Factors Influencing Adults’ Environmental Attitudes and Behaviors and the Role of Environmental Schools in Influencing Their Communities

Abstract

The present study revisits a subfield of environmental education: significant life experiences, which studies the influences that shape the development of environmental stewardship. In the present study, we examine the effect of various formative experiences on a group of adults and analyze the role of school, as a formative influence on the parents of the students. By employing factor analysis, we were able to differentiate between groups of variables influencing attitudes and those influencing behavior. Cluster analysis enabled us to differentiate between types of respondents in accordance to their responsiveness to influencing experiences. The results draw attention to (a) the different pathways by which environmental attitudes and behaviors are influenced; (b) the important role of the inner self (“personality”) in organizing and giving meaning to all other formative influences; and (c) the effectiveness of environmental schools in urban communities in Israel, in influencing the behavior of students’ parents.

Keywords

significant life experiences environmental education influence schools’ influence on communities formative environmental influence environmental influence on adults

Within the field of environmental education research, a subfield of environmental-education-related significant life experiences (SLE) emerged in the 1980s. It was instigated by the work of Tanner (1980) who conducted a retroactive study among a group of environmental activists. This stream of research aims to uncover the sources of proenvironmental attitudes and behaviors. The underlying assumption is that if we can unveil significant experiences that lead to development of environmental stewardship, then perhaps it will be possible to implement these experiences in educational programs and thus enhance the goals of environmental education.

The work of Tanner was followed by international research that was carried out in many countries (e.g., Chawla, 1999; Hsu, 2009; Myers, 1997; Palmer et al., 1998; Palmer & Suggate, 1996; Palmer, Suggate, Robottom, & Hart, 1999). SLE that were identified include the following: education, child/adult natural experiences, work, people, media, and others (Palmer et al., 1999).

In 1999, the journal Environmental Education Research (Volume 5, Issue 4) published a special issue devoted entirely to the debate regarding SLE research and methodology. Among the many issues that were debated in the above journal and in other publications, two critical questions arose, as follows: (a) What are the appropriate methodologies for SLE research? (Chawla, 1998, 2001; Dillon, Kelsey, & Duque-Aristizabal, 1999; A. Gough, 1999; N. Gough, 1999; Payne, 1999) and (b) Can past childhood experiences of adults, be used as a basis for creating influencing experiences for today’s children (A. Gough, 1999; N. Gough, 1999)?

The following sections aim to contribute to the discussion regarding SLE from new perspectives as outlined in what follows.

Research Objectives

The present study is tangential to and diverges from the body of research on SLE both methodologically and conceptually. It also examines new questions that have not previously been dealt with in SLE research, regarding the relationships between environmental schools and urban communities as sources of formative influences on parents of school children.

The following sections introduce the main claims and aims of the present study. Further elaboration will be presented subsequently.

Distinguishing Influences on Attitudes From Influences on Behavior

Previous SLE research did not differentiate between formative influences on attitudes and those affecting behavior (regarding each participant in a given study), neither in the designs of the studies nor in their analyses and results. Similar formative influences were found for both attitudes and behaviors despite the well-accepted notion regarding differences between the two traits. Research in the field has taught us that responsible environmental behavior and attitudes belong to different cognitive compartments and are not directly interrelated (Abelson, 1972; Doyle, 1997; Hines, Hungerford, & Tomera, 1987; Hungerford & Volk, 1990; Wallace, Paulson, Lord, & Bond, 2005). This inconsistency calls into question the sensitivity of the applied methodologies. It is expected that if attitudes are acquired independently from (or not in direct correlation with) environmental behavior, then the variables influencing the two distinct traits will not be identical.

Derived from this concept is the objective to analyze formative influences on attitudes separately from formative influences on behavior.

Identifying Meaningful Interactions Between Formative Influences

We assume that the interaction between formative environmental experiences is at least as important as each individual experience on its own, if not more so. As none of the previously identified formative influences on its own can explain development of environmental stewardship, it is important that any further research on this issue will focus on the interactions between the experiences rather than on identifying discrete experiences.

Derived from this notion is the objective to identify experiences that have strong interactions with each other to the degree that they can be grouped together to form categories (factors) of influence on attitudes and categories of influence on behavior.

Classifying Populations in Accordance to Responsiveness to Formative Influences

It is possible to assume that different people will react differently to the same formative influences (A. Gough, 1999; Hsu, 2009; Payne, 1999). Therefore, it is worthwhile to differentiate between groups of people according to the differences in their reactions to a given set of formative environmental influences.

Derived from this notion is the objective to identify clusters of people, within the sample, who exhibit similarities in their reactions to potentially influential experiences. The analysis needs to be done separately for formative influences on attitudes and for formative influences on behaviors.

Identifying the Role of Personality

The arena in which all influencing experiences meet and interact is the inner self. While most other formative experiences can theoretically be manipulated and measured using experimental methods, the influence of the “self” or “personality” (as it is termed in the present article) is mostly unknown vis-a-vis the present context. If we seek to understand the relative impact of various formative influences, there is a need to gain some understanding regarding the function of the “personality” in constructing and organizing formative environmental experiences.

Derived from this concept are the following objectives: (a) To find out which influencing experiences the variable “personality” interacts with most strongly; and (b) to identify the relative impact of “personality.” The analyses need to be carried out separately for influences on attitudes and influences on behaviors.

The Influence of Environmental Schools on Parents

In Israel and worldwide, more and more schools are becoming environmental schools. Ballantyne, Connel, and Fien (2006) drew attention to the potential role of environmental schools’ students in influencing their communities. If these processes do take place, it is reasonable to expect from environmental schools to become one of the formative influences on parents of children who study in these schools.

Derived from this notion is the objective to analyze the formative influence of environmental schools on their students’ parents, in comparison to the influence of nonenvironmental schools on their students’ parents.

