1. Introduction
Human activity is universally recognized as being one of the major factors that contribute to climate change. Internationally, political concern and a willingness for changing human activities to stem climate change is increasing [
1]. Consequently, the importance of scientific research on the causation of environmental behavior of individual citizens can hardly be overestimated. Going beyond merely describing observed behavior, academic fields such as environmental psychology and environmental education have focused on what factors influence (intended) environmental behavior.
Hungerford and Volk [
2] reviewed research into how environmental behavior is shaped. They found no evidence for the assumption that more knowledge would alter attitudes or grow awareness of environmental problems, which would then motivate people to change their behavior. They concluded that this traditional simple linear model did not capture the complexity of behavioral change. They found that empowerment and ownership are critical to responsible behavior. Empowerment makes individuals feel capable of making a difference. People “own” environmental issues when they find them extremely important. This resonates with research into goals and motivation.
According to Sheldon and Elliot, the kinds of goals that people select affect the extent to which they are capable of maintaining efforts to attain them. The more people feel ownership when pursuing a goal reflecting their interests and values (i.e., in concordance with their self), the more they are likely to put sustained effort into achieving it [
3]. Deci and Ryan further explored how different regulatory processes that underlie the pursuit of such goals, relate to the quality of behavior. This makes their Self-Determination Theory (SDT) well-suited to explain (lack of) sustained effort [
4]. Before moving on to motivation toward the environment, we provide a brief review of SDT.
Self-Determination Theory
Deci and Ryan distinguish intrinsic and extrinsic goals. The former relates to basic need satisfaction and is perceived as more important to the individual per se, whereas the latter is associated with getting approval from others or external signs of worth (e.g., wealth, fame). Attainment of intrinsic goals is more strongly related to well-being than extrinsic aspirations, since these are less linked to basic need satisfaction [
4]. SDT further posits that goal-directed activities can differ in the level to which they are (not) self-determined. Two types of motivation can thus be ordered onto a continuum from autonomous to controlled, whereas amotivation is a third type that stands somewhat apart as it expresses a lack of motivation, be it self-determined or not [
5]. As
Figure 1 shows, amotivation lacks any kind of motivation, extrinsic or intrinsic. Amotivated individuals neither feel competent nor in control. In fact, there is no regulation at all and they feel helpless. Hungerford and Volk [
2] also found that the locus of control is an important variable involved in environmentally responsible behavior. When this locus of control is internal, the individual expects to be successful. External locus of control means that people expect they will not be able to make a difference [
2]. Extrinsic motivation, which is the least autonomous type of motivation, can vary in the degree to which it is autonomous [
4,
5]. In its most extreme form, people behave in a certain way in order to avoid a sanction or to get a reward (external regulation). If these externally sanctioned values become partially internalized, the individual engages in activities wanting to feel proud or avoiding feelings of guilt (introjected regulation). People may also find the result worthwhile (identified regulation), without engaging in activities for the pleasure of that behavior in itself [
4]. When values have become an integrated part of the self, the motivation is still extrinsic, since people triggered by integrated regulation do not act for the pleasure this provides in itself. Finally, intrinsic motivation occurs when individuals engage in activities purely out of interest. They feel competent and find pleasure in the activity per se. In fact, administering external rewards or threats can be deleterious to intrinsic motivation [
5].
According to Vansteenkiste, Lens, and Deci, intrinsic and identified motivation together can compose autonomous motivation, since people act out of free will. Similarly, introjected and external motivations make up controlled motivation [
6]. Introjected motivation is controlled from within the self (internally), through feelings of guilt or pride. Conversely, with external motivation, control is exercised through punishment and reward systems external to the individual [
7,
8]. In addition to autonomy and competence, relatedness also affects intrinsic motivation positively, albeit less powerfully. When people feel connected to others, loved and cared for, and allowed to love and care for others, they are more likely to maintain intrinsically motivated behaviors [
4].
Almost 20 years ago, Deci and Ryan’s self-determination theory or SDT [
5] was applied to motivation for environmental behavior [
9]. One of the perspectives was to develop and validate a questionnaire that also resulted in a measuring instrument, the Motivation Toward the Environment Scale (MTES). Pelletier et al. [
9] reported on four studies that contributed to the construction of the MTES. They asked why respondents were doing things for the environment. Samples were taken from populations of randomly selected citizens in the Cornwall area of the province of Ontario, Canada (study 2), and university students (studies 1, 3, and 4). Convergent and discriminant validity of the MTES was confirmed by Villacorta and colleagues [
10] within a population of Canadian college students of a university in a multicultural and multilingual city. They called for a validation within a population of students in a secondary education setting in order to further gain insight into the validity of the MTES in a population other than university students.
