Content uploaded by Axel Franzen
Author content
All content in this area was uploaded by Axel Franzen on Oct 14, 2019
Content may be subject to copyright.
Acquiescence and the Willingness to
Pay for Environmental Protection: A
Comparison of the ISSP, WVS, and EVS∗
Axel Franzen, University of Bern
Dominikus Vogl, University of Bern
Objectives. This study examines the effect of countries’ wealth on individuals’ will-
ingness to pay for environmental protection. Former studies using the International
Social Survey Programme (ISSP) report a positive effect, while studies using the World
Values Survey (WVS) or the European Values Study (EVS) find the opposite. In this
article, we explain and reconcile these opposing findings. Methods. First, we analyze
the three data sets (ISSP, WVS, and EVS) separately by applying multilevel analyses
and replicate the different findings. Second, we take respondents’ acquiescence into
account and demonstrate that wealth has a positive effect on the willingness to pay in
the combined data set. Results. Respondents in poorer nations in Asia and Eastern
Europe have higher levels of acquiescence than respondents in richer Western nations.
This difference conceals the wealth effect of studies analyzing the WVS or EVS. If ac-
quiescence is properly taken into account, the wealth effect is confirmed. Conclusion.
Theory predicts that wealth and the willingness to protect the environment should
be positively associated. This wealth effect is confirmed by our analyses of the ISSP,
WVS, and EVS.
Concern about the state of the natural environment has been on the public
agenda since the 1970s in many countries. Much of the debate has focused
on the increase of CO2emissions and related global warming. Since the
Intergovernmental Panel on Climate Change (IPCC) was founded in 1988,
it published four reports, which document the increase in global temperature.
According to the latest report, the average global temperature has risen by
0.74◦C during the last 100 years (1906–2006), and the IPCC estimates
that it will increase by another 4–6◦CbytheendofthiscenturyifCO
2
emissions are not reduced (IPCC, 2007). This forecast motivated the EU to
announce cutting CO2emissions by 20 percent from 1990 to 2020 and to
increase the proportion of renewable resources to 20 percent of their total
energy consumption. Simultaneously, the EU announced its willingness to
reduce CO2emissions by 30 percent if other industrial countries would
∗Direct correspondence to Axel Franzen, Institute of Sociology, University of Bern, Lerchen-
weg 36, 3012 Bern, Switzerland franzen@soz.unibe.ch.
SOCIAL SCIENCE QUARTERLY, Volume 94, Number 3, September 2013
C2012 by the Southwestern Social Science Association
DOI: 10.1111/j.1540-6237.2012.00903.x
638 Social Science Quarterly
follow (Commission of the European Communities, 2008). Many scientists
(e.g., the Nobel Prize winner and Minister of Energy of the United States,
Steven Chu) believe that global CO2emissions need to be reduced by at least
half by 2050. The G8 countries, as well as the G5, have, at least in principle,
agreed to these goals (G8 Declaration, 2009).
The realization of these goals demands fundamental behavioral adjust-
ments from citizens and private enterprises. In addition, democratic govern-
ments will need constant support from voters in order to implement relevant
policies. Environmental sociology is, therefore, concerned with questions of
how seriously citizens and voters are concerned with environmental protec-
tion and how to explain the differences in concern found between individ-
uals, as well as between nations. To investigate these issues, a number of
internationally coordinated surveys included questions to measure environ-
mental concern. The International Social Survey Programme (ISSP) in 1993
and 2000 focused on environmental attitudes. The World Values Survey
(WVS) and the European Values Study (EVS) also included some indica-
tors of support for environmental protection in their coordinated surveys.
These surveys allow an assessment of individual support for the environment
as well as international comparisons. However, analyses based on these sur-
veys have come to very different conclusions regarding both the observed
level of environmental concern and the assumed causes of it. On the one
hand, Dunlap and York (2008) and Gelissen (2007) recently presented an
analysis of different waves of the WVS, concluding that support for envi-
ronmental protection is stronger in developing countries than in industrial
societies, which would indeed challenge fundamental assumptions about de-
terminants of environmental concern. Brechin and Bhandari (2011) repeated
the same arguments. On the other hand, Franzen and Meyer (2010) pre-
sented evidence based on the 1993 and 2000 ISSP, finding that environmen-
tal concern is more pronounced in wealthy nations. Thus, the aim of this
article is to reanalyze the three data sources and explain the contradictory
findings.
This study is structured into four sections. In the second section, we will
briefly describe the theoretical arguments for why wealthy individuals and
nations should be more likely to develop stronger environmental concerns.
The third section is concerned with the measurement of environmental con-
cern in the WVS, ISSP, and EVS. Measurements are slightly different in the
surveys but are still comparable. We replicate the results when data sources
are analyzed separately. In the fourth section, we combine the three data
sources and apply a multilevel analysis to the pooled database. Moreover, we
measure respondents’ acquiescence and incorporate it into the analyses. We
demonstrate that the inconclusive evidence of previous research is caused by
differences in respondents’ acquiescence and that the puzzle is resolved when
this is taken into account. In the last section, results are summarized and
discussed.
Willingness to Pay for Environmental Protection 639
The Demand for a Clean Environment
One of the most fundamental laws of economic theory is that an individual’s
consumption of private goods will increase if he/she experiences an increase
in income. Let us assume two private goods or two bundles of private goods
in which an individual is interested. All consumption options can then be
depicted in a diagram in which the amount demanded of one good is indicated
on the X-axis and the amount demanded of the other good on the Y-axis.
The budget line connecting the two axes indicates all consumption options.
Usually, the consumption of a good is subject to decreasing marginal returns
to utility. Thus, an individual prefers a mix of the consumption of the two
goods rather than spending the entire income on one good. The optimal
combination of the two goods is usually indicated by an indifference curve
tangent to the budget line. Increases in income result in a shift of the budget
line away from the origin, which allows the individual to afford more of both
goods.
This basic economic model can be extended to explain the demand for
public goods. Let us assume that one of the bundles consists of all private
goods and that the other consists of all public goods. Pure public goods are
characterized by nonexcludability and nonrivalry of consumption. Thus, an
individual does not need to contribute any resources to the provision of a
public good in order to consume it. Hence, the demand for public goods
is independent of any income restriction. As an example, let us assume that
the individual is interested in coffee (private good) and spending time in a
public park (public good). The consumption of coffee decreases when income
declines, but the use of the public park should not change. However, if we
could theoretically measure the demand for the public good by an individ-
ual’s willingness to pay for it, then, of course, the willingness to pay should
also decrease when income decreases. This relation between the budget re-
striction (or income), the consumption of private goods, and the demand or
willingness to pay for public goods is depicted in Figure 1. In other words,
the theory predicts that the willingness to pay for a public good will increase
if the budget restriction moves away from the origin. In the literature, en-
vironmental concern is usually defined as a multidimensional concept (e.g.,
Franzen and Meyer, 2010; Marquart-Pyatt, 2008; Xiao and Dunlap, 2007),
yet the willingness to pay is a central dimension in every conceptualization.
