Journal of Environmental Psychology: Lucy J. Hawcroft, Taciano L. Milfont

You might also like

Download as pdf or txt
Download as pdf or txt
You are on page 1of 16

Journal of Environmental Psychology 30 (2010) 143–158

Contents lists available at ScienceDirect

Journal of Environmental Psychology


journal homepage: www.elsevier.com/locate/jep

The use (and abuse) of the new environmental paradigm scale over the last 30
years: A meta-analysisq
Lucy J. Hawcroft a, Taciano L. Milfont b, *
a
University of Auckland, New Zealand
b
Centre for Applied Cross-Cultural Research, School of Psychology, Victoria University of Wellington, PO Box 600, Wellington 6410, New Zealand

a r t i c l e i n f o a b s t r a c t

Article history: This paper reports a meta-analysis of studies using the New Environmental Paradigm (NEP) Scale over
Available online 15 October 2009 the last 30 years. A review of 69 studies from 36 countries (including 58,279 participants from 139
samples) shows that there is considerable variation in the way the NEP Scale is used, particularly with
Keywords: regards to the number of items used and the number of points on the Likert scale employed. Results from
New ecological paradigm scale weighted regression analyses reveals that variations in sample type and scale length have a significant
NEP Scale
effect on NEP scores. In particular, environmentalist and white-collar samples scored significantly higher
Environmental attitudes
Meta-analysis
on the NEP Scale than nationally or regionally representative samples, while blue-collar samples scored
significantly lower; and participants scored higher on 6-item versions of the scale than on the revised 15-
item version, and lower on versions of the scale containing 5, 7, 8 or 10 items. Implications of this
research for the comparability of previous studies using the NEP Scale are discussed and guidelines for
future research are presented.
Ó 2009 Elsevier Ltd. All rights reserved.

1. Introduction environmental attitudes in a valid and reliable fashion. Environ-


mental attitudes (EA) are a psychological tendency expressed by
More than 50 years ago commentators and natural scientists evaluating the natural environment with some degree of favour or
began to systematically investigate ways in which human behavior disfavour, and are a crucial construct in the field of environmental
were damaging the natural environment (e.g., Carson, 1962; Osborn, psychology, discussed in more than half of all publications in this
1948). Their work was soon followed by social scientists who hoped area (Milfont, 2007a). However, it has frequently been noted that
to understand more about how humans relate to the environment, much of the previous research on EA has been conducted in an
and thus uncover techniques that could be used to encourage unsystematic way (Heberlein, 1981; Stern, 1992), with researchers
people to live in a more sustainable manner (for a review, see generating new measures of EA for almost every study they con-
Gardner & Stern, 2002). However, although social scientists have ducted. As a result, in their 2002 review Dunlap and Jones estimated
made many significant discoveries, it is still clear that humanity is that at least several hundred measures of EA have been developed
not living in balance with our natural limits. In fact, recent research since the 1960s. Analysts agree that this ‘‘anarchy of measurement’’
shows that humans are consuming the Earth’s resources at (Stern, 1992, p. 279) has been a key factor contributing to the
increasingly unsustainable rates (Millennium Ecosystem Assess- noncumulative and atheoretical nature of much research on EA
ment, 2005). As a consequence of this maladaptive behavior we are (Milfont & Duckitt, 2004b). The proliferation of measures has also
also facing ever more serious environmental issues, such as climate made it difficult for researchers who are not familiar with the field to
change (Intergovernmental Panel on Climate Change, 2007). choose an appropriate standard measure of EA to use when
Given this situation, it is crucial that social scientists gain a better designing a new study, thus compounding the problem.
understanding of why people treat the environment as they do. An However, although many researchers have noted the detrimental
important step towards achieving this goal is to measure peoples’ effect on research in this field of using numerous measures of EA, less
attention has been paid to the question of whether researchers are
using standardized measures of EA consistently. It is also not known
q Portions of this paper were presented at the Annual Meeting of the Society of
whether an inconsistent use of standardized EA measures may
Australasian Social Psychologists, Wellington, New Zealand, March 2008.
* Corresponding author. Tel.: þ64 4 463 6398; fax: þ64 4 463 5402.
constitute a problem for the field. Research into psychometrics
E-mail address: taciano.milfont@vuw.ac.nz (T.L. Milfont). suggests that an inconsistent use of scales may well be problematic.
URL: http://www.milfont.com For example, it is known that even a small change in the wording of

0272-4944/$ – see front matter Ó 2009 Elsevier Ltd. All rights reserved.
doi:10.1016/j.jenvp.2009.10.003
144 L.J. Hawcroft, T.L. Milfont / Journal of Environmental Psychology 30 (2010) 143–158

one item can have a substantial effect on how people respond to the development of the NEP Scale was the authors’ recognition
a scale (Robinson, Shaver, & Wrightsman, 1991). Changes in the that it was possible to identify an emerging ecocentric system of
response format (e.g., variation in the number of points on a Likert beliefs (i.e., humans are seen as being part of natural systems and
scale offered to participants) are also known to effect responses constrained by that fact) that challenged the dominant anthro-
(Aiken, 1987; Krosnick, 1999). Finally, the context in which items are pocentric system of beliefs current in Western societies (i.e.,
presented (i.e., the items or text that come before or after them) can humans are seen as being independent from, and superior to,
also have a considerable effect on the way people respond to scales other organisms in nature) (Dunlap & Van Liere, 1984; Pirages &
(Schwarz, 1999). Therefore, it seems likely that if there is variation Ehrlich, 1974). These two systems were respectively named the
across studies in the way that standardized measures of EA are used New Environmental Paradigm (NEP) and the Dominant Social
this might decrease the comparability of the studies’ results, leading Paradigm (DSP).
to the same problem of noncumulative research noted above. It is worth noting that the NEP and DSP are theoretically related
When trying to measure the effect of such variation in scale use it to Schwartz’s (1999) harmony–mastery cultural value dimension.
is important to control for other factors that might also affect Following the seminal work of Florence Kluckhohn (1953), Schwartz
results. For example, recent research has found that sampling argues that this value dimension is a result of one of the basic issues
fluctuation (e.g., sample size, sample composition) has a significant that societies must confront: the relation of humans to the natural
influence on the internal structure of the values domain (Fischer, and social environment. This issue leaves societies with two solu-
Milfont, & Gouveia, in press; Fontaine, Poortinga, Delbeke, & tions in order to regulate human activity: either to fit harmoniously
Schwartz, 2008). It is thus important to determine whether changes into the world, trying to preserve it (harmony values, or the NEP
in results across studies are caused by variations in scale use rather worldview), or to exploit and change the world (mastery values, or
than by variations in sample composition or other factors that the DSP worldview). Empirical results have supported this rela-
might affect responding. A meta-analysis provides an ideal method tionship by showing that national–level NEP scores correlate posi-
to examine this possibility as it summarizes many aspects of several tively with harmony values across 27 countries (Milfont, Hawcroft,
studies in a quantitative form (Rosenthal & DiMatteo, 2001). This & Fischer, 2008). It therefore seems likely that the NEP Scale taps an
article therefore takes a meta-analytic approach to investigate how important aspect of human beliefs. There are several different
the use of the most widely-used EA measure, the New Environ- extant versions of the NEP Scale, which are described below.
mental Paradigm (NEP) Scale, may have affected the results
obtained by researchers using the scale. By investigating the use of 2.1. The original NEP Scale
one scale in detail, this article will highlight some of the more
problematic aspects of how EA is currently measured. Some The original NEP Scale was published in 1978 by Dunlap and Van
recommendations to researchers as to how the NEP Scale can be Liere, and consists of 12 items (8 pro–trait and 4 con–trait)
best used to avoid such problems in the future are also provided. responded to on a 4–point Likert scale (anchored by strongly agree
to strongly disagree). The items were developed based on both the
2. The NEP Scale literature on the emergence of the NEP worldview and consultation
with environmental experts. The items are intended to tap three
Despite the large number of EA measures available, reviewers main facets of EA: a belief in (1) humans’ ability to upset the
agree that only three have been widely used (Dunlap & Jones, 2003; balance of nature, (2) the existence of limits to growth, and (3)
Fransson & Gärling, 1999). These are the Ecology Scale (Maloney & humans’ right to rule over the rest of nature. Table 1 displays the
Ward, 1973; Maloney, Ward, & Braucht, 1975), the Environmental items. The NEP Scale measures the overall relationship between
Concern Scale (Weigel & Weigel, 1978), and the NEP Scale (Dunlap & humans and the environment, and is normally used as a unidi-
Van Liere, 1978; Dunlap, Van Liere, Mertig, & Jones, 2000). These mensional measure of EA. Higher NEP scores indicate an ecocentric
three scales examine multiple phenomena or expressions of orientation reflecting commitment to the preservation of natural
concern, such as beliefs, attitudes, intentions and behavior. These resources, and lower NEP scores indicate an anthropocentric
scales also examine concerns about various environmental topics, orientation reflecting commitment to exploitation of natural
such as pollution and natural resources. Because they focus on resources.
multiple environmental issues and multiple expressions of concern, To validate the NEP Scale, Dunlap and Van Liere (1978) surveyed
these measures are all multiple–topic/multiple-expression assess- two samples in their 1976 Washington State survey: one general
ment techniques (Dunlap & Jones, 2002). Although widely used, population sample (n ¼ 806, Sample 1), and another sample from
both the Ecology Scale and the Environmental Concern Scale include members of an environmental organization (n ¼ 407, Sample 2).
items referring to specific environmental topics that have become Five main findings are reported demonstrating the reliability and
dated as new issues emerge (Dunlap & Jones, 2002, 2003). The NEP validity of the NEP Scale. First, results indicated that the environ-
Scale avoids this issue by measuring general beliefs about the rela- mentalist sample scored significantly higher (M ¼ 43.8, p < .001)
tionship of human beings to the environment. than the general population sample (M ¼ 36.3), providing known–
The universal nature of the beliefs measured by the NEP Scale group, predictive validity.
may explain why it has become the most widely used measure of Second, high corrected item-total correlations (averaging .46
EA since its publication in 1978 (Dunlap & Jones, 2002, 2003; and .39 for Sample 1 and 2, respectively) and high Cronbach’s
Stern, Dietz, & Guagnano, 1995).1 The theoretical background for alphas (.81 and .76 for Sample 1 and 2, respectively) were found,
suporting the internal consistency of the scale. Third, the first
unrotated principal factor accounted for high percentages of the
1
The authors argue that this scale measures ‘‘a paradigm or worldview that variance (69.2% and 63.3% for Sample 1 and 2, respectively),
influences attitudes and beliefs toward more specific environmental issues’’ providing support for the unidimensionality of the scale. Fourth,
(Dunlap et al., 2000, p. 428). This view indicates that the NEP Scale is not a direct, the NEP Scale was positively correlated with measures of support
specific measure of EA. However, as the authors recognise the NEP Scale has been for funding environmental programs (r ¼ .47), support for envi-
treated ‘‘variously as measuring environmental attitudes, beliefs, values, and
worldview’’ (Dunlap et al., 2000, p. 428). Moreover, the NEP Scale is commonly
ronmental regulations (r ¼ .58), and personal pro-environmental
used and described as an EA measure in the environmental psychology literature behavior (r ¼ .24) among the general population sample. Finally,
(see, e.g., Schultz & Zelezny, 1999), so we follow this trend. construct validity was demonstrated in Sample 1 by showing that
L.J. Hawcroft, T.L. Milfont / Journal of Environmental Psychology 30 (2010) 143–158 145

Table 1
The New Environmental Paradigm (NEP) Scale.