SDT offers insight into what factors may influence the quality of motivation [
4]. A higher quality of motivation might also lead to a better quality of (sustained) pro-environmental behavior. According to Chawla [
11], education becomes an important source of environmental commitment in the junior high through university years. She also found that the influence of family and experience of natural areas is associated more with childhood (i.e., up to 18 years of age). Starting the university years, friends become more important sources [
11]. Therefore, research into motivation toward the environment is required and especially relevant in adolescents as they will grow up to be the citizens and decision-makers of the future. Since it is these individuals’ motivation that will guide their behavior and the decisions they make concerning the environment, the focus of attention is drawn both toward a personal and a collective level.
To our knowledge, research in this field has uniquely made use of a variable-centered approach. Nevertheless, a person-centered approach may offer more possibilities for diagnosis of and distinction between motivational profiles, revealing the interplay of different motives within an individual. Consequently, the combination of both approaches may yield complementary information [
7]. Responding to the call for research within secondary school students outside Canada [
10], and the suggestion to add a person-centered approach [
7], this study will offer a combination of variable-centered and person-centered approaches to motivation toward the environment within secondary school students in Flanders, the Dutch-speaking Community of Belgium.
Canada and Flanders do not only differ geographically, with Flanders covering only 0.1% in surface (i.e., 13,522 km
2) compared to Canada (9,984,670 km
2). Both also show differences in the degree to which the environment is focused on at a societal level: where the Canadian administration protected 8.4% of its total surface in 2003 [
12], Flanders only arrived at a joint effort of 3.4% between government and registered private initiatives in 2011 [
13].
1.1. This Study
Answering Villacorta and colleagues’ [
10] call for a further validation of the MTES in a population other than college students in Canada, we aim to find out whether the MTES can reliably and validly be applied within secondary school students in Flanders. Furthermore, we will seek evidence for MTES profiles, using a person-centered approach as suggested by Vansteenkiste et al. [
7]. Both aims may forward theoretic insight, making the following questions central in this study:
- Research question 1:
Is the MTES a reliable and valid scale for measuring the different types of motivation (SDT) toward the environment within senior secondary students in Flanders?
- Research question 2:
Which MTES profiles toward the environment can be distinguished within senior secondary students in Flanders?
1.2. Literature Review
Since the different types of regulation in SDT lie along a continuum of self-determination, subscales are expected to form a simplex pattern in which scales that are adjacent show a stronger positive correlation than scales that are theoretically more distant, eventually correlating negatively. The three types of motivation, i.e., intrinsic, extrinsic, and amotivation, have been widely confirmed within various domains, such as academic behavior [
14,
15], and motivation toward the environment [
9,
16,
17]. Regardless of some breaches that were found in the simplex structure, the concepts of SDT have produced measuring scales. These scales, tapping into motivation for behavior, were deemed valid and reliable in various domains. One of these is environmental behavior.
1.2.1. Motivation Toward the Environment
Reviewing research into environmental behavior, Pelletier and colleagues concluded that correlates with environmental behavior such as attitudes and knowledge, had not succeeded in explaining why people failed to act pro-environmentally. Consequently, they proposed a motivational approach and more specifically suggested applying SDT to the environmental domain. The core question was why people do things for the environment. Their research included development of the MTES, an instrument for assessing people’s underlying motivation for environmental behavior [
9]. The MTES was further validated by Villacorta et al. [
10], and more recently also by Boeve-de Pauw and Van Petegem [
16].