Therefore, the willingness to pay for a clean environment can be interpreted
as the environmental concern of an individual, or as an important part of it.
Hence, more wealthy individuals should have a higher environmental concern
than less wealthy individuals.
If we aggregate the willingness to pay of all individuals of one society,
then more wealthy nations should, on average, have a higher willingness
to pay for the environment. This predicted positive correlation between the
wealth of nations and the average environmental concern has been empirically
640 Social Science Quarterly
FIGURE 1
The Willingness to Pay for Public Goods and the Demand for Private Goods
confirmed using the 1993 and 2000 ISSP (Diekmann and Franzen, 1999;
Franzen, 2003; Franzen and Meyer, 2010). However, studies using the WVS
show opposing results. Dunlap and York (2008) analyzed three waves of
the WVS (1990–1993, 1995–1998, and 1999–2001) and found negative
correlations between the average reported environmental concern and the gross
national product (GNP) of countries. According to these results, inhabitants
of poorer countries report a higher willingness to pay in order to protect
the environment than citizens of wealthier nations. Most of the correlations
reported by Dunlap and York (2008) are not statistically significant, but even
null results contradict the theory and, as such, Dunlap and York (2008) assert
that their results challenge standard economic and sociological theories.
In addition to the ISSP and WVS, there is a third international data source,
namely, the EVS. Gelissen (2007) combined data from the 1999–2001 WVS
with the 1999 EVS and conducted a multilevel analysis based on the combined
data. His results also show a negative correlation between countries’ wealth
and environmental concern. However, he finds a positive correlation between
countries’ economic growth during the last decade and its citizens’ willingness
to pay. Thus, overall all three analyses using the ISSP, WVS, and EVS reach
contradictory or at least mixed results.
The differences among these results could have three different causes. First,
all three data sets consist of different country samples. The 2000 ISSP con-
tains mainly countries from the OECD and only a few nations from Asia
(Japan and the Philippines) or South America (Mexico and Chile), while the
WVS includes more countries from Asia and Africa, and the EVS contains
Willingness to Pay for Environmental Protection 641
only European countries (see Table 2). Second, the country-specific samples
of respondents were constructed differently in each survey. All the countries
that participated in the ISSP, WVS, and EVS agreed to conduct surveys ac-
cording to similar standards, including using random sampling of the adult
population, presenting the same questionnaire, and conducting either written
or personal (face-to-face) interviews. However, in some of the countries, these
methodological standards are difficult to meet and national reports on sur-
veying methods indicate that certain countries deviated from the agreed upon
standards. It must be considered that it is hard to conduct random sampling
in countries in which a large part of the population is not officially listed or
where a large portion of the population does not have a permanent residential
address. The WVS explicitly allows quota sampling when costs of random
sampling are prohibitively high. Even if random samples were achieved, the
response rates in the national surveys strongly depend on the resources in-
vested. Furthermore, survey results can change based on how the interviews
were conducted, for example, how well the interviewers were trained. Third,
measurement of environmental concern differs in studies conducted with the
ISSP, WVS, and EVS. In this study we concentrate on the first and third
points, the different selection of countries, and the different measurement of
environmental concern. A discussion of how each country constructed the
national sample and how the interviews were conducted is beyond the scope
of this article.
Comparisons of the ISSP, WVS, and EVS
The first concern in trying to compare the three data sets is that they were
not conducted simultaneously. The ISSP on the environment was conducted
in 1993 and 2000, the WVS in 1990–1993, 1995–1998, 1999–2001, and
2005–2008, and the EVS in 1999. Thus, a comparison among the three
surveys can best be accomplished for the 2000 ISSP, the 1999–2001 WVS,
and the 1999 EVS. The second challenge is locating a comparable measure of
environmental concern in the three surveys. The WVS and the EVS contain
two statements that respondents could agree or disagree with on a four-point
scale: “I would give part of my income if I were certain that the money
would be used to prevent environmental pollution” and “I would agree to
an increase in taxes if the extra money were used to prevent environmental
pollution.” Both items were analyzed in Dunlap and York (2008) and Gelissen
(2007) for the combined analysis of the EVS and the WVS. The 2000 ISSP
contains two similar, but not identical, items (see Table 1 for exact question
wording). The second item asked how willing respondents would be to pay
higher taxes in order to protect the environment. However, the first item
differs slightly more from the formulation in the WVS or EVS, asking how
willing respondents are to pay much higher prices in order to protect the
environment rather than to give “part of” their income as in the WVS or
642 Social Science Quarterly
TABLE 1
Items for Measuring Environmental Concern in the ISSP 2000, WVS 1999–2001,
and EVS 1999
Question Wording Answer Categories
Item 1: WVS/ EVS I would give part of my income if I
were certain that the money
would be used to prevent
environmental pollution.
(1) Strongly agree
(2) Agree
(3) Disagree
(4) Strongly disagree
Item 2: WVS/ EVS I would agree to an increase in
taxes if the extra money were
used to prevent environmental
pollution.
(1) Strongly agree
(2) Agree
(3) Disagree
(4) Strongly disagree
Item 1: ISSP How willing would you be to pay
much higher prices in order to
protect the environment?
(1) Very willing
(2) Fairly willing
(3) Neither willing nor
unwilling
(4) Fairly unwilling
(5) Very unwilling
Item 2: ISSP How willing would you be to pay
much higher taxes in order to
protect the environment?
(1) Very willing
(2) Fairly willing
(3) Neither willing nor
unwilling
(4) Fairly unwilling
(5) Very unwilling
EVS. The formulation “much higher prices” might imply a more substantial
amount than the expression “part of income” and could deter respondents
from agreeing with the statement. Nevertheless, the two items have a high
correlation in each survey. In the ISSP, the smallest correlation can be observed
in Japan (0.53) and the highest in Bulgaria (0.82). In the WVS, the two
items correlate lowest in Peru (0.33) and highest in Uganda (0.86), and
in the EVS the correlations vary from 0.40 in Iceland to 0.75 in Great
Britain. Moreover, an explorative factor analysis using the principal component
method with varimax rotation provides a one-dimensional solution for every
country, suggesting that the two items measure the same latent variable.
Therefore, we combined both items for every country by adding them up
to an index. This is of course a rather crude measure of the latent construct
“willingness to pay” or “environmental concern.” However, these are the only
two items that are contained in all three survey programs, thus providing the
only opportunity to compare the three surveys.
