The Effects of Children's Age and Sex on Acquiring Pro-Environmental Attitudes through Environmental Education

Environmental education programs aiming to enhance children's environmental attitudes in a pro-environmental direction require background information, such as age and sex differences, to ensure appropriate design. We used the 2-MEV model with its domains preservation and utilization of nature to assess a four-day program at an educational field center with students 9–10 and 11–13 years of age. A pre- and post-retention test design revealed younger students to be more responsive concerning positive attitude shifts than older students, whereas the sexes were equally influenced. Program developers should consider that education has a stronger effect on young children's environmental attitudes.

Keywords: age differences; comprehensive education; environmental attitudes; gender differences; 2-MEV model; water

Research about attitudes has a long tradition and is one of the most frequently studied concepts in the field of psychology (e.g., Dunlap, van Liere, Mertig, & Jones, 2000; Eagly & Chaiken, [21]; Thurstone, [41]). Environmental psychologists specify attitudes by referring to preferences or the evaluative tendency "a person holds regarding environmentally related activities or issues" (Schultz, Shriver, Tabanico, & Khazian, [34], p. 31).

Education aims at achieving a sustainable, positive enhancement of attitudes in a pro-environmental direction. When we speak of pro-environmental attitudes, we mean attitudes which may lead to actions that either prevent or reduce harm to the environment or that may even benefit the environment (e.g., Steg & Vlek, [37]). Enhancing pro-environmental attitudes of students, however, is an important but inevitably slow process (Bogner, [6]). The importance of intervening to enhance attitudes is neatly demonstrated in the Irish plastic bag example described by Dobson ([17]): The government encouraged the use of reusable bags by imposing a fee on plastic shopping bags. This approach aimed at enhancing individual attitudes toward littering and pollution. Indeed, in the follow up, 90% of all bags were removed from circulation. However, did the attitudes regarding litter and pollution really change or did only the behavior change? What would happen if plastic bags were again free of charge? Most people would certainly use the free plastic bags again, meaning that pro-environmental attitudes had not been promoted. Therefore, only enhancement in attitude can ensure pro-environmental behavior, for example, in the absence of externally imposed sanctions.

Many researchers have developed instruments to measure environmental attitudes. These instruments can be used to investigate the effect of educational programs on the development of pro-environmental attitudes. We will briefly describe the most commonly used constructs to allow a comparison with the construct used in our study. First, we consider the revised New Environmental Paradigm (Dunlap & Van Liere, [19]; Dunlap et al., [20]). This instrument is designed to measure a person's endorsement of an ecological worldview (also referred to as pro-environmental orientation, environmental concern or values). The NEP refers to environmental attitude as a one-dimensional, bipolar construct, which varies from unconcerned to concerned about the environment.

Second, we consider multidimensional constructs that are related to value-based orientations and consist of either two or three dimensions (Milfont & Gouveia, [28]). Thompson and Barton's (1994) instrument measures attitudes with a two-dimensional scale: ecocentric (concern for all living things) and anthropocentric (concern for humans). In comparison, Stern and Dietz's (1994) theory uses a tripartite value orientation: egoistic concern (concern for self), altruistic concern (concern for other people), and biospheric concern (concern for the biosphere). Measurement instruments to capture these orientations are Schwartz's (1977) value items or Schultz's (2001) Environmental Motives Scale.

The model we have chosen for our study is Bogner and Wiseman's (1999, 2002, 2006), Two Major Environmental Values (2-MEV) model, which is based on a theoretical framework comprising environmental attitudes that can be grouped under two distinct higher-order-factor domains: preservation and utilization. The ecological value preservation describes a (more) selfless environmental attitude and is determined by "a biocentric dimension that reflects conservation and protection of the environment." Utilization, on the other hand, describes a self-interested environmental attitude and is determined by "an anthropocentric dimension that reflects the utilization of natural resources" (Wiseman & Bogner, [45], p. 5). From this, it follows that a person who has a high preservation attitude and a low utilization attitude holds pro-environmental attitudes. The theory posits preservation and utilization into four quadrants, not a linear scale. This orthogonal allocation permits a change of an individual's position on one dimension independent of the other dimension. This is important because "people who have strong Preservation (biocentric) attitudes do not necessarily have weak Utilisation (anthropocentric) attitudes" (Johnson & Manoli, [25], p. 87). Assuming this, a conflict between the protection of the environment and the need to make use of natural resources, which can occur in real life, can be excluded (Bogner & Wiseman, [13]).