Since SDT describes motivational types that show some presence of motivation in more detail, with five types ranging from completely autonomous to completely controlled, whereas lack of motivation is represented only in one type, i.e., amotivation, the latter merits more consideration. Moreover, Pelletier et al. [
17] noted that although people expressed stronger concern for the environment, they did not show more frequent, nor more difficult pro-environmental behavior. Consequently, they emphasized the need to look into individuals’ reasons for not engaging in pro-environmental behaviors. This led to the development of an instrument tapping into amotivation toward the environment, i.e., the Amotivation Toward the Environment Scale (AMTES). They further defined amotivation toward the environment, suggesting four kinds of beliefs might be underlying amotivation: global helplessness, strategy, capacity, and effort beliefs. When people are intimidated by scale and severity of the environmental situation, they may not act pro-environmentally because of global helplessness beliefs. They fail to see how their contribution could have a positive effect on such a large-scale problem. Individuals who expect certain strategies to be ineffective could be amotivated because of these strategy beliefs. They expect that their behavior would not produce the desired outcome. People doubting their abilities to perform certain pro-environmental behaviors, would be amotivated because of negative capacity beliefs. Consequently, even if a person believes a certain course of action to be effective, they may still be doubtful that they themselves are capable of producing the required behaviors. Finally, effort beliefs might also induce inactivity. Individuals may feel up to the activities and believe them eventually to produce positive results. However, they think they are incapable of producing and sustaining the required effort to engage in certain behaviors and integrating these into their lifestyle. Generally, Pelletier et al. [
17] found evidence for a preliminary instrument for tapping into people’s amotivation. In order to determine how people can be supported so they can overcome their amotivation toward the environment, insight into why people fail to do something for the environment is essential. This could inform environmental education and awareness campaigns.
The conceptualizations of both motivation and amotivation toward the environment were achieved using a variable-centered approach [
9,
17]. Such an approach focuses on the effects of motivational dimensions on people’s behavior and performance [
7]. Still, since individuals may show different types of motivation at the same time, a person-centered approach would enable one to categorize them in groups with members sharing a similar motivation profile. The insight gained from such research might offer complementary information and enable change agents to design motivational interventions to suit each particular profile [
7].
1.2.2. Motives and Profiling
Research into motivation is interested in both the various types of regulation and how these may be interrelated within individuals. Consequently, two approaches are possible, i.e., a variable-oriented and a person-oriented approach. Neither approach is better than the other, as they yield complementary information [
18]. Whereas the variable-oriented approach focuses on separate variables about individuals on average, the person-oriented approach allows one to address inter-individual as well as intra-individual differences [
19]. Vansteenkiste et al. [
7] applied this approach for studying motivation profiles for learning within high school and college students. Based on scores for autonomous and controlled motivations as described in SDT, they found four motivation profiles: a high quality (high autonomous, low controlled), a low quality (low autonomous, high controlled), a high quantity (high scores on both), and a low quantity profile (low scores on both). They confirmed the predicted most optimal learning patterns of the high quality group in comparison to the other three. In contrast to Vansteenkiste and colleagues [
7], Ratelle et al. [
20] also included amotivation when studying motivation profiles for learning within high school students. They found evidence for three motivation profiles. The first consisted of students scoring high on controlled and amotivation, but low on autonomous motivation. A second category showed high controlled and autonomous motivation and low amotivation. Finally, a third profile consisted of moderate levels of autonomous and controlled motivation, but low amotivation.
3. Results
779 students were asked to fill the questionnaire, that consisted of closed questions only. MTES items were scored on a 7-point Likert-scale. Unfortunately, only 578 fully filled questionnaires were returned. Some technical problems in the schools may have made it impossible for groups to finish the (online) questionnaire. The fact that entire groups returned incomplete answers seems to point in that direction. Some schools also had an extremely multi-lingual population. Language problems have probably made it unfeasible for some students to finish within the time limits. We also noticed a gradual increase in missing items towards the end of the questionnaire. This could be an indication of mental fatigue, especially for the students who indicated they did not speak Dutch at home. One item (MTES10) showed a higher number of missing cases (148 as compared to 131 to 140, gradually increasing towards the end of the questionnaire). In CFA analysis (see
Section 3.1.3) this item also emerged as problematic.
3.1. MTES Reliability and Validity
In order to answer our first research question, we verified the reliability and validity of a Dutch version of the MTES within senior secondary students in Flanders. Within a 95% confidence interval, the assumption of normality was deemed acceptable for all 24 items.