Original NEP items (1978) Revised NEP items (2000)


1. We are approaching the limit of the number of people the earth can 1. We are approaching the limit of the number of people the earth can
support. support.
2. The balance of nature is very delicate and easily upset. 2. Humans have the right to modify the natural environment to suit their
3. Humans have the right to modify the natural environment to suit their needs.
needs. 3. When humans interfere with nature it often produces disastrous
consequences.
4. Mankind was created to rule over the rest of nature.
5. When humans interfere with nature it often produces disastrous 4. Human ingenuity will insure that we do NOT make the earth unlivable.
consequences. 5. Humans are severely abusing the environment.
6. Plants and animals exist primarily to be used by humans. 6. The earth has plenty of natural resources if we just learn how to develop
7. To maintain a healthy economy we will have to develop a ‘‘steady–state’’ them.
economy where industrial growth is controlled. 7. Plants and animals have as much right as humans to exist.
8. Humans must live in harmony with nature in order to survive.
8. The balance of nature is strong enough to cope with the impacts of modern
9. The earth is like a spaceship with only limited room and resources industrial nations.
10. Humans need not adapt to the natural environment because they can 9. Despite our special abilities humans are still subject to the laws of nature.
remake it to suit their needs. 10. The so–called ‘‘ecological crisis’’ facing humankind has been greatly
11. There are limits to growth beyond which our industrialized society cannot exaggerated.
expand. 11. The earth is like a spaceship with very limited room and resources.
12. Mankind is severely abusing the environment. 12. Humans were meant to rule over the rest of nature.
13. The balance of nature is very delicate and easily upset.
14. Humans will eventually learn enough about how nature works to be able
to control it.
15. If things continue on their present course, we will soon experience a major
ecological catastrophe.

Note. Original NEP Scale: Agreement with items 3, 4, 6, and 10 indicate anti–NEP responses. Hypothesized facets: balance of nature (items 2, 5, 8, 12), limits to growth (items 1,
7, 9, 11), and human dominance over nature (items 3, 4, 6, 10, all anti–NEP items). Revised NEP Scale: Agreement with the eight odd–numbered items and disagreement with
the seven even–numbered items indicate pro–NEP responses. Hypothesized facets: the reality of limits to growth (items 1, 6, 11), antianthropocentrism (items 2, 7, 12), the
fragility of nature’s balance (items 3, 8, 13), rejection of exemptionalism (items 4, 9, 14), and the possibility of an eco-crisis (items 5, 10, 15).

the NEP Scale was negatively related to age and positively related to concept of the fragile balance of nature is more dramatically
education and liberalism. This provides support for the validity of expressed in item 5: ‘‘When humans interfere with nature it often
the NEP Scale as many prior and subsequent studies have also produces disastrous consequences’’. Although not much is known
found that younger, better educated individuals with more liberal about the validity and reliability of the 6-item NEP Scale, it has
ideological orientations tend to have higher EA than their coun- been used by many other researchers since its publication (e.g.,
terparts (for reviews, see Dietz, Stern, & Guagnano, 1998; Fransson Knight, 2008).
& Gärling, 1999).
2.2. The revised NEP Scale
2.1.1. The shortened 6-item NEP Scale
In the late 1970s Dunlap developed an abbreviated 6-item In 2000, Dunlap and colleagues reviewed several empirical
version of the original NEP Scale for use in a study by the Conti- studies published since they first developed the NEP Scale that
nental Group (Dunlap et al., 2000). Although an article was never provided support for the content, construct, predictive, and
published describing the psychometrics of this shorter scale (R. E. known–group validity of the scale. Despite this review they
Dunlap, personal communication, June 21, 2008), the items were decided to revise the scale because the original NEP Scale was
used in a series of studies by John Pierce and colleagues (for unbalanced (as two of the three facets contained only pro–trait
a review, see Pierce, Steger, Steel, & Lovrich, 1992). The items were: and one facet contained only con–trait items), and some of its
(1) The balance of nature is very delicate and easily upset by human items were worded in a sexist fashion (e.g., mankind was used
activities, (2) The earth is like a spaceship with only limited room instead of humans). The NEP was therefore revised in an effort to
and resources, (3) Plants and animals do not exist primarily for make it more psychometrically sound and avoid outdated, sexist
human use, (4), Modifying the environment for human use seldom terminology.
causes serious problems, (5) There are no limits to growth for While revising the original scale, Dunlap et al. (2000) also
nations like the United States and Canada, and (6) Mankind was decided to broaden the content of the measure by expanding the
created to rule over the rest of nature. hypothesized facets of an ecological worldview from three to five
These 6 items provided a balanced measure of each of the facets. These five facets are: (1) the reality of limits to growth, (2)
three facets identified in the original NEP Scale: balance of nature antianthropocentrism, (3) the fragility of nature’s balance, (4)
(items 1 and 4), limits to growth (items 2 and 5), and human’s rejection of exemptionalism, and (5) the possibility of an eco-crisis.
right to rule (items 3 and 6). As can be seen from Table 1, items 1, The revised NEP Scale consists of 15 items (8 pro–trait and 7 con–
2, and 6 of this 6-item NEP Scale are virtually identical to items 2, trait) with three items tapping each of these five hypothesized
9, and 4 from the original NEP Scale. Items 3 and 5 of the 6-item facets (see Table 1). These items include 6 from the original NEP
NEP are also very similar to items 6 and 5 from the original scale, Scale, 4 of which were slightly modified.
although their direction is reversed to balance the shortened To validate the revised scale, Dunlap et al. (2000) surveyed
scale. The biggest difference in wording is probably that used in a representative sample of 676 Washington State residents. They
item 4 of the shortened scale. In the original NEP Scale the obtained results very similar to those of the original study. First,
146 L.J. Hawcroft, T.L. Milfont / Journal of Environmental Psychology 30 (2010) 143–158

results again indicated skewed responses (toward pro–NEP) across form their own scaledversions of the scale containing 5, 7, 8 or 10
all 15 items. Second, supporting the internal consistency of the items can be found, as noted below. Researchers normally do this
scale corrected item-total correlations ranged from .33 (item 9) to by first performing a factor analysis of the items and then selecting
.62 (item 10) averaging .46, and Cronbach’s alpha was .83. Third, the ones that load onto one factor. Moreover, many researchers
providing support for the unidimensionality of the scale the first have often used items from the NEP Scale but altered the item
unrotated principal factor accounted for 31.3% of the variance, and wording substantially to fit their own interests (e.g., Yencken, Fein,
all items loaded heavily (from .40 to .73) on this factor. Fourth, the & Sykes, 2000). This variation in the use of the NEP Scale raises the
predictive and construct validity of the revised NEP Scale was interesting question of whether the results of such studies can be
examined by its correlation with several variables. The NEP Scale meaningfully compared or used to support a theoretical position
correlated positively with measures of perceived seriousness of (cf. Milfont, 2007a).
world ecological problems (r ¼ .61), support for proenvironment The last point refers to the title of the scale. In their revision of
policies (r ¼ .57), perceived seriousness of state and community air the scale in 2000, Dunlap et al. changed the scale title from the
and water pollution (r ¼ .31), and personal pro-environmental ‘‘New Environmental Paradigm’’ to the ‘‘New Ecological Paradigm
behavior (r ¼ .31). Moreover, the NEP Scale was also correlated in Scale’’. This new title has been adopted by some researchers (e.g.,
the expected direction, albeit modestly, with several socio- Clark, Kotchen, & Moore, 2003), but many others still use the
demographic variables, such as age (r ¼ .11), income (r ¼ .10), original title (e.g., Schultz, Shriver, Tabanico, & Khazian, 2004;
education (r ¼ .10), and liberalism (r ¼ .32). Verplanken, Walker, Davis, & Jurasek, 2008). Because the scale is
still most commonly known as the New Environmental Paradigm
2.3. The original and revised NEP Scales: points to consider Scale, this original title is used here to avoid confusion.

The reliability and validity of both the original and revised NEP
3. The present study
Scales as a general measure of EA are well–established. The scales
have been shown to discriminate between environmentalists and
It has now been 30 years since the NEP Scale was first pub-
non–environmentalists in many cultures (e.g., Johnson, Bowker, &
lished, and the scale has been widely used ever since. As an
Cordell, 2004; Olli, Grendstad, & Wollebaek, 2001; Pierce, Lovrich,
indication, a search on the ISI Web of Science database in 2008
Tsurutani, & Abe, 1987). The scales also correlate highly with other
showed that the 1978 paper had been cited 379 times, and the
measures of general EA (Dunlap et al., 2000), and have been shown
2000 paper had already been cited 135 times. However, no
to predict pro-environmental behavior in many studies (e.g., Kor-
previous study has attempted to provide a systematic review of
tenkamp & Moore, 2006; Olli et al., 2001). However, before moving
studies using the NEP Scale. Dunlap et al. (2000) provided a review
on to discuss the current meta-analysis, there are three points
of selected findings supporting the validity of the NEP Scale, but
relating to how the NEP Scales are currently used that are worth
because their article focused on the development of a new scale
noting.
this review was not extensive nor quantitative. The drawback of
The first point refers to the dimensionality of the scale. As
research reviews when presented and described only qualitatively
discussed above, the items of the original NEP Scale tap three
is that the discussion of the conflicting findings is often confusing
facets while the items of the revised scale tap five facets. Despite
and inconclusive. For this reason, a meta-analysis is often seen as
presenting these facets, both Dunlap and Van Liere (1978) and
a better alternative because it is a methodology that systematically
Dunlap et al. (2000) actually summed all NEP items to give a single
summarizes a body of quantitative research (Rosenthal & DiMat-
score instead of using these facets as sub-scales. Not surprisingly,
teo, 2001).
therefore, the dimensionality of both the original and revised NEP
The goal of the present study was to use a meta-analysis to
Scales has occasioned disagreement. Many studies have examined
provide an extensive quantitative review of studies that have used
the dimensionality of the original NEP Scale (e.g., Bechtel, Corral-
the NEP Scale. The main objective was to investigate the extent to
Verdugo, Asai, & Riesle, 2006; Bechtel, Corral-Verdugo, & Pinheiro,
which variation in the way that the NEP Scale was used would
1999; Bostrom, Barke, Turaga, & O’Connor, 2006; Corral-Verdugo &
affect the overall NEP score obtained across studies. To represent
Armendáriz, 2000; Geller & Lasley, 1985; Ji, 2004; Noe & Snow,
the diversity of research that has been conducted using the NEP, an
1990; Nooney, Woodrum, Hoban, & Clifford, 2003; Rideout,
effort was made to include studies that have used the scale in
Hushen, McGinty, Perkins, & Tate, 2005), and a few studies have
a wide range of cultural settings. This meant that it was more
also examined the dimensionality of the revised NEP Scale (Dunlap
important to include studies from a wide range of countries than to
et al., 2000; Floyd & Noe, 1996; Hunter & Rinner, 2004; Milfont &
include every single study from a particular country. Of particular
Duckitt, 2004a; Thapa, 2001). These studies rely on factor analytic
importance was to test whether variations in use of the NEP Scale
results to derive factors underlying the NEP items. Sometimes
are widespread, and, if so, whether these variations have a signifi-
distinct factors emerge and relevant items are combined to form
cant effect on samples’ scores on the NEP. Finding significant rela-
sub-scales of the NEP Scale (e.g., Thapa, 2001). However, a large
tionships between mean scores and variations in scale use (after
number of recent studies, especially those using the revised NEP
controlling for other study characteristics) would suggest that the
Scale, typically sum all items into a single measure of EA. They
variations in scale use have a negative effect on the comparability of
treat the items as measuring one construct even if unidimen-
results of studies using the NEP Scale.
sionality is not found (e.g., Dunlap et al., 2000), and often report
the alpha without examining dimensionality (e.g., Schultz, 2001).
The second point refers to other forms of the scale. A NEP Scale 4. Method
for Children, which is a 10-item modified version of the revised NEP
Scale, has also been developed and validated for use with children 4.1. Procedure
aged 10–12 years (Manoli, Johnson, & Dunlap, 2007). This scale has
not yet been widely used and so it will not be described further. The procedure consisted of three steps: (1) a search for studies
Apart from the four main alternative forms of the NEP Scale (i.e., using any version of the NEP Scale, (2) selection of studies that met
original, shortened, revised, and children versions), many criteria for inclusion in the meta-analysis, and (3) coding of relevant
researchers have also selected certain items and combined them to study features.
L.J. Hawcroft, T.L. Milfont / Journal of Environmental Psychology 30 (2010) 143–158 147