Other groups working with the 2-MEV model have independently determined that the psychometric properties of the model are sound (Boeve-de Pauw & Van Petegem, [5]; Drissner, Haase, & Hille, [18]; Milfont & Duckitt, [27]; Munoz, Bogner, Clement, & Carvalho, [29]). The measure has undergone various validation steps and its dichotomous structure has proven to be consistent, reliable, and highly valid in diverse settings (across 22 languages and age barriers), for example, with pre- and in-service teachers in 16 countries in and neighboring Europe, with adolescents in Belgium and the USA, as well as with freshman students in New Zealand (Boeve-de Pauw & Van Petegem, [5]; Drissner et al., [18]; Johnson & Manoli, [24]; Milfont & Duckitt, [27]; Munoz et al., [29]).

The great variety of scales makes it difficult to compare and evaluate intervention studies or educational programs. Therefore, it is necessary to implement the same measure in different studies to allow a comparison across studies (Bogner, [8]; Bogner & Wiseman, [13]). The 2-MEV scale consists of 20 items which are easy to understand and which can be answered quickly via a 5-point-Likert scale, which varies from 1 (strongly disagree) to 5 (strongly agree). These features make this model especially convenient for evaluating educational programs for adolescents and children via repeated measurement designs (Bogner & Wiseman, [13]; Johnson & Manoli, [25]). Reviewing results from previous research allows new outcomes based on the 2-MEV model to be compared.

First, we focus on baseline studies, which only measure the current state of environmental attitudes with the 2-MEV model in study groups differing in location, age, and gender. German university students (mean age ± SD = 22.12 ± 1.55 years; Wiseman, Wilson, & Bogner, [46]) scored about 3.8 on the preservation dimension and 2.1 on the utilization dimension. In the United States, Johnson and Manoli (2011) investigated the environmental attitudes of students 9–12 years of age and found an average preservation score of 3.7 and utilization score of 2.5. Both studies found that respondents show a tendency to agree with preservation and refuse utilization. With reference to gender, studies from Western Europe revealed young women score higher on preservation and lower on utilization compared to young men (Bogner & Wiseman, [10], 2006; Wiseman & Bogner, [45]). The same pattern was found for New Zealand introductory psychology students (Milfont & Duckitt, [27]). Further studies with German pre- and in-service teachers and with Flemish students revealed the same result for the utilization dimension; however, they found no significant differences between the preservation scores of male and females (Boeve-de Pauw & van Petegem, [5]; Oerke & Bogner, [30]).

Intervention studies, like out-of-school learning or environmental education programs, can be designed to foster pro-environmental attitudes. The 2-MEV model has been used as an empirical basis for many previous intervention studies and is therefore useful for comparing differing samples and educational programs. Some programs were able to improve both of the measured environmental attitudes, with preservation scores increasing and utilization scores decreasing (e.g., Bogner & Wiseman, [12]; Johnson & Manoli, [25]). In other studies, only the utilization or the preservation scores improved (e.g., utilization: Bogner, [8]; Drissner et al., [18]; preservation: Bogner, [7]).

