Reliability and Validity of Widely Used International Surveys on the Environment

Abstract

Do existing public opinion surveys provide valid and reliable measures of attitudes towards environmental sustainability? This question is critical given the importance of public support for achieving sustainability. Starting with 28 survey items about the environment drawn from the World Values Survey Waves 5 and 6 and the 2010 International Social Survey Program, we assessed reliability by checking for significant correlations between similar or identical items on different surveys. Next, to assess validity, we evaluated correlations between survey items and 22 objective environmental indicators drawn from the Environmental Performance Index (EPI). As the level of economic development is a likely confound, we also performed partial correlation analyses controlling for GDP per capita. From the initial 28 items, we identified 23 sufficiently reliable items, but many of these were found to have low predictive power in the validity analysis. Items about air and water pollution were valid predictors of objective environmental conditions in these areas. Items asking about the relative importance of environmental problems compared to other social issues were also good positive predictors of progress on perceptible environmental issues. Items asking about general sentiment with no clear referent performed poorly. When controlling for GDP, country-level attitudes were more aligned with country-specific environmental conditions. Finally, nearly half of all EPI indicators were associated with few or no survey items, indicating the existence of 'blind spots' in public awareness. Our findings should offer guidance to both survey developers and users, as well as to policy makers responsible for conveying information about environmental sustainability to the wider public.