More problematic than these differences in the wording of the questions
is the fact that the ISSP uses five answering categories, whereas both the
WVS and EVS use four-point scales. Having more answer categories should
reduce the proportion of those agreeing with a statement since respondents
can disperse more over all available categories. This assumption is clearly
Willingness to Pay for Environmental Protection 643
supported if we look into the proportion of respondents agreeing in the three
largest economies—United States, China, and Japan. According to the results
obtained from the WVS, 62 percent of respondents in the United States agree
with both items, in China the proportion is 72 percent, and in Japan 60
percent. These are very high proportions that are not reached in any country
within the ISSP. The United States and Japan are also in the ISSP and here rates
of agreement are 33 and 38 percent, respectively—more than 20 percentage
points below the rates reported in the WVS. In the WVS, proportions of
respondents that agree to both statements range from 40 percent in Moldova
to 90 percent in Vietnam. In the ISSP, the range varies from 12 percent in
Finland to 38 percent in Japan. If we compare the seven countries that are
listed in the WVS and the ISSP (Canada, Chile, Japan, Mexico, Philippines,
Spain, United States), the average difference in the agreement rate is 31
percentage points. This comparison demonstrates that small changes in the
wording of questions or the number of answer categories can have considerable
consequences.
Also, comparison of results between the ISSP with the EVS reveals substan-
tial differences. The highest willingness to pay within the EVS can be observed
in Sweden where 77 percent of the respondents agreed on average with on
both items. The lowest level is measured in Lithuania where only 23 percent
agreed. There are 15 identical countries in both surveys. On average, the
willingness to pay is 28 percentage points higher in the EVS than in the ISSP.
Comparing all three surveys suggests that the variation in results is due to
the difference in the answering scale. The lack of a middle category in the
WVS and EVS seems to push respondents toward the agreement categories.
However, there are also some differences between the WVS and EVS. This
is demonstrated with the case of Spain, which is the only country that has
participated in all three surveys. Analysis of the ISSP results in an agreement
rate of 24 percent, the EVS reports 48 percent, and the WVS 52 percent.
Thus, the results for Spain reflect the fact that the willingness to pay for the
environment is highest in the WVS, followed by the EVS and the ISSP.
Considering these findings, we next use the complete information of the
answer scales instead of only the two agreeing categories to compare the three
surveys. We did this by treating the items as interval scales and recoding them
so that higher values indicate stronger agreement. We standardized them by
dividing the sum of the values obtained from both items by the sum of the
answer categories available. Thus, the new scale is standardized between 0
and 1, where 1 indicates the highest willingness to pay and 0 the lowest. On
average, the countries in the WVS reach a value of 0.58, the ISSP a value of
0.44, and the EVS a value of 0.51. Next, we calculated the bivariate correlation
between the measured willingness to pay and the purchasing power adjusted
GDP (PPP) for every survey separately (see Figure 2 where the index displayed
on the Y-axis is multiplied by 100).
Using the ISSP data, we find a positive correlation of 0.54, which is statisti-
cally significant (p=0.005), an analysis of the WVS generates a nonsignificant
644 Social Science Quarterly
FIGURE 2
The Correlation Between Wealth and Environmental Concern for the ISSP, EVS,
and WVS
negative correlation of −0.28 (p=0.177), and an analysis of the EVS data
results in a correlation of −0.04 (p=0.85). These results replicate previous
findings using the ISSP (Franzen, 2003; Franzen and Meyer, 2010) and the
WVS (Dunlap and York, 2008). Dunlap and York report a negative corre-
lation of −0.32 using the same measurement of environmental concern and
taking the natural logarithm of the countries’ GNP per capita. In our case,
taking the natural logarithm of the adjusted GDP (PPP) per capita does not
change the reported results. Also, using the Spearman rank correlation instead
of the Pearson correlation does not lead to any substantial differences in the re-
ported correlation or significance levels. The results also remain robust if other
waves of the ISSP (Diekmann and Franzen, 1999; Franzen, 2003; Franzen and
Meyer, 2010) or the WVS (Dunlap and York, 2008) are analyzed. Thus, an
analysis of the three international surveys shows that the ISSP data produce a
positive correlation, the EVS data produce no association, and the WVS data
produce a negative correlation. Hence, curiously, the three surveys generate
all three possible outcomes.
A possible solution to the puzzle is combining all countries from the three
data sets. However, the higher levels of agreement in the WVS and EVS
as compared to the ISSP, which are due to different measurements, are an
obvious problem when pooling the data. In addition, cross-cultural research
has often pointed out that some countries generally have a higher level of
acquiescence than Western cultures. One reason for this is that some countries
(presumably Asian countries) have a more collectivistic value orientation as
compared to the individualistic cultures of Western nations (Bosau, 2009;
Hofstede, 2001; Smith and Fischer, 2008; Van de Vijver and Leung, 1997).
Willingness to Pay for Environmental Protection 645
Thus, the higher level of acquiescence could be an additional explanation for
the higher willingness to pay in the WVS and EVS countries as opposed to
the countries contained in the ISSP. To test this assumption, we calculated
the level of acquiescence for every country in the three surveys. We did this,
as recommended in the literature (e.g., Hofstede, 1980; Matsumoto and Yoo,
2006; Smith, 2004), by selecting as many statements on diverse topics as
possible from each questionnaire, which respondents had to agree or disagree
with.
The 2000 ISSP contains 28 statements respondents could more or less
agree with in addition to the two items measuring the willingness to pay.
Fifteen of these 28 items are related to environmental issues; the other 13 are
related to questions of economic development or technology and science. We
constructed the acquiescence coefficient by first recoding the rating items so
that a value of zero is assigned to the most disagreeing category and a value
of four to the most agreeing category. Second, we summed up all answers
to the 28 items and divided it by the maximum possible sum. This way the
coefficient of acquiescence ranges from 0 to 1 for every respondent. A value
of zero denotes a respondent who never agreed to a statement, irrespective
of the content or formulation of the item. A value of 1 denotes the other
extreme, a respondent who agreed to every statement. The average of all
respondents in a given country is then an estimate of acquiescence in that
country (see Table 2). From the countries included in the ISSP, New Zealand
has the lowest acquiescence with a value of 0.47, and Portugal the highest,
0.66.1On average, the countries contained in the ISSP have a coefficient
of acquiescence of 0.54. Since more than half of the items considered for
measuring acquiescence are related to environmental issues, the index might
not be completely independent from the willingness to pay items. To test
whether the measure of acquiescence depends on the environmental topic of
the 2000 ISSP, we calculated acquiescence using the 1999 ISSP and the 2001
ISSP, which were on the topics of inequality and social networks, respectively.