3.1.1. Reliability and Descriptives
Table A1, in the
Appendix A, provides a summary of means, standard deviations, and Cronbach’s alphas. Items were possible answers to the question: “Why are you doing things for the environment?” Respondents were asked to score their (dis)agreement with each answer on a 7-point Likert scale ranging from “does not correspond at all” to “corresponds exactly”. The six subscales represented the types of motivation as described in the SDT [
4]. Sumscores of items MTES1 to MTES4 provided the overall scores for intrinsic motivation, items MTES5 to MTES8 for integrated motivation, MTES9 to MTES12 for identified motivation, MTES13 to MTES16 for introjected motivation, MTES17 to MTES20 produced the sumscore for external regulation, and items MTES21 to MTES24 did for amotivation. All six subscales showed high reliability, with alphas ranging from 0.87 to 0.93. This means that the items that tapped into these types of motivation showed strong internal consistency between answers. Highest positive answers were given for amotivation (M
= 4.63, sd
= 1.41). Students indicated strongest agreement with the following statements: “I wonder why I’m doing anything about the environment, since the situation isn’t improving” (MTES21); “I feel that doing something for the environment is a waste of time” (MTES22); “I can’t see how my efforts to be environmentally friendly are helping the environment” (MTES23); and “I can’t see what’s in it for me” (MTES24). Respondents agreed least with answers indicating external regulation (M
= 3.09, sd
= 1.43). Statements in this type of motivation were: “Because other people would be mad if I didn’t do anything about the environment” (MTES17); “For the recognition I get for it from others” (MTES18); “Because my friends insist that I do” (MTES19); and “To avoid being criticized”. The second lowest score was observed with integrated motivation (M
= 3.71, sd
= 1.44), where “Because taking care of the environment is an integral part of my life” (MTES5) and “Because being environmentally conscious has become a fundamental part of who I am” (MTES8) both scored lowest. Looking more closely at identified motivation, item MTES10 (“Because it is the way I have chosen to contribute to the environment”) showed a much lower score (M
= 3.92, sd
= 1.54) than the other three answers (MTES 9: “Because it is a sensible thing to do something for the environment”; MTES11: “Because it is a reasonable thing to do something for the environment”; and MTES12: “Because I think it is a good idea to do something about the environment”). With the exception of amotivation (α = 0.87), identified regulation showed the lowest reliability (α = 0.89). We also noted that individual respondents’ answers differed greatly from each other, with standard deviations for each item and subscale all around or exceeding 1.4, ranging from 1.39 for identified motivation to 1.50 for introjected motivation.
3.1.2. Validity: Covariances and Simplex Pattern
As shown in
Table 2, when observing covariances between the latent factors in the CFA, the simplex pattern was largely confirmed, except for covariances of intrinsic motivation with integrated (0.80;
p = 0.00) and identified motivation (0.85;
p = 0.00). With exception of the negative covariance between latent factors identified motivation and amotivation (−0.08;
p = 0.085), all covariances and Pearson’s correlations were significant (
p < 0.001).
3.1.3. Validity: Confirmatory Factor Analyses (CFA)
Confirmatory factor analyses were performed for several models. We looked at modification indices to determine if the original 24-item model of the MTES could be improved. In the course of examining various models, modification indices repeatedly indicated item MTES10 (“Because it is a way I have chosen to contribute to the environment”) was problematic. Because identified motivation also broke up the simplex pattern, we were especially interested in a five-factor model without identified motivation. As shown in
Table 3, comparison of fit indices also pointed at a five-factor solution that included intrinsic, integrated, introjected, externally regulated, and amotivation. This model (see
Figure 2) consisted of twenty items, equally distributed over five subscales (i.e., intrinsic, integrated, introjected, externally regulated, and amotivation). Except for this model, all models showed disturbances in the simplex pattern with identified motivation breaking up the pattern.
3.2. MTES Profiles
For answering our second research question, sumscores were calculated for amotivation, intrinsic motivation, integrated, introjected, and external regulation. No univariate outliers were found, but eight multivariate outliers were removed. Criteria determining the optimal number of clusters were:
the similarity of cases within a specific cluster (i.e., homogeneity within groups),
discriminative value between clusters,
parsimony,
explanatory power of cluster solutions, and
a priori theorizing.
First, a hierarchical cluster analysis was performed in order to assess possible numbers of clusters at various stages of aggregation. These suggested a 7-, 6-, 4-, 3-, and 2-cluster solution. We further examined solutions with two to seven clusters. We then proceeded to assess homogeneity within clusters for various cluster solutions to determine what cluster solution showed optimal similarity within groups. In addition, parsimony was taken into account. A pseudo screeplot (
Figure 3) indicated a clear drop in reduced within-groups sum of squares between a five- and a six-cluster solution, which indicated that adding more than five clusters would not result in increased homogeneity within the groups. In sum, hierarchical cluster analysis pointed towards a three-, four-, or five-cluster solution.
In order to further determine an optimal number of clusters, we performed a k-means cluster analysis. We also examined the discriminant value between clusters and explanatory power for cluster solutions ranging from seven to two clusters, bearing in mind that solutions with five clusters would probably be the maximal number for reasons of parsimony. Post-hoc tests indicated that only in a two- and a four-cluster solution, all cluster pairs differed significantly.
