Gender and Concern for Environmental Issues in Urban China

Abstract

Little research exists on Chinese people's environmental concern, despite China's great global environmental impacts. This study brings four hypotheses of gendered difference in environmental concern commonly found in Western literature to urban China, using a national data set of the 2003 China General Social Survey. Structural equation modeling was applied to examine the interaction effects of gender, employment status, and parental roles, and the intervening effects of environmental knowledge. Results show that, unlike in the West, men in urban China were more concerned with environmental issues than women. However, findings of men's greater environmental knowledge relative to women, higher concerns for pollution issues than other issues among Chinese women, and little effect of employment status and parental role on environmental concern are largely consistent with the Western literature. The applicability of common hypotheses from the West may be less limited than expected in the context of China.

Keywords:

• China General Social Survey
• environmental beliefs
• gender
• structural equation modeling
• urban China

Acknowledgments

An earlier version of this article was presented at 2009 Beijing International Conference of Gender and Social Transformation, July 17–19, Beijing, China. Many thanks to Dr. Celine-Marie Pascale, Dr. Michelle Newton-Francis, the three anonymous reviewers, and the chief editor for their insightful comments and suggestions in revising this article.

Notes

Note. The respondents were asked: “How serious is each environmental problem in your community? (1) Very serious; (2) Fairly serious; (3) Maybe a little serious; (4) Not very serious; (5) Not serious at all/No such issue.” We recoded all items so that higher scores reflected higher degree of perceived seriousness and hence higher level of concern. Missing data were recoded to the middle point “maybe a little serious.” Interfactor correlation was 0.643, which could indicate a uni-factor structure. For conceptual reasons and the need of hypothesis testing, we maintained the separation between the two factors.

Note. Items 1, 3, 5, 7, and 9 are false, and the rest are true. A correct answer earns the respondent one point. All other answers were recoded as wrong.

Note. Superscripts indicate the following: 1, perceptions of pollution issues; 2, perceptions of other environmental issues; 3, index of environmental knowledge; 4, “employed men” is the base category of comparisons, omitted from the equation.

^a Significant at p < .05.

Note. Superscripts indicate the following: 1, perceptions of pollution issues; 2, perceptions of other environmental issues; 3, index of environmental knowledge; 4, significance test not conducted for indirect effects.

^a Signficant at p < .05.

The China General Social Survey was conducted annually from 2003 to 2006. The Environment Module was only administered in 2003. It was difficult to conduct tests for representativeness of this sample due to lack of reliable parameters for the urban population in China. Our confidence in the quality of sample and data was primarily based on the rigorousness of sampling and survey implementation procedures. For details, see http://www.ust.hk/~websosc/survey/GSS2003e0.html

In comparison, the 2003 poverty line of China, set by the World Bank, was roughly US$1,368 (purchasing power parity) for a family of three members, US$1.25 per day per person.

Currently there is no single best fitness statistic (see Bollen 1989 and Byrne 2010 for more details). We thus report a group of model evaluation statistics, a standard practice in literature. Overall, most fitness statistics show that our model has a good fit except the significant chi-squared test, which tends to be sensitive to large sample sizes (Bollen 1989; Byrne 2010).