The correlation of acquiescence is 0.78 when calculated with the 1999 and
2000 ISSP and 0.85 when calculated with the 2000 and 2001. Therefore, our
measure of acquiescence seems to be independent of the specific topic and
very reliable.
The WVS contains a total of 24 items that can be used to calculate the
coefficient of acquiescence. These items concern attitudes on employment,
political issues, gender equality, and religion. As expected, the average acqui-
escence for the countries in the WVS is 0.60 and thus higher than in the ISSP.
The lowest value of 0.53 is found in the United States and the highest value
of 0.68 is found in the Philippines. Finally, the EVS has 34 items suitable
for the calculation of the tendency of agreement that are on topics such as
1Note that Table 2 contains the average acquiescence when a country is contained in more
than one survey. Thus, the reported coefficients in the text might deviate from Table 2 if we
refer to survey specific acquiescence.
646 Social Science Quarterly
TABLE 2
The Willingness to Pay for Environmental Protection in the ISSP, WVS, and EVS
ISSP 2000 WVS 1999–2001 EVS 1999
Willingness Willingness Willingness GDP per
Number Country Nto Pay (%) Nto Pay (%) Nto Pay (%) Acquiescence Capita PPP
1 Albania (ALB) 939 63.37 0.65 3,864
2 Argentina (ARG) 1,252 46.33 0.60 9,189
3 Austria (AUT) 997 21.16 1,493 39.12 0.53 28,632
4 Bangladesh (BGD) 1,394 76.47 0.64 840
5 Belarus (BLR) 930 48.82 0.63 5,071
6 Belgium (BEL) 1,887 47.43 0.59 26,795
7 Bosnia (BIH) 1,170 70.77 1,170 70.77 0.61 4,353
8 Bulgaria (BGR) 946 18.29 944 49.36 0.62 6,200
9 Canada (CAN) 1,091 25.48 1,911 60.49 0.53 28,910
10 Chile (CHL) 1,437 31.32 1,181 65.37 0.58 9,479
11 China (CHN) 907 72.22 0.61 2,372
12 Croatia (HRV) 983 63.48 0.60 10,972
13 Czech Republic (CZE) 1,212 17.66 1,874 64.94 0.58 15,008
14 Denmark (DNK) 1,047 28.75 1,000 68.40 0.52 28,325
15 Estonia (EST) 953 33.68 0.61 9,894
16 Finland (FIN) 1,437 12.25 1,013 51.04 0.52 24,476
17 France (FRA) 1,591 39.03 0.61 25,938
18 Germany (DEU) 947 21.75 1,015 27.09 0.52 26,281
19 Great Britain (GBR) 956 32.11 940 49.68 0.52 25,673
20 Greece (GRC) 1,128 67.82 0.62 18,644
21 Hungary (HUN) 982 36.56 0.64 12,057
22 Iceland (ISL) 959 58.39 0.53 26,890
23 India (IND) 1,498 56.14 0.65 1,446
Willingness to Pay for Environmental Protection 647
TABLE 2—continued
ISSP 2000 WVS 1999–2001 EVS 1999
Willingness Willingness Willingness GDP per
Number Country Nto Pay (%) Nto Pay (%) Nto Pay (%) Acquiescence Capita PPP
24 Ireland (IRL) 1,192 34.82 988 40.59 0.53 28,768
25 Israel (ISR) 1,204 30.98 0.57 20,985
26 Italy (ITA) 1,951 44.95 0.61 24,431
27 Japan (JPN) 1,166 38.16 1,214 59.64 0.53 25,274
28 Kyrgyz (KGZ) 1,042 61.90 0.63 1,335
29 Latvia (LVA) 976 17.62 971 46.65 0.59 7,670
30 Lithuania (LTU) 939 22.90 0.64 8,417
31 Luxembourg (LUX) 1,172 55.89 0.59 55,151
32 Macedonia (MKD) 1,003 68.10 0.62 6,170
33 Malta (MLT) 992 49.09 0.61 18,190
34 Mexico (MEX) 1,249 32.83 1,453 62.84 0.59 10,647
35 Moldova (MDA) 921 40.50 0.61 1,472
36 Netherlands (NLD) 1,583 37.59 1,002 56.79 0.51 29,663
37 New Zealand (NZL) 1,092 31.87 0.47 19,333
38 Norway (NOR) 1,431 24.25 0.49 38,988
39 Peru (PER) 1,476 62.20 0.61 5,055
40 Philippines (PHL) 1,170 28.29 1,194 65.75 0.63 2,333
41 Poland (POL) 1,053 50.71 0.67 10,281
42 Portugal (PRT) 929 18.62 961 45.27 0.64 17,680
43 Romania (ROM) 992 46.88 0.68 6,181
44 Russia (RUS) 1,641 29.37 2,328 53.09 0.60 7,628
648 Social Science Quarterly
TABLE 2—continued
ISSP 2000 WVS 1999–2001 EVS 1999
Willingness Willingness Willingness GDP per
Number Country Nto Pay (%) Nto Pay (%) Nto Pay (%) Acquiescence Capita PPP
45 Serbia (SRB) 1,112 74.01 0.61 5,685
46 Singapore (SGP) 1,497 46.16 0.60 32,787
47 Slovakia (SVK) 1,288 41.77 0.62 11,229
48 Slovenia (SVN) 1,036 32.34 979 63.02 0.59 17,429
49 South Africa (ZAF) 2,804 42.76 0.60 6,552
50 South Korea (KOR) 1,138 55.89 0.59 16,456
51 Spain (ESP) 917 23.56 1,155 52.38 1,154 48.18 0.56 22,296
52 Sweden (SWE) 1,019 20.22 1,012 77.17 0.52 26,336
53 Switzerland (CHE) 975 35.28 0.50 31,001
54 Tanzania (TZA) 1,146 78.10 0.65 778
55 Turkey (TUR) 1,166 59.86 0.67 8,150
56 Uganda (UGA) 994 44.16 0.61 690
57 Ukraine (UKR) 1,110 52.79 0.65 3,317
58 USA (USA) 1,184 33.36 1,196 61.71 0.52 34,776
59 Vietnam (VNM) 961 90.22 0.64 1,420
60 Zimbabwe (ZWE) 955 48.59 0.59 247
Mean 27.12 61.04 50.64
NOTE:If a country is included in more than one survey, we report the average acquiescence.
Willingness to Pay for Environmental Protection 649
employment, religion, family, and political issues. Among the European na-
tions, the Netherlands shows the lowest acquiescence with a value of 0.49 and
respondents from Romania the highest level of 0.68. On average, the index
of acquiescence for countries contained in the EVS is 0.60. Comparing the
coefficient of acquiescence among countries shows high values of agreement
in Asian countries (India, Bangladesh, Vietnam, but not Japan) and in Eastern
European nations (Romania, Poland, and Turkey; see Table 2).
