The Greening of the Globe? Cross-national Levels of Environmental Group Membership

Abstract

Is the environmental movement still growing in members? What explains cross-national levels of environmental mobilisation? This article addresses these questions using data from the new 1999–2002 wave of the World Values Survey. We describe the membership levels in environmental groups across nations, and then examine rival explanations for why membership is concentrated in certain nations although environmental concerns exist globally. We first find that environmental groups represent one of the most common forms of political group membership on a global scale, and membership levels are increasing. We also demonstrate that the combination of social and political conditions in advanced industrial democracies is a strong predictor of environmental group membership levels.

Acknowledgements

The author would like to thank Robert Rohrschneider, David Frank and David Meyer for their contributions to this research, Alix van Sickle for her research assistance and the World Values Survey Project for access to the data from the 1999–2002 wave.

Notes

• Additional information on the WVS samples, fieldwork and the questionnaire are available on the project website: www.worldvaluessurvey.org. These data are available from the Inter-university Consortium for Political and Social Research (ICPSR) (3975) and other national archives. We began with the 59 nations included in the first public release of the data (May 2004) as described in Inglehart et al. (2004).

• We did not include two nations (Bangladesh and Tanzania) in our analyses because of concerns about the representativeness of the sample. In both instances environmental group membership is very high (20.3% and 20.1%, respectively), as is membership in the other social groups listed in the WVS. Our analyses for this article and in other analyses determined that both samples have a very large upper-status, upper-education bias. For instance, the United Nations Development Programme education index gives Tanzania and Uganda equivalent scores, but education levels in the Tanzanian WVS sample are twice as high as in Uganda, and comparable to the level of many European nations. The United Nations Development Programme education statistics for Bangladesh are substantially lower, but the WVS sample had a mean educational level even higher than Tanzania. Excluding these two nations strengthens the impact of economic condition, but has less effect on the other relationships presented below.

Markedly higher levels of membership were registered for labour unions (12.6%), professional associations (6.9%) and educational groups (12.2%).

Neumayer (2002) did a validity check on some of the attitudinal items from the Gallup Health of the Planet survey and found much lower levels of consistency, often falling below minimal levels of statistical significance. However, his analyses focused on measures of environmental opinions, which are more variable in measurement and meaning across surveys, and his reference poll was the Gallup survey. The WVS correlations with the European Social Survey and International Social Survey Program display much stronger consistency.

Information on the European Social Survey (ESS) is available at www.europeansocialsurvey.org/. On average, membership levels are 2.25% higher in the ESS. This is likely due to the more inclusive wording of the European Social Survey question: ‘an organisation for environmental protection, peace or animal rights’.

Information on the ISSP is available at www.issp.org/. Overall, average membership levels are within 2% in both the 1993 and 2000 ISSP surveys. This is a small difference considering the different time frame and question wording in the two surveys. The ISSP wording is ‘are you a member of any group whose main aim is to preserve or protect the environment?’

We did not include the third wave of the WVS (1995–98) because that wave used a different list of groups, which raises issues of the comparability of membership levels. Most of western Europe was also missing from the third wave. Additional data on the prior WVS waves are presented in Inglehart (1997), Norris (2003) and Dalton & Rohrschneider (2002).

These data are taken from United Nations Development Programme (2001). Our indicator is gross national product per capita, adjusted for price parity. Because economic statistics vary across time, and our surveys were conducted in different years, we decided to use data from 2000 as consistent across nations in the adjustments for purchasing price parity (ppp) and other factors, and this year was in close proximity to the sampling dates of the various WVS surveys.

We also explored an alternative measure of socio-economic development. The United Nations' Human Development Index combines economic conditions, literacy and other social factors to go beyond the simple economics of GDP (using the 2000 index scores). This measure is positively related to environmental group membership (r = 0.190).

This index is the average of the two seven-point scales of civil liberties and political rights that the Freedom House reports for each nation. The resulting index was recoded to range from 1 (low democracy) to 7 (high). We use the Freedom House scores for the same year in which the WVS was conducted.

The press freedom measure is from the Freedom House; the corruption index is from Transparency International. Both of these statistics are from the same year in which the WVS was conducted in the nation.

Postmaterial values also overlap with the values of the ‘new environmental paradigm’, which represents a more encompassing, biocentric view of nature. These two value dimensions are empirically related (Milbrath, 1984; Dalton et al., 1999) and, since the postmaterial index is available from the WVS, we focus on this measure of values.

Dunlap et al. (1993) classified their nations into high, medium and low personal income. Sewage was named as a very serious community problem by 46% in the low-income nations, compared to only 12% in the high-income nations. In the low-income nations, 42% cited poor water quality and 34 poor air quality, compared to 13% and 14% respectively in the high-income nations. Concerns about soil, overcrowding and noise problems display much weaker relationships with national income.

• We use the four-item postmaterial index because it is available for a larger number of nations. The question presents four choices to the respondent, and asks them to identify their first and second choice. Materialists select the first and third items, postmaterialists the second and fourth (Inglehart, 1997 In your opinion, which one of these items is most important? And what would be the next most important?

• Maintaining order in the nation

• Giving people more say in important government decisions

• Fighting rising prices

• Protecting freedom of speech

We selected the election adjacent to the WVS and coded the percentage of the vote the green party received in the election. Our data source was www.electionworld.org.

A factor analysis of the four subdimensions and the ESI index for the full set of nations finds that all items load positively on a first dimension. However, the stress index displays the weakest factor loading (0.119).

This may be an indication that stress leads to mobilisation in advanced industrial democracies, but the strong correlation with national affluence (r =  − 0.705) suggests this may also be a spurious relationship. The test for an independent influence of environmental conditions comes from the multivariate analyses in Table 2 below.

The social/institutional capacity dimension does not directly measure environmental conditions, but might be viewed as tapping the potential for environmental action in a nation. To ensure that this did not distort our results, we created an alternative ESI measure that excludes this subdimension. This revised measure is correlated at 0.907 with the original ESI. In addition, neither ESI summary measure displays a statistically significant relationship with environmental group membership. Substituting this reduced measure in the multiple regression analyses of Table 2 does not significantly change the coefficients in the model (ESI β = − 0.085, ESI revised β =  − 0.117). Thus, we utilise the original ESI in our analyses.

The correlations among these four predictors for all nations are given in Table 3.

• Table 3