4.1.1. Search procedure formulas recommended by Aiken (1987). The recalculated, stan-
The aim of the search was to locate studies that had used the dardized scores ranged between 0 and 1 and have the advantage of
NEP Scale from as wide a range of countries as possible. A search allowing direct comparisons of scores across alternative response
was carried out for relevant articles in PsycINFO, Current formats. Following standard meta-analytic procedures, an inverse
Contents and Scopus covering the period 1978 to July, 2007. variance weight was then calculated for each of these standardized
Keywords used were New Environmental Paradigm, New Ecological scores. Weights were calculated for each sample using the formula:
Paradigm and NEP. A secondary search was performed using the w ¼ n/sd2, where w ¼ the inverse variance weight, n ¼ sample size,
ISI Web of Science to locate any article that had cited either the and sd ¼ standard deviation (Lipsey & Wilson, 2001, p. 72).
original (Dunlap & Van Liere, 1978) or revised NEP Scale (Dunlap Weighting the scores by the inverse variance accounts for the
et al., 2000). Reference lists from previous articles on EA were different sample sizes on which the scores were based. This
also consulted, particularly articles including developing coun- procedure thus controls for possible biases in the NEP scores as
tries (e.g., Schultz et al., 2005; Watson & Halse, 2005). In areas a result of sample size.
where there was a lack of published studies on the topic (e.g., the
Middle East and Africa), researchers living in those areas were 4.2.2. Measuring the effect of study characteristics on NEP scores
asked if they knew of any other relevant studies. Finally, the Weighted least squares regressions were performed to measure
reference lists of all studies located during the search were the effect of study characteristics on NEP scores. Specifically, the
reviewed. data was analysed with inverse variance weighted regressions
using a random-effects approach (via maximum likelihood esti-
4.1.2. Study selection mate). Inverse variance weighted regressions are recommended in
The search located more than 300 articles. The abstract or meta-analyses; the use of a random-effects model is further rec-
method section of each article was then examined to determine ommended when cases are highly heterogeneous and allows
whether it met the following criteria for inclusion in the analysis: findings to be generalised beyond the cases included in the meta-
(1) the study had used at least five items from either the original, analyses (Hedges & Vevea, 1998; Lipsey & Wilson, 2001). Random-
shortened or revised NEP, (2) the sample was based on an adult effects models take into account random variability in effect sizes
population (a mean age of at least 16 years), and (3) the mean derived from both subject–level sampling error (i.e., subjects in
score of the study sample on the NEP Scale was either included in each NEP study sampled from a population of subjects) and study–
the article or was available on request. The first criterion was level sampling error (i.e., NEP studies sampled from a population of
used to guarantee that a realistic number of similar items were studies). This means that the relationships between NEP scores and
used across studies, providing a reasonably reliable measure of methodological features reported in this meta-analysis can be
the NEP. Given that the original and revised NEP Scales were generalised to other studies.
designed for use with adults, the second criterion was used to Study characteristics were entered as independent variables and
guarantee that studies inappropriately using the scales with standardized NEP scores as the dependent variable. The study
children were avoided. Where different studies had used the characteristics included were internal reliability, mean age of
same data set (e.g., Schultz & Zelezny, 1999; Schultz, Zelezny, & samples, gender (percentage of males in the sample), year of
Dalrymple, 2000), scores were only included in the analyses once. publication, response format, sample type and scale length. Sample
In total 69 studies including a total of 139 samples met these type and scale length were dummy coded as a set of dichotomous
criteria and were included in the analyses. Summary statistics for variables using the nationally or regionally representative samples
all studies included in the analysis are provided in Appendix A, and 15-item NEP Scale version as baseline. For example, a dummy
and references for all studies included in the meta-analysis are variable was labelled ‘students’ which comprised a comparison of
provided in Appendix B. the NEP scores between representative and student samples
(0 ¼ representative sample, 1 ¼ student sample). Another variable
4.1.3. Coding of study features was labelled ‘12-item NEP’ which comprised a comparison of the
For each sample the following information was recorded: (a) NEP scores between the revised and original versions of the NEP
general study characteristics (year the study was published, Scale (0 ¼ 15-item NEP Scale, 1 ¼12-item NEP Scale).
authors, sample size); (b) culture of sample (defined as either the
ethnic group or country of residence); (c) sample composition 5. Results
(mean age, percentage of males, sample type); (d) methodology of
study, including length of the scale used (i.e., number of items), 5.1. Variations observed in the studies
response format used (i.e., number of points on the Likert scale),
and internal reliability of the NEP Scale; and (e) descriptive statis- 5.1.1. Study characteristics
tics (mean and standard deviation of NEP scores). The analysis includes 139 samples and a total of 58,279 partic-
ipants (for two samples information on sample size was not
4.2. Data analyses provided). Although a few samples in this analysis were collected
and/or published in the late 1970s and early 1980s, the vast
Data analyses consisted of two parts. Analyses were performed majority of samples (86.3%) were published in the last 12 years,
to (1) standardise NEP scores across studies, and (2) to test the since 1997.
effect of study characteristics on NEP scores.
5.1.2. Sample composition
4.2.1. Standardizing NEP scores The samples in the analysis vary greatly in terms of size, age,
The studies included in the meta-analysis varied greatly with gender composition, and sample type. Descriptive statistics relating
regard to the response format offered to participants and also the to sample composition are provided in Table 2.
sample size. To meaningfully compare such studies it was necessary
to standardise the NEP scores. To achieve this, means and standard 5.1.3. Study methodology
deviations of each samples’ scores on the NEP were first recalcu- As shown in Table 2, there is substantial variation in the meth-
lated to compensate for differences in response format using the odology of studies in the analysis: 7 different lengths of scale were
148 L.J. Hawcroft, T.L. Milfont / Journal of Environmental Psychology 30 (2010) 143–158

Table 2 Table 3
Descriptive statistics of samples. Number of samples and subjects by country and region of the world.

Variable % Min Max M SD N Countries and world regions Samples Subjects % of total
Sample size (n) 27 3686 425.39 589.86 137 Asia 12 6143 10.54
China 4a 2335
Age (mean) 18 60.41 –a – 80
Hong Kong 2 769
Gender (% male) 12.90 97.00 46.27 16.74 86 India 2 357
Indonesia 1 225
Sample type 139 Japan 3 1708
Representative 28.78 Maldives 1 199
Student 31.66 South Korea 1 550
Other 18.70
Environmentalists 9.35 North America 62 26,546 45.55
White-collar 8.63 Canada 10 4851
Blue-collar 2.88 USA 52 21,695

Length of scale 139 Latin America 16 2662 4.57


15 items 41.73 Argentina 1 54
6 items 25.18 Brazil 4 680
12 items 16.55 Colombia 1 149
5, 7, 8 or 10 items 16.54 Costa Rica 1 213
Dominican Republic 2 388
Response format 139 Ecuador 1 201
5-point scale 83.45 El Salvador 1 194
4-point scale 7.91 Mexico 1 65
7-point scale 6.48 Panama 1 100
9- or 10-point scale 2.16 Paraguay 1 200
Internal reliability (a) .35 .87 .68 .11 78 Peru 1 224
Venezuela 1 194
a
It was not possible to calculate the overall mean and standard deviation because
many studies did not report the necessary information. Middle East 2 531 .91
Turkey 2 531

Western Europe 24 12,769 21.91


used and 5 different response formats employed. With regards to Germany 2 588
scale length there is no standard version of the scale or number of Netherlands 3 417
items that the majority of researchers used. Although the revised Norway 4 6466
Portugal 1 396
15-item NEP Scale is the most commonly used in this analysis (for
Spain 4 454
41.73% of the samples), a substantial percentage of studies also used Sweden 5 2906
the original 12-item NEP Scale (16.55%) or the 6-item shortened United Kingdom 5 1542
version (25.18%). Studies that used 5, 6, 7, 8, or 10 items from the
Eastern Europe 7 2327 3.99
NEP Scale (representing 16.54% of the total) did not always use the Bulgaria 3 1287
same itemsde.g., although Gooch (1995), Rauwald and Moore Czech Republic 1 113
(2002), and Pierce et al. (1987) all used 6 items, the specific NEP Estonia 1 400
Latvia 1 407
items used vary in each study. This variation was increased by the
Russia 1 120
fact that when choosing the items, some researchers drew on the
original 12-item NEP Scale and others from the 6-item or the 15- Oceania 14 7301 12.53
New Zealand 6 5221
item NEP Scale. For example, although both Blake, Guppy, and
Australia 8 2080
Urmetzer (1997) and Hall and Moran (2006) used 10 items from the
a
Data on number of subjects was not available for two of the samples from China.
scale, the items were quite different in the two studies because they
were drawn from the original 12-item and the revised 15-item
versions of the NEP Scale, respectively.
With regards to response format, the methodology of the
studies in the analysis is much more consistent. All studies used 5.2. Effect of variation in study characteristic on NEP scores
a Likert scale and, although five different numbers of points on the
Likert scale were used, the vast majority of studies (83.45%) used As a first step in the present analysis, correlations between
a 5–point response format. However, there is a wide variability in mean NEP scores and sample characteristics were performed.
the internal reliabilities reported. While the average internal reli- Bivariate correlations might be of interest to some readers and
ability is acceptable for research purposes (average Cronbach’s are reported in Table 4. However, correlation coefficients do not
alpha ¼ .68), 39 of the 78 studies that report internal reliability account for indirect effects between the variables examined and
have a lower alpha level (i.e., .69) than is considered acceptable are therefore less reliable. Take the relationship between NEP
(Nunnally, 1978). scores and year of publication for example. The negative corre-
lation between NEP scores and the year in which a study was
5.1.4. Subjects and samples from each country published suggests that scores from earlier studies are higher
Twelve of the 139 samples in the analysis are made up of ethnic than those from more recent ones. However, previous studies
groups within a country (e.g., Asian Americans). For simplicity of have indicated that pro–EA have increased over the last 20 years
analysis these minorities were coded to the country in which they (e.g., Dunlap et al., 2000; Inglehart & Baker, 2000). It thus seems
lived rather than their country of origin. Using this system the study unlikely that this relationship is actually due to a decrease in EA
includes samples from 36 countries. The number of samples and over time. Multivariate analyses using weighted regression
subjects from each country is presented in Table 3. The majority of analyses were then performed as a more robust way to examine
subjects and samples in the analysis come from North America and the relationship between mean NEP scores and study
Western Europe. characteristics.
L.J. Hawcroft, T.L. Milfont / Journal of Environmental Psychology 30 (2010) 143–158 149

Table 4
Correlation coefficients of mean NEP scores and study characteristics.