Formative Influences as an Outcome of Interactions Between Schools and Communities

There is a growing interest in environmental education as a means of producing a wholesome educational experience in which schools and communities interact together to produce an influencing effect on environmental literacy (Armstrong & Bottomley, 2003; Ballantyne et al., 2006; Hopkins & McKeown, 2002; Organisation for Economic Co-operation and Development, 2001; United Nations Educational, Scientific and Cultural Organization, 2005; Victorian Curriculum and Assessment Authority, 2005). If such effective relationships do exist, they are expected to influence the parents of children who study in environmental schools.

Derived from this notion is the objective to analyze the extent of interaction between schools and their communities with regard to influences on attitudes and behavior of school children’s parents. For comparison, the analysis will be carried out among adults whose children study in environmental schools and among adults who live in the same urban communities and whose children study in regular, not environmental, schools.

Methodological Approach

This section presents the methodological approach applied in our research in relation to previous SLE research approaches. Positioning our methodology in this way allows our approach to emerge as a response to previously applied methodologies, thus creating a “discussion” between the various approaches. Our methodological approach is distinguished by two main characteristics, as discussed in what follows.

Developing Categories Based on the Level of Interactions Between Variables as A Posteriori Outcome Rather Than A Priori Forming Discrete Categories

In previous SLE research, the most commonly applied methodology was using interviews, questionnaires, and biographies, from which researchers derived discrete categories of significant formative experiences. The frequency of experiences pertaining to each category was then calculated in a disconnected manner (Arnold, Cohen, & Warner, 2009; Chawla, 1998; Palmer, 1993; Palmer et al., 1998; Peterson, 1982; Sward, 1999; Tanner, 1980).

This a priori approach to the data is qualitatively mediated by the researcher. Its main drawback is that the formed categories are relatively rigid and do not provide information regarding relationships between the variables that form the categories. In the present study, the allocation of variables into categories is done a posteriori by applying a statistical method of factor analysis to participants’ responses. By applying this method, it becomes possible to form categories (factors) that are based on the levels of correlation between the variables. Each variable in the analysis is a formative influence, and each factor is composed of formative influences that are interconnected.

Analyzing Differences Between People in Their Responsiveness to Formative Experiences Rather Than Differences in Exposure to Experiences

Most previous SLE research focused on uncovering the influencing experiences that shaped the worldviews of environmental activists and educators. (Chawla, 1998, 1999; Corcoran, 1999; Palmer, 1993; Palmer et al., 1998, 1999; Palmer & Suggate, 1996; Sward, 1999; Tanner, 1980). In the present study, we are diverting the attention from the actual experiences to the people who were exposed to the experiences. The SLE literature reports similar significant experiences in most of the studies. These include experiences such as: Extended time spent outdoors in natural areas (often in childhood), adult role models, education and others. These experiences on their own are not unique in the course of many people’s lives. We can assume that many people who did not grow up to be environmental activists spent many hours of childhood outdoors or were exposed to adult role models. A characteristic distinguishing these people from others who became activists might be the different ways in which they responded to these formative experiences. For analyzing how nonpredefined group of participants are distributed with regard to their responsiveness to a given set of experiences, we applied a method of cluster analysis. Unlike factor analysis, which allocates variables into categories, cluster analysis allocates participants into categories, thus allowing us to differentiate groups of respondents according to their responsiveness to a given set of experiences.

Method

Participants

The research was carried out in Israel, in two cities with similar population density. The first city is Haifa, the core city of the northern metropolitan area. Haifa’s population size is 265,000, and its urban density is 4183.2 people per km² (Israel Central Bureau of Statistics, 2009). The second city is Raanana, located at the middle ring of the central metropolitan area of Tel Aviv. Raanana’s population size is 70,000, and its urban density is 4046.5 people per km² (Israel Central Bureau of Statistics, 2009). Data for the research were collected during the years 2009-2010.

The sample was composed of 95 adults. Gender distribution of participants was 64% females and 36% males. Two groups of adults comprised the sample. One group, consisted of parents to students who were attending environmental schools (ENV; N = 67), and the other group, consisted of parents to students who did not attend environmental schools (N-ENV; N = 28), but rather attended regular schools located in the same types of neighborhoods with the same population characteristics as those of the ENV schools. The parents in the ENV group are not necessarily more environmentally oriented compared with the N-ENV group, because enrollment to the various schools is determined by the municipality on the basis of vicinity.

To control for demographic variables that could influence environmental attitudes and behaviors, we attempted to choose a homogeneous sample. For the purposes of the analyses, we used only questionnaires that were filled in by the participants who complied with the following criteria: (a) whose ages range from 30 to 50 years; (b) who have an academic education; (c) who are parents to children in the ages of 11 to 15 years; and (d) who are living in secular Jewish communities with a cluster membership of 8 to 9 in the socioeconomic index (Israel Central Bureau of Statistics, 2006). Parents were requested to fill in questionnaires containing a request for demographic data. We used these data to select suitable participants from the pool of respondents.

Data Collection From the ENV Group

Data collection from the ENV group was done by distributing questionnaires to parents of children in environmental schools, in the two cities. The questionnaires were handed out to students at the schools. Students were asked to deliver the questionnaires to their parents and return them completed. Most of the returned questionnaires were filled in by students’ mothers rather than by their fathers. We distributed 580 questionnaires to students; 157 questionnaires were filled in by the parents and returned; and 67 questionnaires complied with the above criteria and were used for analyses.

Data Collection From the N-ENV Group

Data collection for the N-ENV group was carried out through two methods as follows:

Emailing questionnaires to adults who complied with the above criteria and

Handing out questionnaires to parents as they were coming out of parents’ meetings at schools.

We distributed 150 questionnaires to parents, either personally or by email. Parents filled in and returned 87 questionnaires. Only 28 questionnaires were compatible with the above criteria and were used for analyses.

Questionnaire

The written questionnaire given to the participants was composed of two parts. The first part included a request for demographic details. The second part included three questions, as follows:

Question 1—Self-ranking of environmental attitudes and behavior: Participants were asked to rank on a Likert-type scale between 1 (very low) to 5 (very high) the following two statements: (a) “The degree of importance I attribute to environmental issues is . . .” and (b) “The degree to which I consider my behavior as environmentally friendly is . . .”

