A re-assessment of factors associated with environmental concern and behavior using the 2010 General Social Survey

Abstract

The associations between social and psychological influences and environmental attitudes, intentions and behavior have generated considerable interest, both in the fields of environmental behavior and of environmental education. We use the 2010 General Social Survey (GSS) to study these associations and expand the scope of earlier studies by including additional measures across social demographic, social psychological and behavioral blocks. The findings highlight the relationships between social psychological constructs and environmental concern and behavior, as well as the relationships between social demographic characteristics and environmental values and beliefs. The findings from this study will be useful to environmental educators and communicators interested in studying the social psychological and social demographic bases of environmental concern and behavior.

Keywords:

• pro-environmental behavior
• environmental concern
• General Social Survey

Acknowledgments

We would like to thank three anonymous reviewers for their comments and suggestions on this article.

Notes

¹ Scale range from 1 (somewhat strong) to 3 (strong)

² Scale range from 1 (not spiritual at all) to 4 (very spiritual)

¹ Scale range from 1 (Independent, close to Democrat/Republican) to 3 (strong Democrat/Republican)

1. Although the alpha coefficients for PB and FB fall below the acceptable range, both scales use the same battery of questions as the original Dietz et al. (1998) conceptualization. Nonetheless, this limitation calls into question the validity of the results for these variables in the subsequent analysis.

2. Religion and Religious Strength have a high correlation among Protestants (r = 0.51) and no religion (r = –0.86). Due to the high correlation, this may weaken the statistical significance of these variables. Additionally, the variance inflation factors (VIF) indicated that no religion and religious strength may pose a problem of multicollinearity (Tolerance of .20 and .20, respectively). As such, removing religious sects from the analysis would show that religious strength is statistically significant, and vice versa.

3. Party identification and political ideology (Liberalism) have a high correlation among Democrats (r = 0.43) and Republicans (r = −0.47). This may weaken the statistical significance of these variables. However, the VIF indicated that multicollinearity is not a problem.

4. LCI and LCE have a high correlation (r = 0.47). Due to this high correlation, this may weaken the statistical significance of these variables. However, the VIF indicated that multicollinearity is not a problem.

5. Know cause of environmental issues KC and know solution of environmental issues KS have a high degree of correlation (r = 0.63). This high correlation may weaken the statistical significance of these variables. However, the VIF indicated that multicollinearity is not a problem.

6. We would like to thank an anonymous reviewer for this suggestion.