1 2 3 4 5 6 7 8 9 10 11 12 13 14
1. NEP (M) (139)
2. Internal reliability .05 (78)
3. Age .07 .23 (80)
4. Gender .02 .14 .64*** (86)
5. Year .21* .23* .25* .05 (139)
6. Response format .03 .24* .24* .22* .09 (139)
7. Students .02 .00 .80*** .47*** .11 .27** (139)
8. Blue-collar .44*** .09 .08 .19 .04 .03 .12 (139)
9. White-collar .19* .06 .42*** .51*** .06 .05 .21* .05 (139)
10. Environmentalists .32*** .02 .25* .20 .16 .08 .22* .06 .10 (139)
11. Other samples .11 .07 .24* .09 .22** .06 .33** .08 .15 .15 (139)
12. 12-item NEP .22* .07 .04 .04 .06 .12 .10 .39*** .07 .14 .18* (139)
13. 6-item NEP .45*** .32** .20 .01 .45*** .12 .27** .10 .13 .17* .09 .25** (139)
14. Other length .18* .00 .50*** .34** .19* .01 .24** .08 .12 .17* .18* .21* .26** (139)

Note. Number of samples for each variable is indicated in brackets in the diagonal.
*p < .05. **p < .01. ***p < .001.

Weighted, random-effects regression analyses using SPSS variance in NEP scores. This suggests that, as hypothesized, varia-
macros (Lipsey & Wilson, 2001, Appendix D) were performed. In tion in methodology may affect the accuracy of NEP scores. Most of
a first model, all variables were included as predictors (see Table the variables related to sample type and scale length emerge as
5). However, due to the small number of cases available for some significant predictors, indicating that variation in these variables do
variables, this first model (Model 1) includes only forty samples, affect NEP scores.
which make the findings unstable and less generalizable. To Environmentalist and white-collar samples score significantly
increase the number of samples included in the analyses and thus higher on the NEP Scale than do representative samples, while
increase statistical power, a second regression model (Model 2) blue-collar samples score significantly lower. Nevertheless, it is
was performed excluding internal reliability, mean age and gender important to note that the vast majority (79.14%) of samples in the
from the model. As expected, with fewer variables there is selected studies consist of representative, students or other sample
a reduction in the variance explained by Model 2 (R2 ¼ .48) types, and for these studies the sample type has no significant effect
compared to Model 1 (R2 ¼ .64). More importantly, however, the on NEP scores. Therefore, variation in sample type probably does
number of samples increased to 137 and the residual for Model 2 not have much of an effect on the comparability of NEP scores for
is nonsignificant (Qresidual ¼ 137.63, p ¼ .23)dwhile reaching most studies.
significance for Model 1 (Qresidual ¼ 46.54, p ¼ .06)dwhich indi- In comparison, variations in scale length have a greater effect on
cates that the unexplained variability in the NEP score in Model 2 the comparability of the scores. NEP scores are significantly higher
is no greater than would be expected from sampling error. This for samples responding to the 6-item NEP (completed by 25.18% of
model thus has higher statistical power and the predictors explain samples) than for the baseline 15-item NEP (completed by 41.73% of
more of the variability in the NEP scores, providing more robust samples). On the other hand, NEP scores are significantly lower for
findings. the other (5, 7, 8, and 10 item) scale lengths (completed by 16.54% of
Results from Model 2 are reported in Table 5. As expected, some samples). Therefore, variations in scale length do have a prevalent
of the variables that had significant correlations with NEP scores effect on the comparability of the NEP scores for most studies.
are no longer significant when other factors are taken into account. However, given that some studies using shorter versions of the
Nevertheless, variation in study characteristics explains 48% of the scale did not use the same items, it is important to note that the

Table 5
Summary of inverse variance weighted regression model predicting the effects of study characteristics on mean NEP score (random effects).

Predictor variable N Model 1 (R2 ¼ .64***, Qmodel ¼ 81.42***, k ¼ 47) Model 2 (R2 ¼ .48***, Qmodel ¼ 127.60***, k ¼ 137)

B SE b Z B SE b Z
Internal reliability 78 .02 .05 .06 .46 – – – –
Age 80 .00 .00 .21 .90 – – – –
Gender (being male) 86 .00 .00 .21 1.56 – – – –
Year 139 .00 .00 .28 2.30* .000 .00 .01 .13
Response format 139 .01 .01 .20 1.88 .00 .00 .03 .45
Students (versus representative) 139 .02 .03 .18 .50 .02 .01 .13 1.37
Blue-collar (versus representative) 139 .21 .04 .61 5.63*** .15 .03 .39 5.52***
White-collar (versus representative) 139 .07 .04 .23 1.77 .04 .01 .20 2.87**
Environmentalists (versus representative) 139 .03 .03 .19 1.05 .07 .01 .31 4.61***
Other samples (versus representative) 139 .00 .02 .02 .08 .01 .01 .04 .44
12-item NEP (versus 15 item NEP) 139 .02 .02 .18 1.23 .01 .01 .05 .68
6-item NEP (versus 15 item NEP) 139 .09 .03 .39 3.43*** .05 .01 .34 3.83***
Other length (versus 15 item NEP) 139 .06 .03 .36 1.71 .03 .01 .18 2.19*

Note. Model 1. All variables included as predictors. Model 2. Excluding internal reliability, age and gender from the model. N ¼ number of samples for each variable;
R2 ¼ proportion of variance accounted for; Qmodel ¼ test of whether the regression model explains a significant amount of variability across effect sizes. k ¼ number of samples
included in the analysis; B ¼ raw regression coefficient; SE ¼ standard error for the regression coefficient; b ¼ standardized regression coefficient; Z ¼ z test for the regression
coefficient.
*p < .05. **p < .01. ***p < .001.
150 L.J. Hawcroft, T.L. Milfont / Journal of Environmental Psychology 30 (2010) 143–158

apparent effect of scale length on NEP scores may also be due to the 6.1.2. Sample type
specific items chosen as opposed to length per se. The results with regard to the effect of sample type on NEP
scores are unsurprising. As found in several primary studies (e.g.,
5.3. Variations in content of the 6-item scales Dunlap et al., 2000; Widegren, 1998), our meta-analysis reveals
that environmentalists score higher on the NEP Scale than
Examination of the data used in this analysis suggested that representative samples, which provides known–group, predictive
current usage of the 6-item version of the NEP Scale might (a) validity for the scale. The finding that blue-collar samples score
vary considerably across studies and (b) have a significant effect lower and white-collar samples score higher is also similar to the
on samples’ NEP scores. However, due to the limited number of results of several primary studies indicating that education
samples (n ¼ 33) that had used 6 items, it was not possible to positively correlates with NEP scores (e.g., Dunlap et al., 2000;
carry out a statistical analysis of what effect this variation in item Gooch, 1995).
content might have had on NEP scores. A content analysis of the Nevertheless, these results suggest that future researchers
item variation was performed instead. The content analysis should be aware of the fact that samples from different socio–
showed that of the 33 studies, 13 (39.40%) used the same 6 items economic backgrounds may respond differently to the NEP Scale.
originally developed by Dunlap for the Continental Group in It may even be necessary to develop measures of EA that more
1982, and 10 (30.3%) used slightly modified items to remove adequately tap the concerns of blue-collar samples from low
sexist wording. Of the remaining 10 studies, 5 studies selected 6 income areas (Mohai & Bryant, 1998). However, given that no
items, seemingly at random, from the original 12-item NEP Scale. difference was observed between the NEP scores of the repre-
The remaining 5 studies use 6 items that are very similar to the sentative, student and other samples, it seems that this may only
items developed for the Continental Group, except that the be necessary when working with samples drawn from very low or
direction of one item has been changed. The effect of this high socio–economic groups. It may also be necessary to develop
(seemingly minor change) is that the three sub-scales within the a more appropriate measure of EA for environmentalists given
6-item NEP Scale are no longer balanced. Overall, however, this their tendency to strongly agree with virtually every NEP item,
content analysis suggests that the majority of studies were creating a kind of ‘ceiling effect’ (cf. Widegren, 1998). It would be
consistent in using the original items of the shortened 6-item NEP interesting to see if scales developed specifically for use with
Scale. environmentalists show higher correlation with pro-environ-
mental behavior than does the NEP Scale. Given the benefits of
having results that can be easily compared across studies, it is
6. Discussion recommended that while researchers may want to develop new
scales for use with white/blue-collar/environmentalist groups,
This article used a meta-analytical approach to provide a quan- they should also administer the NEP Scale or another standard EA
titative review of 30 years of research using the New Environmental measure to such samples.
Paradigm (NEP) Scale, including 69 studies, 139 samples, and
58,279 participants. Because the NEP Scale is currently the most 6.1.3. Variation in scale length and item content
widely used measure to assess people’s environmental attitudes Samples tend to score higher on the 6-item version of the NEP
(EA), it seemed useful to investigate whether variations in the Scale and lower on versions using other numbers of items than on
characteristics of studies using the scale would affect samples’ the revised 15-item NEP Scale. This suggests that variations in the
average scores. Our quantitative review indicates problematic ways way the NEP Scale is used may affect both the accuracy and
in which the NEP Scale is being used. It also provides evidence that comparability of research using this scale to measure EA. It is of
variation in sample type and scale length do affect NEP scores. The course possible that despite the difference in mean scores observed
implications of each of these findings will be discussed briefly and responses to the 12–, 6–, and 15-item versions of the NEP Scale
then some recommendations for future research using the NEP might still be highly correlated. In this case, although their mean
Scale will be made. scores might not be directly numerically comparable, studies using
different versions of the NEP Scale could be assumed to be
6.1. Problems associated with the use of the NEP Scale measuring the same construct, and also to correlate similarly to
other variables such as pro-environmental behavior.
6.1.1. Lack of necessary information However, given the variation observed in item content among
The results show that there is a marked tendency for researchers studies using different versions of the scale, this assumption of
to provide less information, both about the NEP Scale and also equivalence of the versions of the NEP Scale cannot be supported.
about the characteristics of their sample, than is recommended by For example, our content analysis of studies that used 6 items from
the American Psychological Association. For example, only 74 of the the NEP Scale indicates that such studies usually did not include
139 samples included in the analyses reported the internal items referring to the possibility of an ‘‘ecological catastrophe’’ or
consistency of their samples’ responses to the NEP Scale. This was ‘‘crisis’’ (phrases that occur in the original 12-item and revised 15-
particularly worrisome since many samples in the meta-analysis item versions). This may well explain why samples tend to score
had lower internal reliability than is generally considered accept- higher on the 6-item NEP Scale than the other versions. Dunlap and
able. Researchers also often did not report the standard deviation of colleagues (2000, Table 1) break–down of responses suggests that
their NEP scores: only 80 of the 139 samples included this crucial samples often agree less with ‘‘catastrophe’’ items than many other
piece of information. Finally, about one third of studies failed to items in the NEP Scale. It seems intuitively logical that more people
give basic information about their samples’ composition, such as are concerned about humans’ treatment of the environment than
the average age or gender composition. While the limited space concerned that humans’ treatment of the environment will lead to
available in journals may make it difficult for researchers to include an eco-crisis. It is also conceivable that including such items in
such details, it is important they do so. In particular, it is essential to a scale might create a context effect (Schwarz, 1999) that would
provide the standard deviation and internal reliability of scores on cause participants who reject the possibility of an eco-crisis to
a scale so that results can be properly interpreted by other respond in a less pro-environmentally fashion to all items in the
researchers. scale.
L.J. Hawcroft, T.L. Milfont / Journal of Environmental Psychology 30 (2010) 143–158 151