Question 2—Sources of influence on attitudes: Question 2 was phrased as follows: “Following is a list of possible sources of influence. Please rank the degree of influence of each source on the development of your attitudes regarding the environment.” Participants were asked to rank on a Likert-type scale ranging from 1 (no influence) to 5 (very influential) the following items: books; newspapers; television; internet; movies; the school in which my child studies; conversations, meetings, and interactions with people; the schools in which I have studied; certain teacher(s); my personality; a leader/a guide; a group activity; parents and family; my child(ren); time spent in nature or certain experiences that are related to nature; youth movements; academic studies or continuation studies; work/occupation; the community in which I live; political or institutional formats; constitutive experience; being a parent; and other.

Question 3—Sources of influence on behaviors: Question 3 was identical to Question 2 above, with one difference: Parents were asked to rank the sources of influence on their behaviors towards the environment.

Data Analyses, Results, and Conclusions

Participants’ Self-Ranking of Environmental Attitudes and Behaviors

We applied descriptive statistics to the data obtained in Question 1 (see above). Figure 1 presents the distribution of participants’ self-reported environmental attitudes and behaviors. Rankings of environmental attitudes were obtained from 46 participants. The rankings of attitudes ranged between low (1 case) to very high (18 cases). The majority ranked high (22 cases). Rankings of environmental behavior were obtained from 44 participants. The rankings of behaviors ranged between low (1 case) and very high (8 cases). The majority ranked high (24 cases).

Figure 1.

Distribution of participants’ self reported levels of proenvironmental attitudes and behaviors.

The results seem to be compatible with and reflective of the relatively high socioeconomic status of the participants and their high level of education. Previous research has established the connection between increased level of education and increased level of concern about the environment (Department of Environment, Climate Change and Water NSW, 2010).

Analysis of Categories of Influences on Attitudes and on Behavior, by Factor Analysis

To analyze formative influences on attitudes and formative influences on behavior, we applied two sets of factor analysis. The first set analyzed factors influencing attitudes (for data source, see Question 2 to participants, above), and the second set analyzed factors influencing behaviors (for data source, see Question 3 to participants, above). The analyses were applied to the total sample comprising the ENV and N-ENV groups pooled together (only valid cases were included; N = 65).

Factor analysis is a statistical test that explores which variables in a data set are most related to one another. In particular, it seeks to discover if the observed variables can be explained largely or entirely in terms of a much smaller number of variables called factors (Gorsuch, 1983). The variables that interact strongly with one another are grouped together to form factors, which are an a posteriori product of the analysis procedure.

Following are the results of the analyses, including the title of each factor, the variables it comprises, and the factor’s influence (presented as percentage of variance). Table 1 presents the analysis of influences on attitudes, and Table 2 presents the analysis of influences on behavior. A glance at the two tables reveals that there are qualitative (as expressed in the factors’ components) and quantitative (as expressed in percentage of variance) differences between factors that influence attitudes and factors that influence behavior.

Table 1.

Factors Influencing Attitudes, by Title, Variables, and Their Influence (% of Variance).

Factor number	Factor title	Variables	Influence (% of variance)
A1	My past and present close relationships and myself as a citizen	My child; my child’s school; political/institutional sources; my community; being a parent; parents and family	15.28
A2	Personal and interpersonal interactions with informal educational agents and with nature	Youth movement; nature; personality; a leader/a guide	13.83
A3	My early formal education	My teacher(s); schools in which I’ve studied	10.31
A4	The media—“fast intake” information sources	Internet; television; newspapers	9.42
A5	Formal, cognitive informative sources	Work/occupation; academic/continuation studies	9.39
A6	Informal, “medium to slow intake” informative sources	Movies; books; discussions/interactions with people	8.32
Total % of variance			66.55

Table 2.

Factors Influencing Behavior, by Title, Variables, and Their Influence (% of Variance).

Factor number	Factor title	Variables	Influence (% of variance)
B1	My early formal education, books and present citizenship and institutional formative sources	Schools in which I’ve studied; my teacher(s); books; political/institutional sources	14.48
B2	The media – “fast intake” formative sources	Newspapers; television; internet; movies	14.28
B3	My present intimate relationships and their related circles	My child; my child’s school; my community; being a parent	11.96
B4	My past close relationships, myself and interpersonal relationships	Discussions/interactions with people; personality; parents and family	11.00
B5	Formal, Cognitive informative sources	Work/occupation; academic/continuation studies	10.52
B6	Interpersonal interactions with informal educational agents and with nature	A leader/a guide; youth movement; nature	9.9
Total % of variance			72.14

The factor analysis reveals that six factors form categories of influences on attitudes, together explaining 66.5% of the variance. Factors A1, A2, and A6 (which together account for 37.43% of the total explained variance) represent a person’s informal, personal, and intimate relationships and experiences. Factors A3 to A5 (together accounting for 29.12% of the total explained variance) represent the more remote and formal circles that influence our lives—past and present formal education (Factors A3 and A5) and the media (Factor A4).

The factor analysis reveals that six factors form categories of influences on behavior, together explaining 72.14% of the variance. Factors B3, B4, and B6 (which together account for 32.86% of the total explained variance) represent a person’s close circles and intimate relationships in the present and in the past. Factors B1, B2, and B5 (together accounting for 39.28% of the total explained variance) represent formal formative influences and the media.

When comparing influences on attitudes to influences on behavior, the factor analyses reveal that with regard to influences on attitudes, a higher percentage of the variance (37.43%) is explained by informal formative experiences that are related to a person’s closest circles and intimate experiences and relationships, whereas with regard to influences on behavior, a higher percentage of the variance (39.28%) is explained by formal formative influences and the media.

The above results support the claim that different factors influence attitudes and behavior. Most of the influences on environmental attitudes incubate by slow processes that involve personal and interpersonal experiences, whereas most of the influences on behavior involve more formal experiences and rather fast intake processes (as expressed by the media).

Exploring Types of Participants by Cluster Analyses

To identify clusters of people who exhibit similarities in their reactions to experiences that can potentially influence their attitudes or their behavior, we applied a method of cluster analysis. The analysis was applied to the rankings of the ENV group (N = 67), with separate analyses for formative influences on attitudes and for formative influences on behaviors.