The intervention programs differed in length, content, as well as the age of participants. Johnson and Manoli ([25]) used outdoor programs for students in grades four to six. The programs dealt with abstract ecological concepts which were made more concrete. Both environmental attitudes, preservation and utilization (modified 2-MEV scale), stayed improved with moderate effect sizes four to six weeks after the interventions. Drissner et al. ([18]) established an experiential learning forum outside school for grades four to five. Their intervention lasted only three hours and students had direct contact with small animals, which led only to an improvement of utilization. They argue that their program might have been too short for an enhancement of preservation, especially since the program content did not specifically target issues of preservation. A three-day program on an endangered migratory bird species for students 11–16 years of age was found to lead to an improved preservation preference (Bogner, [7]); this result is most likely due to the program topic, which focused on bird conservation. Sellmann and Bogner ([36]) conducted a one-day intervention study in a botanical garden with tenth grade students on the topic of global climate change. The program had a positive short-term effect on preservation and utilization directly after the program, but only utilization persisted four to six weeks after the program. They argue that the topic taught mainly dealt with the utilization factor and had no focus on conservation or preservation issues. In conclusion, the enhancement of pro-environmental attitudes probably depends on the topic taught. To improve both preservation and utilization attitudes, it seems necessary to carefully design programs that cover both dimensions sufficiently. Until now, there has been no investigation as to whether students of differing ages or sex are equivalently accessible concerning their environmental attitudes due to participation in environmental education, which is highly relevant for education.

The following four research questions will be addressed: (a) Are we able to confirm the orthogonal structure of the 2-MEV model? (b) Do children in the age groups 9–10 and 11–13 score differently on preservation and utilization? (c) Based on previous literature, can we confirm that female students and male students score differently on the two environmental attitudes? (d) Does our environmental education program enhance the pro-environmental attitudes of (i) both age groups and (ii) both sexes? The answer to these questions may alter how we implement and design effective environmental education according to the time in a student's life and the methods used.

Students from German (Bavarian) schools formed two main subsamples, which consisted of students participating in our educational program or students assigned to an external control group (Table 1). The subsamples were further divided according to grade: fourth grade students (9–10 years of age) from primary schools and sixth grade general-education track students (11–13 years of age).

TABLE 1 Number, Age, and Sex Distribution of the Study Group Divided Into the Subsamples Participants and Control

Overall, our study group consisted of 15 participating classes and had an average group size of M ± SD = 20.8 ± 5.22 students. The participant group consisted of 11 classes, which had already been registered by their schools to spend a week at an educational field center before they were invited to take part in our study. The control group consisted of students from four classes without program participation who only completed our questionnaires. Both groups were convenience samples. Only those students whose parents signed consent forms to participate in the study were included in the data analysis. Each student was assigned a strictly confidential code that was used on the students' T0, T1, and T2 questionnaires to link them for statistical analysis.

Our residential environmental education program was titled "Water in Life-Life in Water" and was adapted to the Bavarian curriculum for fourth grade students and sixth grade general education students. During late spring and summer, students spent four days at the educational field center and received approximately six hours of guidance every day from the same instructor. The time schedule ensured that each participating class underwent the same activities, structure, and time frame. Various methods and subtopics were included in the program design to account for gender preferences concerning learner types, learning style, and interests. For example, girls prefer collaboration and socially relevant topics, whereas males prefer competition and learning about how things work (Weber & Custer, [42]). We addressed both types of gender preferences by including group work and competitive elements as well as socially relevant content and experiments. The program also covered various cognitive and affective aspects as well as formal and informal aspects aimed at improving the environmental attitudes and knowledge of both sexes along with their relationship to nature. Interactive didactic methods such as "learning at workstations" (e.g., Gerstner & Bogner, [23]) or "jigsaw puzzle" (Aronson, [2]) were implemented to foster environmental understanding, to assist students in interactively confronting local and global water problems, as well as to teach students how to lessen their individual impacts on the water supply. Direct multisensory nature-encounters were used to strengthen positive attitudes toward nature (e.g., discovery and perception games and biological water quality assessment). To sum up, we have implemented numerous and differing elements to promote pro-environmental attitudes. To give one example, our program promoted pro-environmental utilization by having the participants explore the negative impact they may have on water; and the program promoted their preservation attitude by having them determine ways of protecting water.