6.2. Recommendations for future studies using the NEP Scale could be created by selecting one pro– and one anti–NEP item
from each of the five hypothesized facets.2 However, selecting
Based on these results some recommendations for how the items from the NEP Scale like this should be avoided as the
NEP Scale can be used more effectively in future research present results do indicate that the number of items selected
are provided. The first recommendation refers to the information impacts the overall NEP score. Researchers should be aware that
provided in published articles. Researchers should be careful there is a trade–off between having brief but less accurate
to provide all the information needed for their research to be measures, or having long but accurate ones. It is therefore strongly
correctly interpreted: explicitly indicating which items they used recommended that researchers use all 15 items of the revised NEP
from the NEP Scale, and the mean, standard deviation, and Scale whenever possible.
internal reliability. Researchers should also provide enough
information on their samples’ characteristics (e.g., sample type,
6.3. Concluding remarks
percentage of female/male, age range and mean, etc.) to allow
their results to be viewed in context.
One possible limitation of this meta-analysis is that it includes
Reporting more information could also increase our under-
studies from a very wide range of countries. These studies may not
standing of the correlation between EA and other variables. For
be representative of the bulk of NEP research as the vast majority of
example, if all researchers were to provide information on the
studies using the NEP Scale have probably been conducted in North
correlation between NEP scores and the gender, age, educational
America. However, the decision to include studies from different
level, or income of respondents this would provide a robust body of
countries can also be seen as a strength as it means that the analysis
information that could be used to test whether associations
represents the true diversity of research using the NEP Scale that is
between these characteristics and EA truly exist. Currently, it is
occurring worldwide. It is also worth noting that 62 of the 139
usually only possible to locate a handful of studies that have
samples included in this analysis do come from North America,
consistently used the same measure of EA and reported the rela-
thus, hopefully, making these results generalizable to most NEP
tionship between scores on the EA measure and some other char-
research.
acteristics of participants. For example, Zelezny, Chua, and Aldrich
Despite the fact that the NEP Scale has attracted widespread
(2000) could locate only six studies between 1988 and 1998 that
use and despite the fact that EA is a crucial construct in the field,
examined gender and EA using the NEP Scale.
there is a lack of empirical and theoretical integration in studies
The second recommendation refers to the response format
employing this scale as a measure of EA. As suggested, this
used. Theoretical and empirical advances in the field of attitude
problem is compounded by the varying ways in which the NEP
measurement have led to guidelines for the optimal development
Scale has been used. More consistent use of the 15–item version of
of direct self–report attitude measures using the Likert technique.
the NEP Scale could help improve this situation.
A recent research review by Krosnick, Judd, and Wittenbrink
Some recent studies have tried to systematize the study of EA
(2005) suggests that data quality improves when 7–point scales
by building on previous studies and provide a better con-
are used. Although the NEP Scale uses a 5–point scale format, our
ceptualisation of the content and structure of EA (e.g., Blaikie,
results did not indicate any effect of the response format on the
1992; Bogner & Wiseman, 1999; Milfont, 2007b; Milfont &
NEP score. For the sake of consistency, it is thus recommended that
Duckitt, 2004b, 2006). Moreover, empirical studies have been
researchers continue using the 5–point scale format of the original
attempting to develop EA measures based on recent theoretical
scale, as the majority (83.45% in this analysis) of researchers are
advances in the field (e.g., Bogner & Wiseman, 2006, 2007). For
already doing.
example, Milfont and Duckitt (in press) developed the Environ-
The third recommendation refers to the use of specific items
mental Attitudes Inventory (EAI) as a psychometrically sound,
from the NEP Scale. Our findings show that studies using either
multi–dimensional inventory to assess EA cross–culturally. The
the 12– or the 15-item versions of the scales are comparable.
EAI comprises a comprehensive list of twelve scales that captures
However, when researchers have used a smaller number of items
both the vertical and horizontal structure of EA, and has shown
from the NEP Scale, this seems to have a negative effect on the
acceptable reliability and validity in samples from Brazil, New
comparability and, possibly, the validity of their scores. To para-
Zealand, and South Africa (Milfont, Duckitt, & Wagner, in press-a,
phrase the title of the well–known article by Van Liere and Dunlap
in press-b).
(1981), it does make a difference how the NEP is measured. There
These theoretical and methodological advances in the area
are situations and settings where time needed for completion is
should bring integration to the field. However, until a gold–stan-
a major constraint in the use of the NEP Scale, and researchers
dard EA measure has been widely accepted, it is probably advisable
therefore select a sample of items. In this case, it is recommended
for researchers to continue using the NEP Scale as a standardized
that they use a balanced set of pro– and anti–NEP items. For
measure of EA. This study quantitatively reviewed the use (and
example, a 10-item short form version of the revised NEP Scale
abuse) of this scale over the last 30 years. As environmental issues
grow increasingly serious, we must ensure that our measures
provide reliable and accurate scores of EA. It is hoped that the
2
As an illustration, this proposed short version was tested with a published results of this research will be helpful to future researchers using
sample (see Milfont & Duckitt, 2004b) of 455 undergraduate students (319 females;
the NEP Scale.
136 males), with ages ranging from 17 to 48 years (M ¼ 20; S.D. ¼ 4.31). Two
balanced items with higher corrected mean-item total correlations reported by
Dunlap and colleagues (2000, Table 1) were selected for each facet (i.e., items 4, 6, 7,
Acknowledgements
8, 9, 10, 11, 12, 13, and 15). The corrected item-total correlations for the full NEP
scale (M ¼ 4.71; S.D. ¼ .67), ranged from .154 (item 6) to .435 (item 1) and averaged
.373, with Cronbach’s alpha of .768. The corrected item-total correlations for the This research was made possible by a Faculty of Arts Honours
proposed 10-item version of the NEP scale (M ¼ 4.71; S.D. ¼ .68), ranged from .125 scholarship from the University of Auckland to Lucy J. Hawcroft,
(item 6) to .502 (item 15) and averaged .339, with Cronbach’s alpha of .675. This and by scholarship BEX 2246/02–3 from the Ministry of Educa-
reduction in Cronbach’s alpha is expected because this coefficient on the number of
items. Overall, however, this short version provides reasonable internal consistency,
tion of Brazil (CAPES Foundation) to Taciano L. Milfont. The
and correlates (r ¼ .29, p < .001) as strongly as the full version (r ¼ .31, p < .001) with authors would like to express their appreciation to Ronald
a 8-item measure of ecological behavior. Fischer for his valuable help on meta-analysis literature and to
Appendix A

152
Summary statistics for all NEP studies and samples included in the meta-analysis.

Study Culture of sample Sample occupation Sample size Gender (% male) Age (m) Scale length Mean NEPa SD NEP a NEP
Amérigo and Gonzalez (2001) Spain Students 165 31 21 15 3.58 .42 .69
Andersson, Shivarajan, and Blau (2005) Indian White-collar 147 – – 5 3.98 .42 .71
Arcury and Christianson (1990) 1. USA Representativeb 441 48 – 5 2.85 .74 .63
2. USA Representative 653 51 – 5 2.95 .74 .62
Barr and Gilg (2006) United Kingdom Representative 1214 – – 6 3.98 .59 –
Berenguer, Corraliza, and Martin (2005) 1. Spain Other 90 – – 15 3.40 – –
2. Spain Other 95 – – 15 3.63 – –
Bjerke, Østdahl, Thrane, and Strumse (2006) Norway Representative 720 49 50 5 3.80 .71 .78
Blaikie (1992) 1. Australia Students 390 53 21 6 3.96 – –
2. Australia Representative 402 51 42 6 3.88 – –
Blake et al., (1997) Canada Representative 1954 – – 10 3.93 – –
Bostrom et al. (2006) 1. Bulgaria Representative 787 45 – 6 3.52 – .45
2. Bulgaria Other 250 50 38 6 4.03 – .45

L.J. Hawcroft, T.L. Milfont / Journal of Environmental Psychology 30 (2010) 143–158


3. Bulgaria Other 250 50 37 6 3.93 – .54
Casey and Scott (2006) Australia Other 292 34 36 15 3.69 .51 .84
Castro and Lima (2001) Portugal Other 396 40 30 15 3.72 .42 –
Chung and Poon (2001) 1. China Representative – – – 12 3.54 – –
2. China Other – – – 12 3.58 – –
3. China Other 204 – – 12 3.75 – –
Chung and Poon (2000) 1. China Other 2131 – – 12 3.60 – –
2. Hong Kong Other 540 – – 12 3.66 – –
Clark et al. (2003) 1. USA Representative 619 70 51 10 3.39 .69 .80c
2. USA Environmentalists 281 58 52 10 3.78 .73 .80c
Cooper, Poe, and Bateman (2004) United Kingdom Students 200 – – 15 3.88 – .72
Cottrell (2003) USA White-collar 226 97 50 12 3.90 – .82
Davey and Vertrees (1999) 1. Canada White-collar 40 85 57 12 4.00 – –
2. USA White-collar 27 85 52 12 3.50 – –
Deng, Walker, and Swinnerton (2006) 1. Anglo–Canadian Representative 160 55 – 15 3.81 .52 –
2. Chinese–Canadian Representative 178 57 – 15 3.67 .53 –
Dunlap et al. (2000) USA Representative 676 – – 15 3.75 – .83
Environment Waikato (2007) 1. New Zealand Representative 1873 – – 6 3.88 – –
2. New Zealand Representative 528 – – 6 3.56 – –
Eriksson, Garvill, and Nordlund (2006) Sweden Representative 922 48 49 15 3.67 .48 .77
Furman and Erdur (1995) 1. Turkey Representative 387 53 – 12 3.84 .44 .58
2. Turkey Students 144 49 – 12 3.83 .45 .62
Gooch (1995) 1. Estonia Representative 400 51 – 6 3.70 – .52
2. Latvia Representative 407 49 – 6 3.69 – .35
3. Sweden Representative 287 53 – 6 4.22 – .58
Hall and Moran (2006) United Kingdom Environmentalists 38 42 – 10 4.13 – –
Hodgkinson and Innes (2000) Australia Students 391 40 21 12 4.05 .63 .80
Holt (2005) USA Other 900 81 – 12 3.69 .73 –
Hunter and Rinner (2004) USA Representative 398 – – 15 3.81 .73 .87
Jager (2006) 1. Netherlands Environmentalists 197 76 47 7 3.70 – .58
2. Netherlands Other 108 – – 7 3.05 – .60
Johnson et al. (2004) 1. African–American Representative 248 – – 10 3.61 .96 .70c
2. Anglo–American Representative 2995 – – 10 3.68 .69 .70c
3. Asian–American Representative 57 – – 10 3.74 .94 .70c
4. Latinos born in USA Representative 169 – – 10 3.67 .86 .70c
5. Latinos born outside USA Representative 44 – – 10 3.31 1.00 .70c
Kaiser, Hubner, and Bogner (2005) Germany Students 468 17 23 15 4.08 .45 .78
Kim, Borges, and Chon (2006) Brazil Other 412 52 – 15 3.59 – –
Kortenkamp and Moore (2006) USA Students 112 29 19 15 3.57 .58 .83
Kotchen and Moore (2007) 1. USA Environmentalists 624 70 51 5 3.44 – .68
2. USA Environmentalists 677 53 60 5 3.46 – .71
Kotchen and Reiling (2000) 1. USA Representative 206 – – 15 3.65 – .83
2. USA Representative 200 – – 15 3.61 – .83
Appendix A (continued )

Study Culture of sample Sample occupation Sample size Gender (% male) Age (m) Scale length Mean NEPa SD NEP a NEP
Kurz, Donaghue, and Walker (2005) Australia Other 93 – – 15 3.62 .44 –
Lai, Brennan, Chan, and Tao (2003) Hong Kong Other 229 52 35 15 3.46 .37 .56
Leung and Rice (2002) 1. Anglo–Australian Other 203 37 35 12 3.81 .57 .71
2. Chinese–Australian Other 98 61 33 12 3.57 .41 .59
Liu and Sibley (2004) 1. New Zealand Students 1924 51 – 8 3.88 .74 .75
2. New Zealand Students 224 – – 6 3.80 .66 .75
Mainieri, Barnett, Valdero, Unipan, USA White-collar 201 42 – 6 3.86 .79 .70
and Oskamp (1997)
Marshall, Picou, and Bevc (2005) 1. USA Representative 118 48 40 8 3.86 – –
2. USA Representative 73 44 42 8 3.55 – –
Mayer and Frantz (2004) USA Other 60 52 31 15 3.94 .50 .75
McFarlane, Stumpf–Allen, Watson (2006) 1. Canada Other 484 – 51 15 3.71 .64 .83c
2. Canada Other 435 – 46 15 3.87 .60 .83c
3. Canada Representative 457 – 45 15 3.67 .60 .83c