Cluster analysis is a method in which items are sorted into groups, such that the degree of association is strong between members of the same cluster and weak between members of different clusters (http://www.nature.com/).

In the present study, k-means clustering was used for creating two clusters. With this method, the number of clusters is predetermined. The k-means algorithm assigns each point to the cluster whose center is nearest. The center is the average of all the points in the cluster. Following are the results of the analyses.

Results of Cluster Analyses

Two clusters were obtained. Cluster 1 was termed the social type, and Cluster 2 was termed the private type. The “social” type is more influenced by variables related to interpersonal relationships, whereas the “private” type is more influenced by variables that are not directly related to social interactions, such as the media. The results for attitudes and behavior are presented in Figures 2 and 3, in accordance.

Figure 2.

Attitude clusters: Distribution of the means of the variables, by cluster.

Figure 3.

Behavior clusters: Distribution of the means of the variables, by cluster.

Within each pair of clusters, each cluster differs from the other quantitatively and qualitatively. The quantitative aspect is revealed by the fact that in Cluster 1, the mean scores are almost consistently higher than the mean scores of Cluster 2. That is to say, people who belong to Cluster 1 are generally more open to accepting influences on their attitudes and on their behaviors, compared with people who belong to Cluster 2. The qualitative differences between the clusters are expressed by the differences in the types of variables that have high levels of influence.

Cluster analysis of influences on attitudes

Following are the analysis results. Figure 2 reveals that with regard to influences on attitudes, among the “social” types (Cluster 1), the variable “personality” (mean ranking of 4.37) is most influential. Next, in degree of influence are the variables “my child” (4.34) and “nature” (4.26). Contrary to the “social” type, the “private” type (Cluster 2) is mostly influenced by “television” (3.85) and “newspapers” (3.85). The most extreme differences between the two clusters are in the effects of family and social influences—“being a parent,” “parents and family,” “my community,” and “youth movement.” The smallest differences are in the influences of “newspapers,” “television,” “academic/continuation studies,” and “work/occupation.”

Cluster analysis of influences on behavior

Following are the analysis results. Figure 3 reveals that with regard to influences on behavior, among the “social” types (Cluster 1), the variables “my child” (mean ranking of 4.35), “personality” (4.17), and “nature” (4.11) are most influential. Among the “private” types (Cluster 2), variables related to media—“newspapers” (3.93), “television” (4.13), and “internet” (3.67)—are most influential. The largest gaps between the two clusters are in influences of “my child’s school,” “being a parent,” and “academic/continuation studies.”

Analysis of distribution of participants between clusters

The following tables represent the distribution of participants between the clusters, separately for attitudes (Table 3) and for behavior (Table 4).

Table 3.

Distribution of Cases Between Clusters Regarding Influences on Attitudes.

Cluster	No. of cases	% of cases
1. Social type	38	59
2. Private type	26	41
Invalid cases	3	—
Total valid cases	64	100

Note: There is an 18% difference between the number of participants in Cluster 1 and the number of participants in Cluster 2.

Table 4.

Distribution of Cases Between Clusters Regarding Influences on Behavior.

Cluster	No. of cases	% of cases
1. Social type	48	72
2. Private type	15	28
Invalid cases	3	—
Total valid cases	64	100

Note: There is a 44% difference between the number of participants in Cluster 1 and the number of participants in Cluster 2.

Tables 3 and 4 reveal that in both attitudes and behaviors, substantially more participants belong to Cluster 1—the “social type.” The result implies that the majority of the sample is open and receptive to influences. The result is consistent with the results shown in Figure 1 regarding the participants’ self-ranking of environmental attitudes and behavior.

With regard to influence on behavior, the difference (44%) between the two clusters is substantially larger (2.4 times greater) than the difference between the two clusters of attitudes (18%). This result implies that the adults in the sample are more open to accepting influences on their behavior than they are to accepting influences on their attitudes. The Discussion section will elaborate on this finding.

Analysis of the Variable “Personality” With Regard to Its Interactions and Relative Impact

This section presents analyses that were carried out to answer the following questions: (a) With which other variables does the variable “personality” mostly interact? and (b) What is the relative impact of the variable “personality”?

Analysis of Interactions Between “Personality” and Other Variables

Interactions between “personality” and other variables were analyzed in the following two ways:

Analysis of results obtained by the above factor analyses; and

Analysis of correlations between the variable “personality” and the other variables.

Following are the results of the analyses.

Inspection of factor analyses results

Variables that belong to the same factor exhibit strong interactions with one another. With regard to influences on attitudes, the variable “personality” is included in Factor A2, “personal and interpersonal interactions with informal educational agents and with nature.” It is in strong interaction with the variables “youth movement,” “nature,” and “a leader/a guide.”

The variable “nature” has been found in previous research to be one of the strongest predictors of environmentalism. The strong interaction between “personality” and “nature” suggests that on its own, exposure to nature cannot predict development of environmental attitudes. For environmental attitudes to develop, there is a need for personality to mediate such formative experiences. Further research is required to elucidate the specific characteristics of “personality” with regard to the present context.

With regard to influences on behavior, the variable “personality” is included in Factor B4, “highly personal and interpersonal relationships.” It is in strong interaction with the variables “discussions/interactions with people” and “parents and family.”

Analyses of correlations

An alternative approach to factor analysis is measuring closeness between variables by measuring correlations. While factor analysis presents groups of interrelated variables, correlation measurements can expose discrete variables that are in significant correlation with the variable “personality” regardless of the interactions among the other variables.

The association between each pair of variables was measured by calculating Pearson correlation coefficients. The significance of the correlation was determined by applying single-tailed t tests. The outcome of the comparisons was a matrix of results of correlation tests, in which each variable was compared to another. The variables that were in significant correlation with the variable “personality” were singled out (Table 5). Table 5 presents correlations of “personality” with variables influencing attitudes and with variables influencing behavior.

Table 5.

Significant Correlations Between the Variable “Personality” and Other Variables, for Influences on Attitudes and for Influences on Behavior.