In order to compare individuals’ willingness to pay for the environment
cross-culturally, we now take the different levels of acquiescence into con-
sideration. Note that acquiescence might be due to two reasons. The first
is the methodological effect: respondents have a higher tendency to agree
when confronted with four answer categories instead of five. Second, there
are obvious cultural differences in respondents’ tendency to agree with a state-
ment. In order to compare all 60 countries, we took the standardized index
of a country’s willingness to pay and weighted it by the inverse of the coeffi-
cient of acquiescence. Thus, we measure the willingness to pay in relation to
respondents’ overall tendency of agreement. Theoretically, the ratio can vary
between zero and infinity. A value of zero indicates a respondent who answered
“strongly disagree” or “very unwilling” to both items. The value approaches
infinity if a respondent disagrees with all rating questions but shows the max-
imum willingness to pay.2For example, the United States has an average
willingness-to-pay-value of 0.53.3This willingness to pay is compared to the
general agreement tendency (acquiescence) by multiplying it by the inverse of
the acquiescence level (0.53/0.52 =1.02). Thus, for the United States, this
weighting results in a value very close to 1, indicating a similar willingness to
pay as compared to the general acquiescence. The correlation of the weighted
willingness to pay with countries’ wealth (purchasing power adjusted GDP
per capita) for all 59 countries from the ISSP, WVS, and EVS4results in
a statistically significant value of r=0.46. This correlation is depicted in
Figure 3.5
The data sets for the 60 countries do not only contain information about
environmental concern but also data on the sociodemographic characteristics
of respondents. In addition, more statistical information about the countries’
characteristics is available from the United Nations Development Programme
or the European Commission. The data can be analyzed using multilevel
analysis (Snijders and Boske, 1999; Rabe-Hesketh and Skrondal, 2008). At
the individual level (level 1), previous studies (e.g., Franzen and Meyer, 2010)
2Theoretically, an undefined division of zero can occur. However, the lowest value of
acquiescence observed in the data is 0.02.
3The average of the willingness to pay from the ISSP and the WVS is 0.53.
4Luxembourg is an outlier with respect to GDP and is therefore not included in Figure 3.
5A different measurement of acquiescence and a different weighting procedure is used in
Franzen and Vogl (2011). The methods used here use the full information of the answering
scales whereas Franzen and Vogl (2011) treat the agreement items used for the measurement
of acquiescence as ordinal scales. However, the results are affected only marginally and do not
depend on the difference in measurement of acquiescence.
650 Social Science Quarterly
FIGURE 3
The Correlation Between Environmental Concern and Wealth in the 59 Countries
from the ISSP, WVS, and EVS
show that respondents’ income, education, and age should affect environmen-
tal concern. The wealth effect discussed above should not only affect environ-
mental concern at the macro level of countries as demonstrated in Figure 3, but
it should also explain the interindividual differences found within countries.
In the analyses that follow, we calculate individuals’ household equivalent
income by dividing the households’ income by the square root of the number
of individuals living in one household. This procedure has the advantage of
measuring the standard of living more accurately than if only personal incomes
were included. Taking the household income also allows us to measure the
standard of living for respondents who are not active in the labor market or
have no personal income. Instead of trying to adjust these personal incomes
by the purchasing power, we conducted a z-standardization with the income
variable and measured the standardized difference from the country’s average
income for every respondent. This procedure allows us to measure the income
position of a respondent relative to the average income of the country in which
he/she lives.
In addition, environmental concern depends on the level of education. Well-
educated respondents are more informed about the state of the environment,
which should also increase the concern for and understanding of environmen-
tal protection. People with little information and knowledge about the state
of the planet will not be as concerned with the environment. Furthermore,
Willingness to Pay for Environmental Protection 651
older people should have lower concern for the environment than younger
people. This should be due to a cohort effect rather than to the effect of aging
itself. When growing up, younger cohorts were more exposed to environmen-
tal concern through public discussions, political debates, and media reports
than older cohorts. Accordingly, younger generations should be more sensitive
to environmental issues. The literature also points to the effect of gender on
environmental concern (see Blocker and Eckberg, 1997; Wilson et al., 1996).
Some studies find concern higher among women than men. Previous studies
also report higher levels of environmental concern among individuals with
a postmaterialistic value orientation (Inglehart, 1995) and we also include a
measure of postmaterialistic values (see Table 3).
At the macro level, certain characteristics of countries could influence en-
vironmental concern in addition to wealth. Since some of these variables are
correlated with wealth, they must be included in the analysis as control vari-
ables in order to avoid an estimation bias with respect to the wealth effect. In
addition to overall wealth, the distribution of it could matter as well. A high
level of inequality could direct public attention more toward economic issues
and redistribution. Such goals could be in competition with environmental
issues. To test this hypothesis, we calculated the Gini-coefficient from the sur-
veys’ income data. We expect that the Gini-coefficient is negatively correlated
with the willingness to pay for better environmental protection.
Environmental quality could also affect environmental concern. Low local
environmental quality should sensitize respondents and increase the willing-
ness to pay to improve it. We therefore included the environmental sustain-
ability index from 2001 calculated jointly by the Yale Center of Environmental
Law and Policy (YCELP), the Center for International Earth Science Informa-
tion Network of Columbia University (CIESIN), the World Economic Forum
(WEF), and the European Commission (EU). The index consists of different
subscales from which we took a country’s index of air and water quality as
well as the index of biodiversity and soil erosion. A high value indicates that a
country’s environmental quality is above average. Hence, if objective environ-
mental quality matters, it should be negatively correlated with the willingness
to pay.
In most countries environmental quality is not evenly distributed but is
worse in urban areas than in the sparsely populated countryside. As a general
rule, heavily populated regions are supposed to have poorer air and water
quality. Therefore, we also included the population density of a country as well
as the proportion of the population living in cities. The denser the population
and the higher the proportion of respondents living in urban areas, the worse
the objectively and subjectively perceived environmental quality should be.
Thus, higher population density and a higher proportion of inhabitants living
in cities should increase the willingness to pay for environmental protection.