L.J. Hawcroft, T.L. Milfont / Journal of Environmental Psychology 30 (2010) 143–158


Milfont, Duckitt, and Cameron (2006) New Zealand Students 455 30 20 15 3.51 .48 .77
Olli et al., (2001) 1. Norway Environmentalists 2000 – – 15 4.09 – .71
2. Norway Representative 3686 – – 15 3.69 – .71
Pahl, Harris, Todd, and Rutter (2005) 1. United Kingdom Students 45 – 21 15 3.31 – .86c
2. United Kingdom Environmentalists 45 – 21 15 4.22 – .86c
Palmgren, Morgan, De Bruin, and Keith (2004) USA Other 120 36 52 15 3.36 – .84
Pierce et al. (1987) 1. Japanese Representative 684 – – 6 4.05 – .69
2. Japanese White-collar 436 – – 6 4.25 – .62
3. Japanese Environmentalists 588 – – 6 4.08 – .67
4. USA Representative 524 – – 6 3.78 – .62
5. USA White-collar 139 – – 6 3.97 – .68
6. USA Environmentalists 235 – – 6 3.98 – .71
Rahelizatovo and Gillespie (2004) USA Other 124 – 51 15 3.22 .61 –
Rauwald and Moore (2002) 1. Dominican Republic Students 267 43 19 6 3.49 .52 –
2. USA Students 257 26 28 6 3.99 .54 –
Rideout et al. (2005) 1. USA Other 305 45 – 15 3.42 .58 .83
2. USA Other 659 35 – 15 3.60 .56 .82
Schultz et al. (2005) 1. Brazil Students 208 27 27 15 3.55 .36 .56
2. Czech Republic Students 113 34 24 15 3.80 .42 .74
3. Germany Students 120 25 26 15 4.02 .34 .64
4. India Students 210 34 20 15 3.50 .40 .60
5. New Zealand Students 217 33 25 15 3.74 .44 .72
6. Russia Students 120 84 18 15 3.64 .40 .68
Schultz, Unipan, and Gamba (2000) 1. Latino low acculturation, USA Blue-collar 63 32 28 12 3.58 – .74c
2. Latino medium acculturation, Blue-collar 62 32 28 12 2.87 – .74c
USA
3. Latino high acculturation, Blue-collar 28 32 28 12 2.57 – .74c
USA
4. USA Blue-collar 39 31 38 12 2.75 – .77
Schultz et al. (2004) USA Students 160 – – 15 3.57 .46 .78
(continued on next page)

153
Appendix A (continued )

154
Study Culture of sample Sample occupation Sample size Gender (% male) Age (m) Scale length Mean NEPa SD NEP a NEP
Schultz and Zelezny (1999), Schultz et al. 1. Argentina Students 54 41 24 15 3.83 .45 .65
(2000), and Zelezny et al. (2000) 2. Canada Students 96 28 23 15 4.11 .40 .74
3. Columbia Students 149 20 24 15 4.01 .40 .60
4. Costa Rica Students 213 36 26 15 4.08 .45 .67
5. Dominican Republic Students 121 44 21 15 3.75 .37 .47
6. El Salvador Students 194 36 26 15 3.69 .45 .55
7. Ecuador Students 201 47 28 15 3.93 .43 .69
8. Mexico Students 65 63 37 15 3.88 .40 .51
9. Panama Students 100 32 24 15 3.94 .36 .54
10. Peru Students 224 30 24 15 3.89 .41 .56
11. Paraguay Students 200 40 25 15 3.75 .38 .62
12. Spain Students 104 36 23 15 3.83 .45 .70
13. USA Students 245 29 21 15 3.67 .53 .81
14. Venezuela Students 194 41 25 15 3.91 .39 .63

L.J. Hawcroft, T.L. Milfont / Journal of Environmental Psychology 30 (2010) 143–158


Shafer (2006) USA Students 302 57 34 15 3.47 .77 .82
Shin (2001) South Korea Other 550 56 29 12 3.77 .10 .69
Slimak and Dietz (2006) 1. USA Representative 291 59 51 5 3.57 – –
2. USA White-collar 113 77 50 5 3.80 – –
3. USA White-collar 148 70 48 5 4.03 – –
4. USA White-collar 62 66 51 5 3.72 – –
Steel, List, Lach, and Shindler (2004) 1. USA White-collar 155 – 48 6 4.30 .49 .81c
2. USA White-collar 167 – 48 6 3.97 .64 .81c
3. USA Other 118 – 49 6 4.40 .63 .81c
4. USA Other 192 – 56 6 3.99 .95 .81c
Steel, Lovrich, Lach, and Formenko (2005) USA Representative 3204 – 51 6 3.77 .77 .79
Steel, Smith, Opsommer, Curiel, and USA Representative 1176 – 51 6 3.89 .75 .79
Warner-Steel (2005)
Steg, Dreijerink, and Abrahamse (2005) Netherlands Other 112 46 40 15 3.50 .40 .73
Steger, Pierce, Steel, and Lovrich (1989) 1. Canada Representative 600 – – 6 4.04 – –
2. Canada Environmentalists 447 – – 6 4.44 – –
3. USA Representative 476 – – 6 3.88 – –
4. USA Environmentalists 566 – – 6 4.29 – –
Tarrant and Cordell (1997) USA Representative 1220 39 43 12 3.69 50 .75
Thapa (1999) USA Students 540 55 – 15 3.48 – –
Vikan, Camino, Biaggio, and 1. Brazil Students 60 50 23 15 3.73 .42 .66
Nordvik (2007) 2. Brazil Students 60 50 23 15 3.76 .37 .53
3. Norway Students 120 50 24 15 3.60 .42 .72
Watson and Halse (2005) 1. Australia Students 211 13 22 12 3.99 .47 –
2. Indonesia Students 225 14 21 12 3.71 .40 –
3. Maldives Students 199 20 21 12 3.44 .40 –
Widegren (1998) 1. Sweden Representative 1008 – – 6 4.04 – –
2. Sweden Environmentalists 446 – – 6 4.47 – –
3. Sweden Environmentalists 243 – – 6 4.70 – –

Notes: – ¼ no data available. a mean NEP scores are standardized to a 5–point scale. b representative ¼ representative sample of nation or region. c Represents the average internal reliability of several samples from the same
study.
L.J. Hawcroft, T.L. Milfont / Journal of Environmental Psychology 30 (2010) 143–158 155

thank Riley E. Dunlap for his helpful comments on an earlier draft relationship in two Great Lakes areas of concern. The Toledo
of this article. Journal of Great Lakes’ Law, Science, & Policy, 13, 13–32.
18. Deng, J., Walker, G. J., & Swinnerton, G. (2006). A comparison of
Appendix B environmental values and attitudes between Chinese in Canada
References to studies included in the meta-analysis and Anglo–Canadians. Environment and Behavior, 38, 22–47.
19. Dunlap, R. E., Van Liere, K. D., Mertig, A. G., & Jones, R. E. (2000).
1. Amérigo, M., & Gonzalez, A. (2001). Los valores y las creencias Measuring endorsement of the new environmental paradigm:
medioambientales en relación con las decisiones sobre dilemas A revised NEP scale. Journal of Social Issues, 56, 425–442.
ecológicos. [The relationship of values and environmental 20. Environment Waikato. (n.d.). EW people’s environmental atti-
beliefs to decisions about ecological dilemmas]. Estudios de tudes. Retrieved June, 10, 2007, from: http://www.ew.govt.
Psicologı́a, 22, 65–73. nz/enviroinfo/indicators/community/communities/p2c/report.
2. Andersson, L., Shivarajan, S., & Blau, G. (2005). Enacting htm
ecological sustainability in the MNC: A test of an adapted 21. Eriksson, L., Garvill, J., & Nordlund, A. M. (2006). Acceptability
Value–Belief–Norm framework. Journal of Business Ethics, 59, of travel demand management measures: The importance of
295–305. problem awareness, personal norm, freedom, and fairness.
3. Arcury, T. A., & Christianson, E. H. (1990). Environmental Journal of Environmental Psychology, 26, 15–26.
worldview in response to environmental problems: Kentucky 22. Furman, A., & Erdur, O. (1995). Environmental attitudes among
1984 and 1988 compared. Environment and Behavior, 22, Bogazici university students: Testing the ‘‘environmentalist’’
387–407. hypothesis. Review of Social, Economic and Administrative
4. Barr, S., & Gilg, A. (2006). Sustainable lifestyles: Framing envi- Studies, 9, 109–122.
ronmental action in and around the home. Geoforum, 37, 23. Gooch, G. D. (1995). Environmental beliefs and attitudes in
906–92. Sweden and the Baltic states. Environment and Behavior, 27,
5. Berenguer, J., Corraliza, J. A., & Martin, R. (2005). Rural–urban 513–539.
differences in environmental concern, attitudes, and actions. 24. Hall, C., & Moran, D. (2006). Investigating GM risk perceptions:
European Journal of Psychological Assessment, 21, 128–138. A survey of anti–GM and environmental campaign group
6. Bjerke, T., Østdahl, T., Thrane, C., & Strumse, E. (2006). Vege- members. Journal of Rural Studies, 22, 29–37.
tation density of urban parks and perceived appropriateness 25. Hodgkinson, S. P., & Innes, J. (2000). The prediction of ecological
for recreation. Urban Forestry & Urban Greening, 5, 35–44. and environmental belief systems: The differential contribu-
7. Blaikie, N. W. H. (1992). The nature and origins of ecological tions of social conservatism and beliefs about money. Journal of
world views: An Australian study. Social Science Quarterly, 73, Environmental Psychology, 20, 285–294.
144–165. 26. Holt, D. T., & Lofgren, S. T. (2005). Examination of the new
8. Blake, D. E., Guppy, N., & Urmetzer, P. (1997). Canadian public environmental paradigm scale in a military sample. Perceptual
opinion and environmental action: Evidence from British and Motor Skills, 100, 791–794.
Columbia. Canadian Journal of Political Science, 30, 451–472. 27. Hunter, L. M., & Rinner, L. (2004). The association between
9. Bostrom, A., Barke, R., Turaga, R. M., & O’Connor, R. E. (2006). environmental perspective and knowledge and concern with
Environmental concerns and the new environmental paradigm species diversity. Society & Natural Resources, 17, 517–532.
in Bulgaria. Journal of Environmental Education, 37(3), 25–40. 28. Jager, W. (2006). Stimulating the diffusion of photovoltaic
10. Casey, P. J., & Scott, K. (2006). Environmental concern and systems: A behavioural perspective. Energy Policy, 34,1935–1943.
behaviour in an Australian sample within an ecocentric– 29. Johnson, C. Y., Bowker, J., & Cordell, H. K. (2004). Ethnic varia-
anthropocentric framework. Australian Journal of Psychology, tion in environmental belief and behavior: An examination of
58, 57–67. the New Ecological Paradigm in a social psychological context.
11. Castro, P., & Lima, M. L. (2001). Old and new ideas about the Environment and Behavior, 36, 157–186.
environment and science: An exploratory study. Environment 30. Kaiser, F. G., Hubner, G., & Bogner, F. X. (2005). Contrasting the
and Behavior, 33, 400–423. theory of planned behavior with the value–belief–norm model
12. Chung, S. S., & Poon, C. S. (2000). A comparison of waste in explaining conservation behavior. Journal of Applied Social
reduction practices and the new environmental paradigm in Psychology, 35, 2150–217.
four southern Chinese areas. Environmental Management, 26, 31. Kim, H., Borges, M. C., & Chon, J. (2006). Impacts of environ-
195–206. mental values on tourism motivation: The case of FICA, Brazil.
13. Chung, S. S., & Poon, C. S. (2001). A comparison of waste- Tourism Management, 27, 957–967.
reduction practices and new environmental paradigm of rural 32. Kortenkamp, K. V., & Moore, C. F. (2006). Time, uncertainty, and
and urban Chinese citizens. Journal of Environmental Manage- individual differences in decisions to cooperate in resource
ment, 62, 3–19. dilemmas. Personality and Social Psychology Bulletin, 32,
14. Clark, C. F., Kotchen, M. J., & Moore, M. R. (2003). Internal and 603–615.
external influences on pro–environmental behavior: Partici- 33. Kotchen, M. J., & Moore, M. R. (2007). Private provision of
pation in a green electricity program. Journal of Environmental environmental public goods: Household participation in green-
Psychology, 23, 237–246. electricity programs. Journal of Environmental Economics and
15. Cooper, P., Poe, G. L., & Bateman, I. J. (2004). The structure of Management, 53, 1–16.
motivation for contingent values: A case study of lake water 34. Kotchen, M. J., & Reiling, S. D. (2000). Environmental attitudes,
quality improvement. Ecological Economics, 50, 69–82. motivations, and contingent valuation of nonuse values: A case
16. Cottrell, S. P. (2003). Influence of sociodemographics and study involving endangered species. Ecological Economics, 32,
environmental attitudes on general responsible environmental 93–107.
behavior among recreational boaters. Environment and 35. Kurz, T., Donaghue, N., & Walker, I. (2005). Utilizing a social–
Behavior, 35, 347–375. ecological framework to promote water and energy conserva-
17. Davey, A. L., & Vertrees, R. L. (1999). Comparative analysis of tion: A field experiment. Journal of Applied Social Psychology, 35,
stakeholders’ attitudes toward the human–environment 1281–130.
156 L.J. Hawcroft, T.L. Milfont / Journal of Environmental Psychology 30 (2010) 143–158