	Influences on attitudes			Influences on behavior
Variable	Pearson correlation coefficient with “personality”	Significance (one-tailed)	N	Pearson correlation coefficient with “personality”	Significance (one-tailed)	N
Community	0.557	0.000	61	0.321	0.06	62
Nature	0.509	0.000	61	0.369	0.002	61
Youth movement	0.502	0.000	61	0.333	0.005	60
Books	0.409	0.000	62	0.250	0.026	61
My child	0.340	0.004	60	0.306	0.008	62
My parents and family	0.316	0.007	60	0.348	0.003	61
Political/institutional sources	0.281	0.015	60	—	—	—
Being a parent	0.262	0.022	59	0.252	0.026	60
A leader/guide	0.257	0.022	62	0.222	0.044	60
School(s) in which I’ve studied	—	—	—	0.332	0.004	62
Discussions/interactions with people	—	—	—	0.459	0.000	62

Table 5 reveals that the variable “personality” interacts closely with most of the variables that were examined in the questionnaire. Eight variables were significantly correlated with personality in influencing both attitudes and behavior.

The following variables were not in significant interaction with “personality”: (a) “Internet,” “newspapers,” and “television”—variables that belong to the factor “the media”; (b) “work/occupation,” and “academic/continuation studies”—variables that belong to the factor “cognitive informative sources”; and (c) the variable “my teacher(s),” which belongs to the factor “my early formal education.”

Negative and significant correlation was found between the variables “personality” and “television” (r = −0.239*) P = 0.025 with regard to influences on attitudes.

The differences between interactions regarding influences on attitudes and influences on behaviors are as follows: The variable “political/institutional sources” interacts with “personality” in influencing attitudes and does not significantly interact with “personality” in influencing behavior; the variables “school(s) in which I’ve studied” and “discussions/interactions with people” interact with “personality” in influencing behavior and do not interact in influencing attitudes.

Analysis of the Relative Impact of the Variable “Personality”

The relative impact of “personality” was determined separately for each cluster. The clusters represent types of respondents, the “social” type and the “private” type. An analysis was carried out to determine to what degree “personality” influences each type of respondent, and within each type, to what degree it influences respondents’ attitudes and to what degree it influences their behavior.

Figures 2 and 3 present the mean rankings of each variable for each cluster. Of the 20 variables that were ranked, 6 variables that received the highest mean rankings by cluster and by influence aspect were singled out and depicted in a comparison table (see Table 6). Table 6 shows, for each cluster, the variables with the highest impact on attitudes and those with the highest impact on behavior. The relative importance (or impact) of personality is determined by the ranking of the variable among the six most influential variables.

Table 6.

The Six Most Influential Variables, by Cluster (“Social,” “Private”) and by Aspect of Influence (Attitudes, Behavior).

Cluster 1: Social				Cluster 2: Private
Attitudes (N = 38)		Behavior (N = 48)		Attitudes (N = 26)		Behavior (N = 15)
Variable	Mean ranking	Variable	Mean ranking	Variable	Mean ranking	Variable	Mean ranking
Personality	4.40	My child	4.40	Newspapers	3.40	Television	4.10
My child	4.30	Personality	4.20	Television	3.90	Newspapers	3.90
Nature	4.30	Nature	4.10	Nature	3.50	Internet	3.70
Internet	4.20	Being a parent	3.90	My child	3.40	Personality	3.50
Newspapers	4.03	My child’s school	3.77	Internet	3.20	My child	3.20
Television	4.03	My parents and family	3.75	Personality	3.20	Nature	3.10

Table 6 reveals that for both clusters, the variable “personality” has a substantial effect on both attitudes and behavior; it is among the six most influential variables. However, the impact of personality on “social” types is different from its impact on “private” types. Among the social types, personality is the most important variable influencing attitudes and the second most important variable influencing behavior. Among the “private” types, personality is still an important determinant, but it is sixth in importance for attitudes and fourth for behavior.

These results are consistent with the results of the previous analysis, which reveal that personality interacts with most other variables to produce an influencing effect, thus highlighting the importance of the inner self (personality) in organizing and giving meaning to all other formative experiences.

Analysis of the Relative Influence of Environmental Schools in Comparison to Nonenvironmental Schools

Environmental schools can influence parents’ attitudes and behavior by two primary means, as follows: (a) by directly influencing parents through activities that involve parents and through correspondence with them; and/or (b) by indirectly influencing parents through intergenerational influences of students who internalize their learning and pass it on to their parents. Therefore, when examining the impact of environmental schools in Israel, it is important to look both at direct and at indirect influences.

We applied the following two analyses to examine the influence of environmental schools in Israel:

Examining the relative effects of the variables “my child’s school” and “my child” in ENV schools, with regards to influences on attitudes and on behavior, by cluster. The analysis relies on the data presented in Table 6.

Comparison of rankings of the ENV group to the rankings of the N-ENV group to determine whether environmental schools in Israel make a significant difference compared with nonenvironmental schools.

Analysis of ENV Schools’ Direct and Indirect Influence

When examining Table 6, the following observations emerge, regarding the direct influences of ENV schools and the indirect influences through intergenerational influences by children:

Influences on attitudes.

The variable “my child’s school” does not appear as one of the six most influential variables on parents’ attitudes either in Cluster 1 (“social”) or in Cluster 2 (“private”). There is no indication of ENV schools’ direct influence on the children’s parents.

The variable “my child” appears among the six most influential variables in both Clusters 1 (“social”) and 2 (“private”). Therefore, there is an indication of intergenerational influence on attitudes, among parents of ENV schools.

Influences on behavior.

The variable “my child’s school” appears among the six most influential variables on the behavior of parents who belong to Cluster 1 (“social”), but not among those of parents who belong to Cluster 2 (“private”). ENV schools appear to be highly effective in exerting direct influence on the behavior of parents who belong to the “social” type.

The results imply that environmental schools in Israel are more successful at influencing parents’ behaviors than at influencing parents’ attitudes towards the environment. As for intergenerational influence, there is an equal high degree of influence on both attitudes and behavior.