We selected 16 out of the 20 items from Bogner and Wiseman's 2-MEV scale (2006; original German version; see Appendix). We adapted the questionnaire to the cognitive capacity of our younger students by omitting two items from each scale. We embedded the eight preservation and the eight utilization items into a larger, paper-and-pencil questionnaire containing 31 additional items (on environmental knowledge and connectedness to nature) which took about 15 minutes for students to complete. The students responded to the scale items on a 5-point-Likert scale, ranging from 1 (strongly disagree) to 5 (strongly agree), with an "undecided" category (3). This "undecided" category does not force a response, which is especially important for maintaining young children's interest (Johnson & Manoli, [25]). We administered the questionnaire three times, always presenting the items in a differing order: T0 (pre-test at school two weeks before participation in the program), T1 (post-test directly after the program at the educational field center), and T2 (follow-up test at school four to six weeks later). The control group completed the questionnaires without any program participation before T2. Teachers whose classes took part in the study were advised not to teach their students anything on the curricular topic of water until after T2. The percentage of missing answers was low (1.47%) for all test times. Only those students who completed all three questionnaires and at least 80% of each individual 2-MEV scale were included in the analyses.

To assess the dichotomous structure of the 2-MEV model, we conducted a confirmatory principal-axis factor analysis with oblique rotation (direct oblimin; post-test data; see also Bogner & Wiseman, [13]; Johnson & Manoli, [25]; Munoz et al., [29]). A bivariate correlation of preservation with utilization was calculated.

We used the sum mean of each person for the calculation of overall preservation and utilization preferences for both age groups and both sexes. For the baseline studies (age and gender), we merged the pre-test data of the intervention and control group. According to the central limit theorem, we assume normality and apply parametric tests (Wilcox, [44]) to determine any changes in preservation and utilization scores due to the environmental education program.

The scale reliability of the pre-test data of the whole study group expressed as Cronbach's α was moderate with α =.62 for preservation and α = 0.60 for utilization. The inter-relatedness of the items is high (Pearson correlation for preservation: T0–T1: r =.485; T0–T2: r =.548, T1-T2: r =.697; for utilization: T0–T1: r =.691; T0–T2: r =.651, T1–T2: r =.760).

The principal-axis factoring analysis clearly supports the proposed dichotomous structure of the 2-MEV model because both dimensions fall on different axes (Figure 1). Preservation accounted for 18.5% and utilization for 14.3% of the total variance. The bivariate correlation between utilization and preservation show both factors to be almost completely independent (pre-test: r = –0.034, p =.557; post-test: r = –0.119, p =.051; retention test: r =.192, p =.002).

Graph: FIGURE 1 Orthogonal structure of the 2-MEV preservation and utilization dimensions. Note: Post-test results given as an example homologous to the pre- and retention test results.

The baseline scores of the students 9–10 years of age differ significantly from the students 11–13 years of age for preservation (t = 4.90, df = 265, p <.001, r =.29) and utilization (t = −2.14, df = 265, p =.033, r =.13). With 4.34 ± 0.44 on preservation and 2.47 ± 0.63 on utilization, the younger students show a higher preservation and a lower utilization preference than the older students (preservation: 4.03 ± 0.57, utilization: 2.63 ± 0.58).

We calculated program-induced enhancement of preservation and utilization for both age groups. In the following, when speaking of improvement, we refer to a change toward more pro-environmental attitudes with preservation preference rising and utilization preference falling. Over all test times, we find the younger students 9–10 years of age differ significantly from the older students 11–13 years of age in the preservation dimension (F (1, 577) = 88.58, p <.001) and in the utilization dimension (F (1, 577) = 22.82, p <.001). Differences between the test times are shown in Table 2. The program positively promoted the preservation and utilization preference of our younger students over the short term (T0 to T1) and long term (T0 to T2). However, only the improvement in the preservation dimension persisted after the program, while the improvement in the utilization dimension significantly decreased. Significant effects were also seen for the older students' preservation dimension; however, the effects persisted just over the short term with a medium effect size (Table 2). We detect no enhancement in the utilization dimension for the older students. We find no effects from filling in questionnaires as determined by comparison with the control group (preservation: F (2, 211) = 0.45, p =.642, utilization: F (2, 211) = 0.13, p =.876).

TABLE 2 Enhancement of Environmental Attitudes With Regard to Student Age

Baseline results on the environmental attitudes of the sexes show no differences between female and male students, neither for the younger (preservation: t = –0.64, df = 175, p =.526; utilization: t = –0.24, df = 175, p =.810), nor for the older students (preservation: t = –0.89, df = 88, p =.377; utilization: t = 0.15, df = 88, p =.882).

