Abstract
How internalized and integrated into the person's cognitive and goal structures are norms guiding environmentally desirable behavior? In two surveys (N = 206 and N = 200), subjective social norms and personal norms for a specific behavior (the purchase of organic food or recycling) as well as self-reported behavior and the person's reasons and motives for performing the behavior were measured. The number and types of associations differ depending on the strength of the person's norms and the two types of norms differ in their embeddedness in the person's cognitive structures. With the partial exception of really low-cost behaviors (e.g., recycling in many contexts), environmentally responsible behavior is guided by what seems to be truly internalized and integrated (personal) norms.
ACKNOWLEDGMENTS
The research was supported by a grant from the Danish Social Science Research Council. I am grateful to Folke Ölander, two anonymous reviewers, and the handling editor for helpful comments on an earlier draft.
Notes
1See http://www.mim.dk/ministeren/fokus/groentansvar.htm and www.groentansvar.dk.
3See http://www.mostofus.org.
4As remarked by one anonymous reviewer, irrespective of whether the low reliability is a substantive result, it is a problem that the construct is methodologically flawed in a way that favors the hypotheses. For example, if the subsequent analyses confirm that there is a stronger impact of personal norms than subjective social norms on behavior, it is unclear whether this reflects the true relationships or is an artifact of the lower reliability of the subjective social norm measures. To test the robustness of the findings, and place both types of norms on a “level playing field,” I ran a supplemental set of analyses using only one item to represent each of the two norm constructs. I used the subjective social norms item that seems to represent the injunctive norm most clearly (the item mention first in the Variables section) and the personal norms item which has the wording most parallel with that subjective social norms item (the item mentioned last in the Variables section). The results of these supplemental analyses are consistent with the findings reported in this article, the only substantial difference being that when using single-item instead of multi-item norm constructs bivariate correlations, regression coefficients and explained variances are attenuated. This is as expected, and it suggests that the reported differences between subjective social norms and personal norms in terms of relationships with other constructs are not an artifact produced by differences in the reliabilities of the two measures. To conserve space, the suplemental analyses are not reported here, but they can be acquired from the author.
Note. Inter-coder reliability: Abstraction level: 84%, category: 89%. N = 206.
Note. Stepwise dummy variable regression analysis, standardized solution. N = 206.
5That this was the case for the last two is surprising given that they were not significant predictors of subjective norms in Table . However, they both correlated bivariately with subjective norms (rs = −.14 and .15, p < .05) and in neither of the cases was much attenuation needed to make them nonsignificant (less than |.13|).
Note. Stepwise dummy variable regression analysis, standardized solution, N = 200.
6Before the t test, the 6-point behavior scale used in Study 1 was rescaled by means of this formula: nv = ((ov − 1)∗6/5) + 1. (Re-scaled) means, behavior: 4.83 (organic) and 6.05 (recycling). Means, PN: 3.58 (organic) and 6.01 (recycling).