The above findings raise the following question: In comparison to nonenvironmental schools, do environmental schools in Israel influence children’s parents differently? The following analysis refers to these differences.

Analysis of Differences Between ENV and N-ENV Schools

For the purposes of the present analysis, the study sample comprising the ENV group (N = 67) was compared with the control sample comprising the N-ENV group (N = 28).

Comparison of influences was done by comparing the mean rankings of the ENV group with the mean rankings of the N-ENV group and plotting them against each other. This was done for comparison of influences on attitudes (Figure 4) and for comparison of influences on behaviors (Figure 5). We tested for significant differences between groups, for the variables “my child’s school” and “my child.”

Figure 4.

Comparison of mean rankings for influences on attitudes of ENV and N-ENV groups, by influencing variables.

Figure 5.

Comparison of mean rankings of influences on behavior for the ENV group and the N-ENV group, by influencing variables.

A dark arrow in Figures 4 or 5 signifies a significant difference between the means of the corresponding variable. A dashed arrow signifies a borderline statistical significance between the means.

The results regarding influences on attitudes indicate that the differences between the ENV and N-ENV groups in the mean rankings of the variables “my child’s school” (direct influence) and “my child” (indirect, intergenerational influence) are statistically significant. Environmental schools in urban communities in Israel have a significantly greater influence on parents’ attitudes than do nonenvironmental schools.

Figure 5 presents the mean rankings for influences on behavior of the ENV and N-ENV groups, by variable.

The results regarding influences on behavior indicate that the difference between the ENV and N-ENV groups in the mean rankings of the variable “my child’s school” (direct influence) is significant, and the difference between the mean rankings of the variable “my child” (indirect, intergenerational influence) approaches statistical significance. This lack of significance can be attributed to the small sample size.

The compilation of results of the two analyses in the present section reveals that environmental schools in the two examined urban communities in Israel do affect parents’ attitudes and parents’ behavior, both through direct and through indirect formative influences. Yet with regard to the direct influence, ENV schools are more effective in influencing behavior than they are in influencing attitudes.

Analysis of the Extent of Interaction Between the Variables “My Child’s School” and “My Community”

For analyzing the extent of interaction between the variables “my child’s school” and “my community” among environmental schools, we calculated for the ENV and N-ENV groups, correlations between the variable “my child’s school” and other variables. We compared the results, by ENV versus N-ENV, and by attitudes versus behavior. Table 7 presents the results of the comparisons.

Table 7.

Significant Correlations of the Variable “My Child’s School” With Other Variables, by School Type and Influence Aspect.

	The variable “my child’s school”
	Attitudes		Behavior
School type	Variable	Pearson correlation coefficient	Variable	Pearson correlation coefficient
ENV	My child	0.568	My child	0.595
			My community	0.402
N-ENV	Schools in which I’ve studied	0.580	My teacher(s)	0.744
	My teacher(s)	0.431	Schools in which I’ve studied	0.666
	Political/institutional sources	0.451	Discussions/interactions with people	0.604

Table 7 reveals that environmental schools in Israel do interact with their communities to form a linked influential experience on parents’ behavior. With regard to influences on attitudes, the interactions between environmental schools and their communities are weaker. Among the nonenvironmental schools, there are no signs of any such interactions between schools and their communities. This implies that nonenvironmental schools in urban communities in Israel do not interact with their communities to form an interconnected influential experience on environmental attitudes or behavior.

Additional information that can be obtained by examining Table 7 is as follows: In ENV schools, close ties are created between the school, the child, and the community. In N-ENV schools the influence of the child’s school is more strongly associated with that of the parents’ past formal educational history and of external formal political/institutional sources. The intimacy and connectedness to the “here and now” that are observed among parents in the ENV schools are not observed among those in the N-ENV schools.

Discussion

The Discussion section addresses the following issues that emerge from the analyses: (a) the present study’s implications for previous SLE research; (b) the role of environmental schools as sources of environmental influence on their communities; and (c) differences between influences on attitudes and influences on behaviors.

Implications for Previous SLE Research

The present article refers to previous SLE research by putting forward the following claims:

The validity of previous SLE research that aims to expose influences on environmental stewardship is compromised by a lack of differentiation between types of influence processes. Processes of influence on attitudes are not the same as processes of influence on behavior. When the two types of processes are considered as one, the derived picture becomes blurred by a mixture of two types of influencing vectors.

When formative experiences with regard to environmental stewardship are analyzed as isolated entities, their explanation capacity and predictive value are reduced. Some useful information can be gained by analyzing factors of influence and focusing attention on interactions between variables.

The focus of attention on interaction between variables rather than on discrete variables implies that among different types of people different associations between variables are expected to be found.

The variable “personality” is a fundamental determinant in organizing and mediating most other variables. Before it becomes possible to gain predictive value from SLE research, much research is required to understand how these processes occur.

The above claims are supported by the following findings:

The present study reveals that different variables interact to produce different degrees of influence on participants’ attitudes and on their behaviors. The differences between influences on attitudes and influences on behavior are revealed in multiple analyses as follows: (i) factor analyses (see Tables 1 and 2) revealed that most of the influences on attitudes are derived from slow processing of interpersonal experiences, whereas most of the influences on behaviors are derived from more remote sources such as the workplace and the media; (ii) analysis of distribution of participants between clusters revealed that it is easier to influence adults’ behaviors than to influence adults’ attitudes (see Tables 3 and 4). This conclusion is based on the finding that the difference between the two clusters (in terms of the percentage of the sample) was 2.4 times greater for influences on behaviors compared with influences on attitudes; (iii) analysis of the six most influential variables, which was aimed at identifying the influence of the variable “my child’s school” on the parents (see Table 6), revealed that there are differences between schools’ influences on attitudes and on behaviors; and (iv) finally, analysis of significant correlations between variables related to schools’ influence (see Table 7) revealed that there are differences in interactions between variables, with regard to attitudes and with regard to behaviors.

The profound differences that were found suggest that in previous SLE research, which ignored the above differences, important information might have been masked by a lack of differentiation between influences on attitudes and influences on behaviors.