We used a multivariate test with the variables sex and grade to reveal program-induced differences. As described previously (see Table 2), we again find a differing program effect on the younger compared to the older students (preservation: F (3, 187) = 14.82, p <.001; utilization: F (3, 187) = 4.52, p =.004). Male and female students, however, show no difference related to program participation (preservation: F (3, 187) = 0.80, p =.498; utilization: F (3, 187) = 0.51, p =.679). We also find no interrelationship between age and sex (preservation: F (3, 187) = 2.54, p =.058; utilization: F (3, 187) = 0.72, p =.544). The program had no significant sex effects with regard to the enhancement of students environmental attitudes.

Our study contributes to the field of attitude research by considering the initial attitudes and the effects of participation in a four-day environmental education program on students of different ages and sexes. We confirmed the dichotomous orthogonal 2-MEV's structure with regard to its principal domains preservation and utilization with a principal-axis factor analysis. The model allows for a possible change of one attitude independent from the other, which is also supported by our correlation result and those results reported by Boeve-de Pauw and van Petegem ([5]). The explained variance of our study sample, 32.8%, accounted for by preservation and utilization was in line with the results from literature (Bogner and Wiseman [13]: overall explained variance 38.5%).

Regarding the baseline scores, our participants' scores were slightly higher for preservation than the baseline scores specified in earlier studies and similar for utilization (Bogner & Wilhelm, [9]; Johnson & Manoli, [25]; Wiseman et al., [46]). Our high initial scores, especially on the preservation scale, limit the potential space for improvement in a positive direction (ceiling effect). Studies that rely on the self-reporting of young children also need to consider social desirability when interpreting results as younger children were found to show a higher social desirability than their older counterparts (Boehnke, Silbereisen, Reynolds, & Richmond, [4]; Oerke & Bogner, [31]). For instance, Oerke and Bogner ([31]) investigated the influence of "socially acceptable" answers on students' responses on the 2-MEV scale, similarly portraying a significant effect of age on preservation. This was consistent with Wiseman and Bogner ([45]), who also found Lie scores to be a highly significant predictor for preservation scores. Both studies found no influence on the utilization dimension. Therefore, we must consider that our initial preservation scores may very likely have been increased due to social desirability, especially in our younger cohort. A previous publication, based on the same environmental education program as the current study, finds that younger students have a higher connectedness to nature than the older students (Liefländer, Fröhlich, Bogner, & Schultz, 2012). The authors argue that the onset of puberty and the accompanying psychological development of the older students might explain the change toward less favorable connectedness to nature. Piaget's developmental theory (as described in Berk, [3]) can explain the differences we encountered between our two age groups (9–10 and 11–13 year old students) as attitudes are subject to change. Our younger students are in the "concrete operational stage," ranging from around seven to eleven years of age. In this stage, children are oriented toward role models like parents or teachers. This may explain why our program had a stronger impact on the younger participants. Our older students, however, have probably reached the consecutive "formal operational stage" which is entered at around eleven years of age. Adolescents, for example, develop the ability of abstract scientific thinking (Berk, [3]). Within this stage, a detachment process from parents takes place which leads to emotional autonomy (Parra & Oliva, [32]). The presence of peers may influence the responses to questionnaire items dramatically (Borgers, de Leeuw, & Hox, [14]). This may account for the comparably poor effect of our program on the 11–13-year-old general education students.

On the basis of the orthogonal structure of the 2-MEV model, an environmental education program may influence only one or both environmental attitudes. Although we intended to cover both environmental attitudes through the program design, we found a more sustainable influence on the preservation dimension than on the utilization dimension. Following the assumptions of former intervention studies (e.g., Bogner, [7]; Drissner et al., [18]; Sellmann & Bogner, [36]), this is probably due to the program content. Despite our careful design, it is possible that the program content had a more positive overall influence on preservation. Another consideration points to the questionnaire itself: Inspecting the items for utilization and preservation reveals that three preservation items explicitly deal with the topic water, whereas there is only one utilization item (see Appendix: Addition to the original wording). Therefore, our program content and survey items had a major influence on the findings.