Because we expect both individual effects (level 1) and country-specific
effects (context effects or level 2), hierarchical linear regression analysis is a
suitable statistical tool that allows the simultaneous estimation of micro- and
652 Social Science Quarterly
TABLE 3
Description of the Variables
Variable Min. Max. Description Data Source
Individual-level
variables
Sex 0 1 0 =male, 1 =female ISSP, WVS, EVS
Age 18 80 Age in years ISSP, WVS, EVS
Postmaterialism 0 2 Number of postmaterilistic goals a country should have from a list of four:
0=none, 1 =one, 2 =two
ISSP, WVS, EVS
Education 1 4 Education is classified into four categories: ISSP, WVS, EVS
1=Primary education and no formal education
2=Secondary education
3=Degree before university entrance
4=University education
Relative household
income
−1.98 13.62 Household income divided by the square root of the number of persons living
in the household, z-transformed
ISSP, WVS, EVS
Acquiescence 0.18 1.00 Sum of answered categories of four- or five-point scales divided by the highest
possible score
ISSP, WVS, EVS
Country-level variables
GDP per capita (PPP) 0.25 38.99 Per capita GDP 2000 converted to measure the purchasing power in each
country in 1,000s of US$
International Monetary Fund (IMF) World
Economic Outlook Database, April 2009
Population density 0.003 5.88 Number of 1,000 inhabitants per square kilometer of country United Nations: World Population
Prospects
Proportion urban
population
12 100 Proportion of population living in areas classified as urban according to the
criteria used by each country
UN-DATA, WHO-Data
Environmental quality −0.81 1.36 Part of the Environmental Sustainability Index 2001 (Environmental Systems
Component), measures air quality, water quality, water availability,
biodiversity, and terrestrial systems; mean of these five sectors
YCELP, CIESIN, WEF, and the Joint
Research Centre of the European
Commission
z-transformed
Gini-coefficient 0.21 0.69 Measures income inequality: 0 =equal distribution, Own calculations with ISSP, WVS, and
EVS data
1=maximally unequal distribution
Willingness to Pay for Environmental Protection 653
macro effects (Gelman and Hill, 2007; Snijders and Bosker, 1999; Rabe-
Hesketh and Skrondal, 2008). We apply a varying-intercept model and es-
timate coefficients via the maximum likelihood method.6The willingness
to pay Yij depends on the characteristics (x1through x8) of the individu-
als (ito n) as denoted by Equation (1). The country-specific characteristics
(z1through z5) are incorporated by varying the intercept β0jdepending
on the macro-level variables of the jto kcountries. This is formulated in
Equation (2):
Yij =β0j+β1x1ij +···+β8x8ij +εij (1)
β0j=γ00 +γ01z1j+ ···+γ05 z5j+ζj.(2)
In principle, the model could be extended in order to also consider cross-
level interaction effects so that the slope of the individual effects may depend
on context effects (varying slope model). However, our main focus is the
estimation of the macro-level effect of countries’ wealth on the willingness
to pay and not the possibly varying slopes of individual effects. Furthermore,
investigating cross-level effects requires well-founded theoretical hypotheses.
Otherwise, the number of all empirically possible cross-level effects is too large
and the selection arbitrary. Therefore, the analyses are restricted to the varying-
intercept model, which suits our purpose and the theoretical hypotheses.
We start the statistical analyses using the 2000 ISSP data (first column of
Table 4). The first step in multilevel analysis is to test via a likelihood-ratio
test for any variance at the context level. If the null hypothesis of no variance
cannot be rejected, then there is no need to consider any context effects. In
our case the χ2-value is highly significant (p=0.00) indicating that some
context effects should be taken into account. Next, we estimate the intraclass
correlation coefficient (ICC) of the null model. The ICC indicates how much
variation of the dependent variable is due to the macro variables and how much
of it is due to the micro variables. The ICC of the null model results in a value
of 0.06 when taking only the 24 countries of the ISSP into consideration.7
The coefficient indicates that a relatively small proportion (6 percent) of the
overall variance is due to the macro-level variation and that the rest is due
to the variation at the micro level. The results of the first model using only
the ISSP data indicate that respondents with higher educational degrees and
higher relative income are more willing to pay for environmental protection.
Gender and age do not have statistically significant effects on the willingness
to pay. Furthermore, Model 1 shows that postmaterialistic values are also
positively related to the willingness to pay. These results replicate existing
6We used the statistical software STATA 11.1.
7The ISSP 2000 had 25 participating nations. Some data (household size) are missing for
Ireland, which is necessary for calculating the income variable. The sample is therefore reduced
to 24 countries.
654 Social Science Quarterly
TABLE 4
Multilevel Analysis of the Willingness to Pay (WtP) for the Environment
Model 2 Model 3 Model 4
Model 1 WtP: ISSP/ WtP: ISSP/ Weighted WtP
WtP: ISSP WVS/EVS WVS/EVS ISSP/WVS/EVS
Individual-level variables
Sex (1 =female) −0.40 0.40∗0.50∗3.75∗∗∗
(0.38) (0.20) (0.20) (0.47)
Age in years (18–80) −0.006 −0.02∗∗∗ −0.05∗∗∗ −0.32∗∗∗
(0.01) (0.007) (0.007) (0.02)
Secondary education 4.49∗∗∗ 3.53∗∗∗ 3.96∗∗∗ 11.73∗∗∗
(0.50) (0.27) (0.27) (0.64)
High school diploma 6.63∗∗∗ 5.20∗∗∗ 5.88∗∗∗ 17.09∗∗∗
(0.64) (0.29) (0.29) (0.69)
University degree 10.73∗∗∗ 8.47∗∗∗ 9.45∗∗∗ 28.91∗∗∗
(0.62) (0.35) (0.35) (0.82)
Relative income 2.07∗∗∗ 1.06∗∗∗ 1.22∗∗∗ 3.72∗∗∗
within country (0.21) (0.10) (0.10) (0.25)
Postmaterialism 6.02∗∗∗ 4.48∗∗∗ 4.72∗∗∗ 11.58∗∗∗
(0.33) (0.17) (0.17) (0.40)
Acquiescence 28.45∗∗∗
(1.12)
Data set WVS 11.84∗∗∗ 10.91∗∗∗ −0.33
(0.43) (0.43) (1.01)
Data set EVS 10.96∗∗∗ 9.76∗∗∗ −3.42∗∗∗
(0.34) (0.34) (0.79)
Country-level variables
GDP (PPP) in 1,000 0.38∗0.18 0.28∗1.52∗∗∗
(0.15) (0.12) (0.12) (0.30)
Proportion urban pop. 0.06 0.02 0.02 0.05
(0.09) (0.06) (0.06) (0.16)
Population density 27.52∗∗ −1.39 −1.80 −7.67∗∗
(10.47) (1.16) (1.20) (2.91)
Environmental quality −2.03 −2.87 −2.28 0.06
(2.92) (2.10) (2.17) (5.30)
Gini-coefficient 22.67 15.56 17.64 44.00
(14.72) (9.29) (9.59) (23.34)
Constant 13.31 27.28∗∗∗ 9.20 50.16∗∗
(8.27) (6.22) (6.46) (15.63)
Standard deviation
Country level 4.20∗∗∗ 5.82∗∗∗ 6.01∗∗∗ 14.65∗∗∗
Individual level 25.77∗∗∗ 24.78∗∗∗ 24.65∗∗∗ 58.49∗∗∗
Intraclass correlation (ICC)
Null model 0.061 0.058 0.058 0.12
Model with covariates 0.026 0.052 0.056 0.059
Explained variance
Country level 0.61 0.16 0.10 0.56
Individual level 0.05 0.06 0.07 0.07
Number of countries 24 51 51 51
Number of observations 19,300 64,341 64,341 64,341
NOTE:Reported are the unstandardized regression coefficients. Numbers in parentheses
denote the standard errors of the coefficients.