Explanatory power was then assessed by calculating explained variance per MTES subscale. Explained variance ranged from 74.2% (introjected regulation) in a seven-cluster solution to 12.6% (amotivation) in a solution with two clusters. With the exception of a two- and a three-cluster solution, all solutions’ clusters explained at least 50% of variance of the motivational subscales, which was the threshold value. Consequently, four, five, six, and seven clusters all showed sufficient explanatory power.
Reviewing the criteria set to determine the optimal number of clusters, only a four-cluster solution showed positive results on all criteria. This solution had shown optimal within-group homogeneity with all four clusters also differing significantly, explaining between 55.4% (amotivation) and 65.3% (intrinsic motivation) of variance in motivation types.
Moreover, it was also considered parsimonious, since it was the most economical solution meeting the criteria of within-cluster similarity, between-cluster discriminant value, and explanatory power. As can be seen in
Table 4, the four-cluster solution revealed significant results with a first cluster containing 24% of all cases (
n = 132; 24.1% of male; 23.3% of female) scoring low on all motivation types, except amotivation. The second cluster represented 49% of students (
n = 268; 52.8% of male; 38.7% of female) with moderate scores on all five MTES subscales. Containing 18% of cases (
n = 99; 12.7% of male; 31.3% of female), cluster three showed high scores on all subscales except externally regulated motivation. Finally, cluster four was the smallest group, including 9% of students (
n = 51; 10.4% of male; 6.7% of female). This group scored high on all motivation types except amotivation. Similar proportions of both genders were found in the unmotivated group, whereas a larger proportion of female respondents (31.3%) were members of the inconsistently motivated group as compared to the proportion of male respondents (12.7%). The consistently motivated group was the smallest for both genders, albeit proportionally slightly more populated by men (10.4%) than women (6.7%).
Figure 4 visualizes a four-cluster solution.
Taking into account all criteria, our findings pointed to four clusters being an optimal solution. One group, which we call “consistently motivated”, showed high scores on all types of motivation except for amotivation. A second profile was found with moderate scores on all subscales, i.e., the group of “moderately motivated”. Cluster three, the “inconsistently motivated”, scored high on all types of motivation except for externally regulated motivation. Finally, a fourth group consisted of students showing low scores on all motivation types except amotivation. We called them “the unmotivated”.
5. Conclusions
In sum, our evidence suggested that a five-scale version of the Dutch MTES (including intrinsic, integrated, introjected, external, and amotivation) is both a reliable and valid instrument for measuring senior secondary school students’ motivation toward the environment in Flanders. Furthermore, four MTES profiles emerged: two smaller groups of consistently motivated and inconsistently motivated students, a large group of moderately motivated students (nearly one in two), and a relatively large group of unmotivated students (one in four). Low levels of amotivation were only found within consistently motivated individuals, who also scored high on external regulation.
Since evidence in this study indicated that senior secondary school students show distinct motivation profiles, environmental awareness and change programs can build on these findings. Such programs have the option either to differentiate approaches, catering for each separate profile, or to focus on promoting membership of the most desirable, i.e., the consistently motivated profile, which showed low levels of amotivation. Still, since such a vast portion of adolescents were amotivated, all programs are well-advised to provide opportunities for autonomy, relatedness, and competence [
4]. Additionally, change agents would do well providing their target audience not only with essential and relevant knowledge about climate change, but also with viable solutions for mitigation in terms of supporting their efficacy, strategy, and effort beliefs [
17]. Based on our findings concerning identified motivation, we suggest intervention programs focus on making the effect of actual behavior visible in a relatively short time. This would allow people to enjoy the fruits of their efforts, which may boost their effort beliefs so that they feel less helpless and more self-determined. Moreover, in order to promote a more intrinsic motivation, elements of fun should be introduced in pro-environmental activities. Thus, participants would experience that behaving pro-environmentally in itself provides joy and well-being.
Furthermore, we would like to call for inclusion of amotivation in research into motivational profiles, since a vast proportion of respondents (91%) showed moderate to high levels of amotivation, and even self-determined and internally controlled individuals showed high levels of amotivation, possibly preventing them from behaving pro-environmentally.
This study has contributed to a cross-cultural validation of the MTES, generally finding it a useful instrument for tapping into motivation toward the environment. Adding a person-oriented approach to a variable-oriented approach provided complementary information, showing that the inclusion of amotivation in a cluster analysis produces interesting information for environmental awareness and change programs.