36. Lai, C. L. J., Brennan, A., Chan, H. M., & Tao, J. (2003). Disposition 54. Schultz, P. W., & Zelezny, L. C. (1999). Values as predictors of
toward environmental hazards in Hong Kong Chinese: Vali- environmental attitudes: Evidence for consistency across 14
dation of a Chinese version of the environmental appraisal countries. Journal of Environmental Psychology, 19, 255–265.
inventory (EAI–C). Journal of Environmental Psychology, 23, 55. Schultz, P. W., Zelezny, L. C., & Dalrymple, N. J. (2000). A
369–384. multinational perspective on the relation between Judeo–
37. Leung, C., & Rice, J. (2002). Comparison of Chinese–Australian Christian religious beliefs and attitudes of environmental
and Anglo–Australian environmental attitudes and behavior. concern. Environment and Behavior, 32, 576–591.
Social Behavior and Personality, 30, 251–262. 56. Shafer, W. E. (2006). Social paradigms and attitudes toward
38. Liu, J. H., & Sibley, C. G. (2004). Attitudes and behavior in social environmental accountability. Journal of Business Ethics, 65,
space: Public good interventions based on shared representa- 121–147.
tions and environmental influences. Journal of Environmental 57. Shin, W. S. (2001). Reliability and factor structure of a Korean
Psychology, 24, 373–384. version of the new environmental paradigm. Journal of Social
39. Mainieri, T., Barnett, E. G., Valdero, T., Unipan, J. B., & Oskamp, S. Behavior & Personality, 16, 9–18.
(1997). Green buying: The influence of environmental concern 58. Slimak, M. W., & Dietz, T. (2006). Personal values, beliefs, and
on consumer behavior. Journal of Social Psychology, 137, ecological risk perception. Risk Analysis, 26, 1689–1705.
189–204. 59. Steel, B. S., List, P., Lach, D., & Shindler, B. (2004). The role of
40. Marshall, B. K., Picou, J. S., & Bevc, C. A. (2005). Ecological scientists in the environmental policy process: A case study
disaster as contextual transformation: Environmental values in from the American west. Environmental Science & Policy, 7, 1–13.
a renewable resource community. Environment and Behavior, 60. Steel, B. S., Lovrich, N. P., Lach, D., & Fomenko, V. (2005).
37, 706–728. Correlates and consequences of public knowledge concerning
41. Mayer, F. S., & Frantz, C. M. (2004). The connectedness to nature ocean fisheries management. Coastal Management, 33, 37–51.
scale: A measure of individuals’ feeling in community with 61. Steel, B. S., Smith, C., Opsommer, L., Curiel, S., & Warner-Steel, R.
nature. Journal of Environmental Psychology, 24, 503–515. (2005). Public ocean literacy in the United States. Ocean &
42. McFarlane, B. L., Stumpf–Allen, R. C. G., & Watson, D. O. (2006). Coastal Management, 48, 97–114.
Public perceptions of natural disturbance in Canada’s national 62. Steg, L., Dreijerink, L., & Abrahamse, W. (2005). Factors influ-
parks: The case of the mountain pine beetle (Dendroctonus encing the acceptability of energy policies: A test of VBN
ponderosae Hopkins). Biological Conservation, 130, 340–348. theory. Journal of Environmental Psychology, 25, 415–425.
43. Milfont, T. L., & Gouveia, V. V. (2006). Time perspective and 63. Steger, M. A., Pierce, J. C., Steel, B. S., & Lovrich, N. P. (1989).
values: An exploratory study of their relations to environmental Political culture, postmaterial values, and the new environ-
attitudes. Journal of Environmental Psychology, 26, 72–82. mental paradigm: A comparative analysis of Canada and the
44. Olli, E., Grendstad, G., & Wollebaek, D. (2001). Correlates of United States. Political Behavior, 11, 233–254.
environmental behaviors: Bringing back social context. Envi- 64. Tarrant, M. A., & Cordell, H. K. (1997). The effect of respondent
ronment and Behavior, 33, 181–208. characteristics on general environmental attitude–behavior
45. Pahl, S., Harris, P., Todd, H. A., & Rutter, D. (2005). Comparative correspondence. Environment and Behavior, 29, 618–637.
optimism for environmental risks. Journal of Environmental 65. Thapa, B. (1999). Environmentalism: The relation of environ-
Psychology, 25, 1–11. mental attitudes and environmentally responsible behaviors
46. Palmgren, C. R., Morgan, M. G., De Bruin, W. B., & Keith, D. W. among undergraduate students. Bulletin of Science, Technology,
(2004). Initial public perceptions of deep geological and & Society, 19, 426–438.
oceanic disposal of carbon dioxide. Environmental Science & 66. Vikan, A., Camino, C., Biaggio, A., & Nordvik, H. (2007).
Technology, 38, 6441–645. Endorsement of the new environmental paradigm: A compar-
47. Pierce, J. C., Lovrich, N. P., Tsurutani, T., & Abe, T. (1987). Culture, ison of two Brazilian samples and one Norwegian sample.
politics and mass publics: Traditional and modern supporters Environment and Behavior, 39, 217–228.
of the new environmental paradigm in Japan and the United 67. Watson, K., & Halse, C. M. (2005). Environmental attitudes of
States. The Journal of Politics, 49, 54–79. pre–service teachers: A conceptual and methodological
48. Rahelizatovo, N. C., & Gillespie, J. M. (2004). Factors influ- dilemma in cross–cultural data collection. Asia Pacific Education
encing the implementation of best management practices in Review, 6, 59–71.
the dairy industry. Journal of Soil and Water Conservation, 59, 68. Widegren, O. (1998). The new environmental paradigm and
166–175. personal norms. Environment and Behavior, 30, 75–10.
49. Rauwald, K. S., & Moore, C. F. (2002). Environmental attitudes 69. Zelezny, L. C., Chua, P. P., & Aldrich, C. (2000). Elaborating on
as predictors of policy support across three countries. Envi- gender differences in environmentalism. Journal of Social Issues,
ronment and Behavior, 34, 709–739. 56, 443–457.
50. Rideout, B. E., Hushen, K., McGinty, D., Perkins, S., & Tate, J.
(2005). Endorsement of the new environmental paradigm in
systematic and e–mail samples of college students. Journal of References
Environmental Education, 36(2), 15–23.
51. Schultz, P. W., Gouveia, V. V., Cameron, L. D., Tankha, G., Aiken, L. R. (1987). Formulas for equating ratings on different scales. Educational and
Psychological Measurement, 47, 51–54.
Schmuck, P., & Franek, M. (2005). Values and their relationship
Bechtel, R. B., Corral-Verdugo, V., & Pinheiro, J. Q. (1999). Environmental belief
to environmental concern and conservation behavior. Journal of systems: United States, Brazil, and Mexico. Journal of Cross-Cultural Psychology,
Cross–Cultural Psychology, 36, 457–475. 30, 122–128.
52. Schultz, P. W., Shriver, C., Tabanico, J. J., & Khazian, A. M. (2004). Bechtel, R. B., Corral-Verdugo, V., Asai, M., & Riesle, A. G. (2006). A cross-cultural
study of environmental belief structures in USA, Japan, Mexico, and Peru.
Implicit connections with nature. Journal of Environmental International Journal of Psychology, 41, 145–151.
Psychology, 24, 31–42. Blaikie, W. H. (1992). The nature and origins of ecological world views: an
53. Schultz, P. W., Unipan, J. B., & Gamba, R. J. (2000). Acculturation Australian study. Social Science Quarterly, 73, 144–165.
Blake, D. E., Guppy, N., & Urmetzer, P. (1997). Canadian public opinion and envi-
and ecological worldview among Latino Americans. Journal of ronmental action: evidence from British Columbia. Canadian Journal of Political
Environmental Education, 31(2), 22–27. Science, 30, 451–472.
L.J. Hawcroft, T.L. Milfont / Journal of Environmental Psychology 30 (2010) 143–158 157