The enhancement of environmental attitudes also differs between younger and older students, confirming our expectations. In contrast to the positive effects on the younger students, the preservation dimension of the older students only increased over the short term and the utilization dimension was not impacted at all due to program participation. The differing developmental stages of our age groups, as we have mentioned previously, might also explain the limited effect of the program on the older cohort.

Referring to gender, we find no baseline differences between boys and girls. This is inconsistent with the majority of literature, which mostly finds female participants to score lower on utilization and higher on preservation, indicating they have stronger pro-environmental attitudes than male participants do (e.g., Bogner & Wiseman, [12], 2006; Wiseman et al., [46]; and review by Zelezny, Chua, & Aldrich, [47]). Some other studies, however, also report no differences in environmental concern between the sexes (Arcury & Christianson, [1]; Widegren, [43], as cited by Zelezny, Chua, & Aldrich, [47]). The impact of participation in our program did not enhance pro-environmental attitudes in girls and boys differently, which may indicate a suitable program design regarding gender. The reason may be the diverse program topics and methods, which we used to address the various possible preferences in learning style and content.

To conclude, we give recommendations on future research, partly arising from study limitations: One important factor influencing environmental attitudes is the level of education (Buttel, [15]). Our general education track students from grade 6 may differ in attitudes from their same-age university track counterparts, whom we have not tested. To further clarify the enhancement of pro-environmental attitudes through environmental education, future research should therefore examine whether low- and high-achieving students also differ in their environmental attitudes. The 2-MEV scale is an established scale and its wording was not adjusted to the content of our educational program since the success of a program should be rather independent of the topic taught. The 2-MEV items are suitable for young children (like our young 9–10 year old participants) as they avoid using negations. This is important for children in the "concrete operational phase" (up to 11 years of age) as they were found to have problems with understanding negatively phrased items (Borgers et al., [14]). It seems advisable to use the whole scale consisting of 20 items to ensure high scale reliability as our Cronbach's alpha values are below the commonly accepted values of alphas between.70 and.90. Tavakol & Dennick ([39], p. 54) claim that "A low value of alpha could be due to a low number of questions, poor inter-relatedness between items or heterogeneous constructs." We had reduced the original, established, and valid 2-MEV scale from 20 to 16 items. The implemented scale therefore had a lower number of items (8 items per subscale), while the inter-relatedness was high and its unidimensionality was confirmed by the factor analysis. We conclude that the low number of items can be held responsible for the moderate Cronbach's alpha. We think it might also be feasible to include additional items based on the environmental attitudes conveyed by the program. These questions outside the established 2-MEV scale can be used to estimate the program-specific outcome on attitudes. If these items are found to have a high quality and are unidimensional with preservation or utilization, they could be included in a future version of the 2-MEV. Especially, more "difficult" items are needed to prevent a ceiling effect, which we encountered for the preservation attitudes of our young participants.

Due to its orthogonal structure, the empirical 2-MEV model allows shifts in the environmental attitudes preservation and utilization to be recorded. Our article advances the current literature by using the 2-MEV model to record and show the attitudinal differences between the age groups 9–10 and 11–13 years of age: Our younger students obtained and maintained higher pro-environmental attitudes than their older counterparts. The effect of environmental education which aims at enhancing pro-environmental attitudes may be more effective with younger children and may become less effective and more difficult to implement with increasing age. The factor sex does not necessarily have an influence on the enhancement of pro-environmental attitudes if content and methods are used that are appropriate for both sexes. Researchers and educators need to be aware that environmental education programs (lasting four or more days) that aim to enhance environmental attitudes, may be more effective for younger children than for older children regardless of gender.

We wish to thank all participating students and teachers and we acknowledge the continuous support of the educational field center "Schullandheim Weißenstadt" during the implementation phase, in particular, Cornelia Riedel und Anne Gebhardt.

The 2-MEV items applied in the study (retrieved from Bogner & Wiseman, [13]).