∗p<0.05; ∗∗p<0.01; ∗∗∗ p<0.001.
Willingness to Pay for Environmental Protection 655
findings (Franzen and Meyer, 2010; Xiao and Dunlap, 2007; Marquart-Pyatt,
2008).
At the macro level, the finding that countries’ wealth (measured by the pur-
chasing power adjusted GDP) is positively related to environmental concern is
also replicated in our analysis (Franzen and Meyer, 2010). Thus, respondents
from wealthier countries are more willing to pay higher prices or higher taxes
for improved environmental protection. However, the effect of GDP is rather
small. An increase of $1,000 per capita results in an increase of the willingness
to pay by 0.38 points on a scale that varies from 0 to 100. In comparison,
population density has a strong positive influence on the willingness to pay.
In contrast, inequality, environmental quality, and the proportion living in
urban areas are not statistically related to the willingness to pay.
Next, Model 2 takes all 51 countries from the three surveys into consid-
eration.8At the individual level, all findings remain robust as in Model 1.
Additionally, women have a slightly higher willingness to pay, whereas older
people are less willing to do so. Thus, the pooled data analysis of more than
64,000 respondents in 51 countries shows the expected effects of income,
education, and age on the willingness to pay. At the macro level, the wealth
of nations has a positive, though insignificant, effect on the willingness to
pay and the effect of population density disappears when the whole country
sample is taken into account.
Model 3 in Table 4 takes acquiescence into account by including it as an
individual-level variable. As described above, the variable contains an acqui-
escence value for every individual that ranges from 0 to 1. Acquiescence has
a strong influence on the willingness to pay. Since it is coded between 0 and
1, the coefficient shows the difference of an individual with the lowest and
highest acquiescence with respect to the willingness to pay. However, more
importantly, the effect of GDP per capita is statistically significant when we
control for acquiescence. Thus, the multivariate analysis mirrors the bivari-
ate correlation results between GDP and the willingness to pay as shown in
Figure 3.
Finally, Model 4 demonstrates an alternative procedure for incorporating
acquiescence. It controls for acquiescence by weighting the dependent variable
instead of including it as an independent variable. The weighting of the
individual responses was done analogously to the weighting of countries in
Table 2. Hence, we took a respondent’s willingness to pay (index from 0 to
1) and multiplied it by the inverse of his/her acquiescence value. The results
basically remain the same as the results of Model 3. Note, however, that
the size of coefficients cannot be compared to Model 3 since the dependent
variable is not the willingness to pay but the weighted willingness to pay. The
8Bosnia, Kyrgyz Republic, Luxembourg, Malta, Macedonia, Serbia, and Tanzania miss data
on environmental quality, population density, and income and had to be dropped from the
analysis. Vietnam and Bangladesh show extremely high willingness to pay values. They were
identified as outliers by a whiskers box plot and were therefore dropped from the analyses.
656 Social Science Quarterly
difference between Models 3 and 4 is that the former estimates the influence
of acquiescence on the willingness to pay while the latter assumes that the
willingness to pay should be seen in relation to it. Therefore, Model 3 makes
less far reaching assumptions than Model 4.
Examining the explained variances of the different models at the macro
level shows that the countries’ GDP explains a relatively large proportion of
the variance (e.g., 61 percent of the ISSP differences and 10 percent if all
countries are included). Explained variance on the individual level is much
lower (7 percent).
Conclusion and Discussion
This article analyzes why different studies scrutinizing the determinants of
environmental concern in cross-cultural perspective come to different conclu-
sions. On the one hand, studies based on the ISSP support the wealth effect
(Diekmann and Franzen, 1999; Franzen, 2003; Franzen and Meyer, 2010).
On the other hand, Dunlap and York (2008) find higher environmental con-
cern in poorer nations when analyzing the WVS or EVS; thus refuting the
wealth effect. In this article, we measure environmental concern using two
items that ask respondents whether they would be willing to pay higher prices
and higher taxes in order to improve the environment. These two items are
contained in an almost identical format in all three surveys, and therefore
allow comparison across the three surveys. We first analyze the three data
sets separately and replicate previous findings, showing that there is a positive
correlation between countries’ wealth and inhabitants’ environmental concern
using the ISSP, a negative correlation using the WVS, and no correlation using
the EVS. Willingness to pay is comparatively higher in the WVS and EVS
than in the ISSP, likely due to a slight variation in the answering scales of the
surveys (four-point vs. five-point scales), differences in the sample of countries
in each survey, and varying levels of acquiescence in each country.
If acquiescence is taken into consideration, the analysis of the pooled data of
59 countries shows a positive and statistically significant correlation between
the countries’ wealth and their environmental concern. This fundamental
result is robust when we apply multilevel analysis to the data and take further
individual- and country-level effects into consideration. On the individual
level, respondents’ relative income position, their education, and age affect
their willingness to pay. At the macro level, willingness to pay is determined
by countries’ wealth.
Overall, our analyses of the pooled data from the ISSP, WVS, and EVS
support the wealth hypothesis and refute the conclusions of Dunlap and
York (2008). The puzzle of contradictory findings is thus resolved when
the countries’ acquiescence is incorporated into the analysis. Many poorer
nations in Asia and Eastern Europe have a stronger collective culture, which is
manifested as a high tendency to agree to all kinds of statements. This tendency
Willingness to Pay for Environmental Protection 657
of general agreement can also be observed in industrialized countries. However,
the industrialized nations generally have a stronger individualistic culture and
respondents seem to be less sensitive to acquiescence. These cultural differences
have to be taken into consideration in cross-cultural comparisons. When doing
so, the combined data clearly shows that wealth increases the willingness to
pay for environmental protection as predicted by standard economic theory.
It is often argued (e.g., Brechin and Kempton, 1994) that “willingness to
pay” is not an ideal measurement of environmental concern. Critics point out
that inhabitants of developing countries are so poor that they have no financial
resources left that could possibly be devoted to environmental protection.