The present study focuses attention on the useful information that can be gained by analyzing interactions between variables rather than focusing on discrete variables. Analyses of interactions between variables revealed the following: (i) for each aspect of influence, six factors form categories of influences. The relative contribution of each category to the explained variance was calculated and presented (see Tables 1 and 2); (ii) analyses of correlations between variables provided insights into the important role of “personality” in organizing, constructing and giving meaning to most other influencing variables (see Table 5); (iii) the analysis of interactions between the variable representing the schools revealed that environmental schools in Israel interact with their urban communities to form an interconnected influential effect on children’s parents. Nonenvironmental schools showed no such interactions with their urban communities. The analysis of interactions also revealed that environmental schools, their communities, and the parents’ children form an association of influences that are intimate, relevant, and presently in progress, whereas in nonenvironmental schools, the sources of influences are in the parents’ past or derived from remote sources (see Table 7).

In the present study, two types of respondents were identified through a method of cluster analysis. The differences between the respondents were not merely quantitative but also qualitative. In a study conducted in Eastern Taiwan, Hsu (2009) found quantitative differences between environmental activists and those apathetic to environmental protection. The differences between the two groups were in the mean scores of the variables. The present study suggests that the differences between differing groups are also qualitative, as expressed by the different associations that can be found among variables. In the present study, the sample sizes were too small to enable factor analysis for each of the obtained clusters. Table 6 provides indication of these differences by presenting a comparison of the six most influential variables in each cluster. A. Gough (1999) drew attention to the importance of not treating the responses as coming from a homogeneous population. By using a combination of cluster analysis followed by factor analysis within each cluster, future studies will be able to yield insights into the unique ways in which the influential variables are associated with one another within each group of respondents.

Within the SLE debate, A. Gough (1999), N. Gough (1999), and Payne (1999) directed attention to the role of the inner self in constructing and giving meaning to SLE. The same experiences can be constructed in many ways and lead to an array of attitudes and behaviors towards the environment. The present research confirms the fundamental role of “personality” by applying a statistical method that confirms that “personality” is in significant interaction with most other variables and thus acts as a covariant that affects the ways in which other variables influence attitudes and behaviors.

The relative importance of the variable “personality” is even greater in Cluster 1 than in Cluster 2. Cluster 1 is composed of people who are more open to accepting influences regarding the environment. It is expected that the environmental activists and educators who constituted the samples of previous SLE research would be compatible with Cluster 1 of the present study, because they were chosen for those studies on the basis of their interest in environmental issues. As “personality” is revealed in the present study as such a crucial determinant of influences among the Cluster 1 types, it becomes even more important to understand the characteristics of “personality” if we wish to gain any real insight into the makeup of environmental activists and educators, or even to gain more in-depth understanding of how influence processes work and shape an environmental worldview.

The Role of Environmental Schools as Sources of Environmental Influence on Their Urban Communities

Chawla (2001) noted that results of SLE research suggest that formal education, and particularly primary school, is a weak source of formative environmental influence (Hsu, 2009). The present study indicates otherwise. Our study suggests that environmental schools in two urban communities in Israel are particularly successful in influencing the behaviors of parents who belong to Cluster 1. The difference that was found between environmental schools and nonenvironmental schools was significant. More research is required on a larger sample for validating the above results.

Differences Between Influences on Attitudes and Influences on Behavior

As discussed above and revealed by the study, influence is not a unified process. Our results support our proposition that attitudes and behaviors are influenced by different processes. The distinction made here between influences on attitudes and influences on behavior has implications for a wide range of issues in environmental education research and practice. If attitudes and behaviors are influenced by different processes, it follows that when designing and evaluating environmental programs, different strategies need to be employed for obtaining acquisition of environmental attitudes and for obtaining acquisition of environmental behaviors. More research is required to identify effective strategies for influencing each of these aspects. It also follows that when evaluating influences, a clear distinction needs to be made between evaluation of influences on attitudes and evaluation of influences on behaviors. Another implication of the above results is that the media and formal education would be expected to be more effective in influencing changes in behaviors, rather than in influencing changes in attitudes. Whereas influences on attitudes mature slowly through “slow-intake” experiences, influences on behavior are relatively easier to achieve through “medium-fast” experiences that can be operationalized in a preplanned manner by the media and institutions. Our results imply that the adults in the sample are more open to accepting influences on their behavior and less open to accepting influences on their attitudes.

The study challenges the traditional environmental education model from the 1980s, which held that acquisition of environmental behavior is based on acquisition of environmental attitudes as a prerequisite. According to the present study, it is easier to influence adults’ behaviors than to influence their attitudes. These results are not surprising when viewed in light of some well-known behavioral campaigns worldwide, such as campaigns to buy greener products. Intensive behavioral campaigns can gain demonstrable success in a relatively short time. Their success can be reinforced by law, regulations, or social pressure. However, attitudinal changes, as they emerge from the study, are more complex cognitive and affective processes that are slower to develop.

Implications of the Study for Future Research

The influence factors we obtained are an a posteriori statistical reflection of variables that interact with each other to create influences. We assume that these factors reflect the sample from which the data were gathered. The sample is composed mostly of females, in their midcareer, from middle to high socioeconomic status, well educated, and mostly with high awareness towards the environment. A reflection of the above characteristics can be found in the composition of the factors. For example, Factor A1 for attitudes—“my past and present close relationships and myself as a citizen”—reflects people who are firmly woven into their societies. They represent high interconnectedness with the immediate (their children and children’s schools) and remote social environment (political institutions). These characteristics are further reflected in the composition of the other factors. It is expected that when applying the same methods to samples of people from different backgrounds, different types of interactions between variables will arise. For example, in low-income and low-education urban areas, we might find that the media has a higher impact and that “political/institutional sources” will not be in interaction with “my child.” These considerations need to be taken into account in future research.

Footnotes

Declaration of Conflicting Interests

The authors declared no potential conflicts of interest with respect to the research, authorship, and/or publication of this article.

Funding

The authors received no financial support for the research, authorship, and/or publication of this article.

Author Biographies

Efrat Eilam is a researcher in the field of Education for Sustainability (EfS). Her research focuses on EfS evaluation and pedagogy, climate change education, and community influence processes.