Bogner, F. X., & Wiseman, M. (1999). Toward measuring adolescent environmental Maloney, M. P., Ward, M. P., & Braucht, G. N. (1975). Psychology in action: a revised
perception. European Psychologist, 4, 139–151. scale for the measurement of ecological attitudes and knowledge. American
Bogner, F. X., & Wiseman, M. (2006). Adolescent’s attitudes towards nature and Psychologist, 30, 787–790.
environment: quantifying the 2-MEV model. Environmentalist, 26, 231–237. Manoli, C. C., Johnson, B., & Dunlap, R. E. (2007). Assessing children’s environmental
Bostrom, A., Barke, R., Turaga, R. M. R., & O’Connor, R. E. (2006). Environmental worldviews: modifying and validating the new ecological paradigm scale for
concerns and the new environmental paradigm in Bulgaria. Journal of Envi- use with children. Journal of Environmental Education, 38, 3–13.
ronmental Education, 37, 25–40. Milfont, T. L. (2007a). An integrative framework for the study of the dimensionality
Carson, R. (1962). Silent spring. Boston: Houghton Mifflin. of environmental attitudes. Unpublished manuscript.
Clark, C. F., Kotchen, M. J., & Moore, M. R. (2003). Internal and external influences on Milfont, T. L. (2007b). Psychology of environmental attitudes: A cross-cultural study of
pro-environmental behavior: participation in a green electricity program. their content and structure. Unpublished doctoral dissertation. Auckland, New
Journal of Environmental Psychology, 23, 237–246. Zealand: University of Auckland.
Corral-Verdugo, V., & Armendáriz, L. I. (2000). The ‘‘new environmental paradigm’’ Milfont, T. L., & Duckitt, J. (2004a). Investigating the dimensionality of the
in a Mexican community. Journal of Environmental Education, 31, 25–31. revised new environmental paradigm (NEP) scale. Australian Journal of
Dietz, T., Stern, P. C., & Guagnano, G. A. (1998). Social structural and social Psychology, 56, 82.
psychological bases of environmental concern. Environment and Behavior, 30, Milfont, T. L., & Duckitt, J. (2004b). The structure of environmental attitudes: a first-
450–471. and second-order confirmatory factor analysis. Journal of Environmental
Dunlap, R. E., & Jones, R. E. (2002). Environmental concern: conceptual and Psychology, 24, 289–303.
measurement issues. In R. E. Dunlap, & W. Michelson (Eds.), Handbook of Milfont, T. L., & Duckitt, J. (2006). Preservation and utilization: understanding the
environmental sociology (pp. 482–524). Westport, CT: Greenwood Press. structure of environmental attitudes. Medio Ambiente y Comportamiento
Dunlap, R. E., & Jones, R. E. (2003). Environmental attitudes and values. In Humano, 7, 29–50.
R. Fernández-Ballesteros (Ed.), Encyclopedia of psychological assessment, Vol. 1 Milfont, T. L., & Duckitt, J. The environmental attitudes inventory: A valid and
(pp. 364–369). London: Sage. reliable measure to assess the structure of environmental attitudes. Journal of
Dunlap, R. E., & Van Liere, K. D. (1978). The new environmental paradigm. Journal of Environmental Psychology, in press.
Environmental Education, 9, 10–19. Milfont, T. L., Duckitt, J., & Wagner, C. A cross-cultural test of the value-attitude-
Dunlap, R. E., & Van Liere, K. D. (1984). Commitment to the dominant social paradigm behaviour hierarchy. Journal of Applied Social Psychology, in press-a.
and concern for environmental quality. Social Science Quarterly, 65, 1013–1028. Milfont, T. L., Duckitt, J., & Wagner, C. The higher order structure of environmental atti-
Dunlap, R. E., Van Liere, K. D., Mertig, A., & Jones, R. E. (2000). Measuring tudes: A cross-cultural examination. Interamerican Journal of Psychology, in press-b.
endorsement of the new ecological paradigm: a revised NEP scale. Journal of Milfont, T. L., Hawcroft, L. J., & Fischer, R. (2008). A meta-analysis of societal variables
Social Issues, 56, 425–442. associated with environmental attitudes. Unpublished manuscript. New Zealand:
Fischer, R., Milfont, T. L., & Gouveia, V. V. Does social context affect value structures? Centre for Applied Cross-Cultural Research, Victoria University of Wellington.
Testing the within-country stability of value structures with a functional theory Millennium Ecosystem Assessment. (2005). Ecosystem and well-being: Synthesis
of values. Journal of Cross-Cultural Psychology, in press. report. Washington, DC: Island Press.
Floyd, M. F., & Noe, F. P. (1996, May). An assessment of the revised new ecological Mohai, P., & Bryant, B. (1998). Is there a ‘‘race’’ effect on concern for environmental
paradigm scale among visitors at two National Park settings. Paper presented at quality? Public Opinion Quarterly, 62, 475–505.
the Sixth International Symposium on Society and Resource Management. Noe, F. P., & Snow, R. (1990). The new environmental paradigm and further scale
University Park, PA: Pennsylvania State University. analysis. Journal of Environmental Education, 21, 20–26.
Fontaine, J. R. J., Poortinga, Y. H., Delbeke, L., & Schwartz, S. H. (2008). Structural equiv- Nooney, J. G., Woodrum, E., Hoban, T. J., & Clifford, W. B. (2003). Environmental
alence of the values domain across cultures: distinguishing sampling fluctuations worldview and behavior: consequences of dimensionality in a survey of North
from meaningful variations. Journal of Cross-Cultural Psychology, 39, 345–365. Carolinians. Environment and Behavior, 35, 763–783.
Fransson, N., & Gärling, T. (1999). Environmental concern: conceptual definitions, Nunnally, J. C. (1978). Psychometric theory (2nd ed.). New York: McGraw-Hill.
measurement methods, and research findings. Journal of Environmental Olli, E., Grendstad, G., & Wollebaek, D. (2001). Correlates of environmental behav-
Psychology, 19, 369–382. iors: bringing back social context. Environment and Behavior, 33, 181–208.
Gardner, G. T., & Stern, P. C. (2002). Environmental problems and human behavior Osborn, F. (1948). Our plundered planet. Boston: Little, Brown.
(2nd ed.). Boston: Pearson Custom Publishing. Pierce, J. C., Lovrich, N. P., Tsurutani, T., & Abe, T. (1987). Culture, politics and mass
Geller, J., & Lasley, P. (1985). The new environmental paradigm scale: a reexami- publics: traditional and modern supporters of the new environmental para-
nation. Journal of Environmental Education, 17, 9–12. digm in Japan and the United States. Journal of Politics, 49, 54–79.
Gooch, G. D. (1995). Environmental beliefs and attitudes in Sweden and the Baltic Pierce, J. C., Steger, M. E., Steel, B. S., & Lovrich, N. P. (1992). Citizens, political
states. Environment and Behavior, 27, 513–539. communication, and interest groups: Environmental organizations in Canada and
Hall, C., & Moran, D. (2006). Investigating GM risk perceptions: a survey of anti-GM and the United States. Westport, CT: Praeger.
environmental campaign group members. Journal of Rural Studies, 22, 29–37. Pirages, D. C., & Ehrlich, P. R. (1974). Ark II: Social responses to environmental
Heberlein, T. A. (1981). Environmental attitudes. Zeitschrift für Umweltpolitik, 4, imperatives. San Francisco, CA: W.H. Freeman.
241–270. Rauwald, K. S., & Moore, C. F. (2002). Environmental attitudes as predictors of policy
Hedges, L. V., & Vevea, J. L. (1998). Fixed- and random-effects models in meta- support across three countries. Environment and Behavior, 34, 709–739.
analysis. Psychological Methods, 3, 486–504. Rideout, B. E., Hushen, K., McGinty, D., Perkins, S., & Tate, J. (2005). Endorsement of
Hunter, L. M., & Rinner, L. (2004). The association between environmental the new ecological paradigm in systematic and e-mail samples of college
perspective and knowledge and concern with species diversity. Society and students. Journal of Environmental Education, 36, 15–23.
Natural Resources, 17, 517–532. Robinson, J. P., Shaver, P. R., & Wrightsman, L. S. (1991). Measures of personality and
Inglehart, R., & Baker, W. E. (2000). Modernization, cultural change, and the social psychological attitudes. San Diego: Academic Press.
persistence of traditional values. American Sociological Review, 65(1), 19–51. Rosenthal, R., & DiMatteo, M. R. (2001). Meta-analysis: recent developments in quan-
Intergovernmental Panel on Climate Change. (2007). Climate change 2007: Impacts, titative methods for literature reviews. Annual Review of Psychology, 52, 59–82.
adaptation and vulnerability. Geneva: World Meteorological Organization. Schultz, P. W. (2001). The structure of environmental concern: concern for self,
Ji, C.-H. C. (2004). Factor structure of the new environmental paradigm scale: other people, and the biosphere. Journal of Environmental Psychology, 21,
evidence from an urban sample in Southern California. Psychological Reports, 94, 327–339.
125–130. Schultz, P. W., & Zelezny, L. C. (1999). Values as predictors of environmental atti-
Johnson, C. Y., Bowker, J., & Cordell, H. K. (2004). Ethnic variation in environmental tudes: evidence for consistency across 14 countries. Journal of Environmental
belief and behavior: an examination of the new ecological paradigm in a social Psychology, 19, 255–265.
psychological context. Environment and Behavior, 36, 157–186. Schultz, P. W., Gouveia, V. V., Cameron, L. D., Tankha, G., Schmuck, P., & Franek, M.
Kluckhohn, F. R. (1953). Dominant and variant value orientations. In C. Kluckhohn, (2005). Values and their relationship to environmental concern and conserva-
H. A. Murray, & D. M. Schneider (Eds.), Personality in nature, society and culture tion behavior. Journal of Cross-Cultural Psychology, 36, 457–475.
(2nd ed.). (pp. 342–366) New York: Knopf. Schultz, P. W., Shriver, C., Tabanico, J. J., & Khazian, A. M. (2004). Implicit connec-
Knight, A. J. (2008). ‘‘Bats, snakes and spiders, Oh my!’’ How aesthetic and nega- tions with nature. Journal of Environmental Psychology, 24, 31–42.
tivistic attitudes, and other concepts predict support for species protection. Schultz, P. W., Zelezny, L. C., & Dalrymple, N. J. (2000). A multinational perspective
Journal of Environmental Psychology, 28, 94–103. on the relation between Judeo–Christian religious beliefs and attitudes of
Kortenkamp, K. V., & Moore, C. F. (2006). Time, uncertainty, and individual differ- environmental concern. Environment and Behavior, 32, 576–591.
ences in decisions to cooperate in resource dilemmas. Personality and Social Schwartz, S. H. (1999). A theory of cultural values and some implications for work.
Psychology Bulletin, 32, 603–615. Applied Psychology: An International Review, 48, 23–47.
Krosnick, J. A. (1999). Survey research. Annual Review of Psychology, 50, 537–567. Schwarz, N. (1999). Self-reports: how the questions shape the answers. American
Krosnick, J. A., Judd, C. M., & Wittenbrink, B. (2005). The measurement of attitudes. Psychologist, 54, 93–105.
In D. Albarracı́n, B. T. Johnson, & M. P. Zanna (Eds.), The handbook of attitudes Stern, P. C. (1992). Psychological dimensions of global environmental change.
(pp. 21–76). Mahwah, NJ: Lawrence Erlbaum. Annual Review of Psychology, 43, 269–302.
Lipsey, M. W., & Wilson, D. W. (2001). Practical meta-analysis. Thousand Oaks, CA: Stern, P. C., Dietz, T., & Guagnano, G. A. (1995). The new ecological paradigm in
Sage. social-psychological context. Environment and Behavior, 27, 723–743.
Maloney, M. P., & Ward, M. P. (1973). Ecology: let’s hear it from the people. An Thapa, B. (2001). Environmental concern: a comparative analysis between students
objective scale for measurement of ecological attitudes and knowledge. Amer- in recreation and park management and other departments. Environmental
ican Psychologist, 28, 583–586. Education Research, 7, 39–53.
158 L.J. Hawcroft, T.L. Milfont / Journal of Environmental Psychology 30 (2010) 143–158

Van Liere, K. D., & Dunlap, R. E. (1981). Environmental concern: does it Weigel, R., & Weigel, J. (1978). Environmental concern: the development of
make a difference how it’s measured? Environment and Behavior, 13, 651–676. a measure. Environment and Behavior, 10, 3–15.
Verplanken, B., Walker, I., Davis, A., & Jurasek, M. (2008). Context change and travel Widegren, O. (1998). The new environmental paradigm and personal norms.
mode choice: combining the habit discontinuity and self-activation hypotheses. Environment and Behavior, 30, 75–100.
Journal of Environmental Psychology, 28, 121–127. Yencken, D., Fein, J., & Sykes, H. (2000). Environment, education, and society in the
Watson, K., & Halse, C. M. (2005). Environmental attitudes of pre-service teachers: Asia–Pacific. London: Routledge.
a conceptual and methodological dilemma in cross-cultural data collection. Asia Zelezny, L. C., Chua, P.-P., & Aldrich, C. (2000). Elaborating on gender differences in
Pacific Education Review, 6, 59–71. environmentalism. Journal of Social Issues, 56, 443–457.

You might also like