Thus, they might be concerned but are unable to pay anything. This argument
might indeed apply to some very poor countries such as Uganda, which had
a PPP of $690 per capita in 2000. However, our sample of countries does not
consist of either very poor or very rich countries, and the correlation reported
in Figure 3 holds cross-nationally for the entire wealth range.
However, willingness to pay is only one dimension of environmental con-
cern. There are other dimensions, such as cognitive and affective components.
Previous research demonstrates that the positive correlation of wealth and en-
vironmental concern holds even if the analysis is restricted to items that do not
deal with the willingness to pay, at least with the ISSP data (see Franzen and
Meyer, 2010). Thus, there is some evidence suggesting that our results hold
even if different measurements of environmental concern are used. Unfortu-
nately, the comparison of the ISSP, WVS, and EVS as intended in this article
is only possible with respect to the willingness to pay items. Therefore, an
empirical investigation of to what extent the wealth effect holds with respect
to alternative measurements must be left to future research.
It is difficult to judge what this finding implies for the future state of the en-
vironment. On the one hand, the perspective that people will be more willing
to pay for environmental protection if they become wealthier is promising for
anybody who is concerned with environmental protection. Since the world has
grown richer in the past and will most likely continue to do so in the future,
further increases in wealth should be accompanied by increases in environ-
mental concern. On the other hand, if economic development is a prerequisite
for higher levels of environmental concern, and given that economic growth
is usually linked to environmental destruction, the environment might decay
before concern increases. The challenge is, therefore, to disconnect economic
growth from environmental destruction.
REFERENCES
Blocker, T. Jean, and Douglas Lee Eckberg. 1997. “Gender and Environmentalism: Results
from the 1993 General Social Survey.” Social Science Quarterly 78:841–58.
Bosau, Christian. 2009. Arbeitszufriedenheitsmessungen im interkulturellen Vergleich. Disserta-
tion, University of Cologne, Germany.
658 Social Science Quarterly
Brechin, Steven R., and Medani Bhandari. 2011. “Perceptions of Climate Change Worldwide.”
WIREs Climate Change 2:871–85.
Brechin, Steven R., and Willett Kempton. 1994. “Global Environmentalism: A Challenge to
the Postmaterialism Thesis?” Social Science Quarterly 75:245–69.
Commission of the European Communities. 2008. 20 20 by 2020. Europe’s Climate Change
Opportunity. Brussels. Available at http://ec.europa.eu/commission_barroso/president/pdf/
COM2008_030_en.pdf.
Diekmann, Andreas, and Axel Franzen. 1999. “The Wealth of Nations and Environmental
Concern.” Environment and Behavior 31:540–49.
Dunlap, Riley, and Richard York. 2008. “The Globalization of Environmental Concern and the
Limits of the Postmaterialist Values Explanation: Evidence from Four Multinational Surveys.”
Sociological Quarterly 49:529–63.
Franzen, Axel. 2003. “Environmental Attitudes in International Comparison: An Analysis of
the ISSP Surveys 1993 and 2000.” Social Science Quarterly 84:297–308.
Franzen, Axel, and Reto Meyer. 2010. “Environmental Attitudes in Cross-National Perspective:
A Multilevel Analysis of the ISSP 1993 and 2000.” European Sociological Review 26:219–34.
Franzen, Axel, and Dominikus Vogl. 2011. “Pitfalls of International Comparative Research:
Taking Acquiescence into Account.” Journal of Economics and Statistics 231:761–82.
G8 Declaration. 2009. Chair´s Summary of the G8 L’Aquila Summit. Presse- und Information-
samt der Bundesregierung. Available at http://www.bundesregierung.de/Content/DE/Artikel/
2009/G8/2009-07-08-dokumente.html.
Gelissen, John. 2007. “Explaining Popular Support for Environmental Protection: A Multilevel
Analysis of 50 Nations.” Environment and Behavior 39:392–415.
Gelman, Andrew, and Jennifer Hill. 2007. Data Analysis Using Regression and Multi-
level/Hierarchical Models. Cambridge: Cambridge University Press.
Hofstede, Gert H. 1980. Culture’s Consequences: International Studies of Work-Related Values.
Newbury Park, CA: Sage Publications.
———. 2001. Culture’s Consequences: Comparing Values, Behaviors, Institutions and Organiza-
tions Across Nations. Thousand Oaks, CA: Sage Publications.
Inglehart, Ronald. 1995. “Public Support for Environmental Protection. Objective Problems
and Subjective Values in 43 Societies.” Political Science and Politics 28:57–72.
Intergovernmental Panel on Climate Change. 2007. Climate Change 2007: The Physical Science
Basis: Contribution of Working Group I to the Fourth Assessment Report of the Intergovernmental
Panel on Climate Change. Cambridge: Cambridge University Press.
Marquart-Pyatt, Sandra T. 2008. “Are There Similar Sources of Environmental Concern?
Comparing Industrial Countries.” Social Science Quarterly 89:1312–35.
Matsumoto, David, and Seung Hee Yoo. 2006. “Toward a New Generation of Cross-Cultural
Research.” Perspectives on Psychological Science 1:234–50.
Rabe-Hesketh, Sophia, and Anders Skrondal. 2008. Multilevel and Longitudinal Modeling Using
Stata, 2nd ed. College Station, TX: Stata Press.
Smith, Peter B. 2004. “Acquiescence Response Bias as an Aspect of Cultural Communication
Style.” Journal of Cross-Cultural Psychology 35:50–61.
Smith, Peter B., and Ronald Fischer. 2008. “Acquiescence, Extreme Response Bias and Culture:
A Multilevel Analysis.” Pp. 288–311 in Fons J. R. van de Vijver, Dianne A. van Hemert, and
Ype H. Poortinga, eds., Multilevel Analysis of Individuals and Culture. New York: Erlbaum.
Willingness to Pay for Environmental Protection 659
Snijders, Tom A., and Roel J. Bosker. 1999. Multilevel Analysis: An Introduction to Basic and
Advanced Multilevel Modeling. London: Sage.
Van de Vijver, Fons, and Kwok Leung. 1997. Methods and Data Analysis for Cross-Cultural
Research. Thousand Oaks, CA: Sage Publications.
Wilson, Margo, Martin Daly, Stephen Gordon, and Adelle Pratt. 1996. “Sex Differences in
Valuations of the Environment?” Population and Environment 18:143–59.
Xiao, Chenyang, and Riley E. Dunlap. 2007. “Validating a Comprehensive Model of Envi-
ronmental Concern Cross-Culturally.” Social Science Quarterly 88:471–93.