioral control (PBC), motivations and barriers, and subjective norms associated with possible actions (PBC, subjective norms, and a measure of energy curtailment attitudes were analyzed in a separate study). Respondents who had already taken a specific action were asked to report motivations for taking the action(s). Whenever respondents were asked to report reasons for taking or not taking action, they were presented with a list of items indicating several potential reasons or motivations and were instructed to “check all that apply.” All respondents received questions with respect to the perceived threat of an energy crisis and climate change, attitudes toward conserving energy, and perceptions of descriptive norms regarding energy conservation.

These items are similar to those from the Introjected Regulation factor of the Motivation toward the Environment Scale (Pelletier, Tuson, Green-Demers, Noels, & Beaton, 1998; Villacorta, Koestner, & Lekes, 2003), which measures environmental behavior motivations related to self-esteem, guilt, and/or anxiety (e.g., “I would feel bad if I didn't …”, “I would feel guilty …”).

By using the dummy coding procedure, the tests of the b coefficients for descriptive norms in the regression analyses provided the tests of the simple slopes for Conservatives (see Table 4 and Table 5). We then recoded political orientation (1 = Conservative, 0 = Liberal) and conducted the same regression analyses to test the significance of the slopes for Liberals (data not shown in the tables).