Tamar Trop is a senior lecturer at the Faculty of Architecture and Town Planning at the Technion–Israel Institute of Technology. Her fields of research include environmental education, environmental policy and natural resources management.

References

Abelson

R. P.

(1972). Are attitudes necessary? In King

B. T.

McGinnies

(Eds.), Attitudes, conflicts, and social change (pp. 19-32). New York, NY: Academic Press.

Armstrong

Bottomley

(2003). Sustainable Schools Program: A one-stop shop for sustainability education. Paper presented at the Annual Conference of the North American Association for Environmental Education, Anchorage, AK. Retrieved from http://naaee.org/conferences/alaska/proceedings_2003.pdf

Arnold

H. E.

Cohen

F. G.

Warner

(2009). Youth and environmental action: perspectives of young environmental leaders on their formative influences. Journal of Environmental Education, 40(3), 27-36.

Ballantyne

Connel

Fien

(2006). Students as catalysts of environmental change: A framework for researching intergenerational influence through environmental education. Environmental Education Research, 12, 413-427.

Chawla

(1998). Significant life experiences revisited: A review of research on sources of environmental sensitivity. Environmental Education Research, 4, 369-382.

Chawla

(1999). Life paths into effective environmental action. Journal of Environmental Education, 31(1), 15-26.

Chawla

(2001). Significant life experiences revisited once again: Response to Vol. 5(4) “Five critical commentaries on significant life experience research in environmental education.” Environmental Education Research, 7, 451-461.

Corcoran

P. B.

(1999). Formative influences in the lives of environmental educators in the United States. Environmental Education Research, 5, 207-220.

Dillon

Kelsey

Duque-Aristizabal

A. M.

(1999). Identity and culture: Theorizing emergent environmentalism. Environmental Education Research, 5, 395-405.

10.

Doyle

J. K.

(1997). The cognitive psychology of systems thinking. System Dynamics Review, 13, 253-265.

11.

Gorsuch

R. L.

(1983). Factor analysis. Hillsdale, NJ: Erlbaum.

12.

Gough

(1999). Kids don’t like wearing the same jeans as their Mums and Dads: So whose “life” should be in significant life experiences research? Environmental Education Research, 5, 383-394.

13.

Gough

(1999). Surpassing our own histories: Autobiographical methods for environmental education research. Environmental Education Research, 5, 407-418.

14.

Hines

J. M.

Hungerford

H. R.

Tomera

A. N.

(1987). Analysis and synthesis of research on responsible environmental behavior: A meta-analysis. Journal of Environmental Education, 18(2), 1-8.

15.

Hopkins

McKeown

(2002). Education for sustainable development: An international perspective. In Tilbury

Stevenson

R. B.

Fien

Schreuder

(Eds.), Education and sustainability: Responding to the global challenge (pp. 13-24). Gland, Switzerland: IUCN, Commission on Education and Communication.

16.

Hsu

S. J.

(2009). The effects of an environmental education program on responsible environmental behavior and associated environmental literacy variables in Taiwanese college students. Journal of Environmental Education, 35, 37-48.

17.

Hungerford

H. R.

Volk

T. L.

(1990). Changing learner behavior through environmental education. Journal of Environmental Education, 21, 8-21.

18.

Israel Central Bureau of Statistics. (2009). Retrieved from http://www.cbs.gov.il/reader

19.

Myers

(1997). Significant life experiences and the choice of major among undergraduate minorities and nonminority students majoring in environmental studies and other disciplines. Paper presented at the Annual Conference of the North American Association for Environmental Education, University of British Columbia, Vancouver, Canada.

20.

NSW and the Department of Environment, Climate Change and Water NSW (2010). Who cares about the environment in 2009. A survey of NSW people’s environmental knowledge, attitudes and behaviours. Department of Environment, Climate Change and Water NSW. Sydney South 1232.

21.

Organisation for Economic Co-operation and Development. (2001). Schooling for tomorrow. What schools for the future? Paris, France: Author.

22.

Palmer

(1993). Development of concern for the environment and formative experiences of educators. Journal of Environmental Education, 24(3), 26-30.

23.

Palmer

Suggate

(1996). Influences and experiences affecting the pro-environmental behaviour of educators. Environmental Education Research, 2(1), 109-121.

24.

Palmer

J. A.

Suggate

Bajd

Hart

R. K. P.

Ofwono-Orecho

J. K. W.

. . .Van Staden

(1998). An overview of significant influences and formative experiences on the development of adults’ environmental awareness in nine countries. Environmental Education Research, 4, 445-464.

25.

Palmer

J. A.

Suggate

Robottom

Hart

(1999). Significant life experiences and formative influences on the development of adults’ environmental awareness in the UK, Australia and Canada. Environmental Education Research, 5(2), 181-200.

26.

Payne

(1999). The significance of experience in SLE research. Environmental Education Research, 5, 365-377.

27.

Peterson

(1982). Developmental variables affecting environmental sensitivity in professional environmental educators (Unpublished master’s thesis). Southern Illinois University, Carbondale, IL.

28.

Sward

L. L.

(1999). Significant life experiences affecting the environmental sensitivity of El Salvadoran environmental professionals. Environmental Education Research, 5(2), 201-206.

29.

Tanner

(1980). Significant life experiences: A new research area in environmental education. Journal of Environmental Education, 11(4), 20-24.

30.

United Nations Educational, Scientific and Cultural Organization. (2005). United Nations decade of education for sustainable development (2005-2014). International Implementation Scheme. Paris, France: UNESCO, Department for Education for Sustainable Development (ED/PEQ/ESD).

31.

Victorian Curriculum and Assessment Authority. (2005). Sustainability perspectives in the VELS. Retrieved from http://vels.vcaa.vic.edu.au/support/crosscurricular/sustainability.html#top

32.

Wallace

D. S.

Paulson

R. M.

Lord

C. G.

Bond

C. F.

(2005). Which behaviors do attitudes predict? Meta-analyzing the effects of social pressure and perceived difficulty. Review of General Psychology, 9, 214-227.