Self-perceived action competence for sustainability: the theoretical grounding and empirical validation of a novel research instrument

Abstract

This study contributes to an operationalization of the concept of action competence for sustainability through the theoretical development and empirical validation of a new 12-item Likert-scale questionnaire: the Self-Perceived Action Competence for Sustainability Questionnaire, SPACS-Q. Other scales in environmental and sustainability education (ESE) typically measure concepts such as pro-environmental and sustainability attitudes and behaviors, and therefore do not fully cover the concept of action competence for sustainability. An action differs from a ‘mere’ behavior in that it is voluntary and targeted at bringing about change, which is the overarching goal of ESE. We define action competence as a latent capacity among individuals to act sustainably. We introduce a novel scale measuring this seminal concept. Totally, 614 Swedish adolescents aged 12-19 participated in this study. The scale includes three latent subconstructs: i) knowledge of action possibilities, ii) confidence in one’s own influence, and iii) the willingness to act. Confirmatory factor analyses, reliability measures and investigation of convergent validity reveal a questionnaire instrument with excellent psychometric quality. We put forward that the SPACS-Q is a novel and theory-driven, empirically reliable and valid, instrument, and encourage fellow researchers to use the SPACS-Q when investigating people’s action competence for sustainability in various contexts.

Keywords:

• Action competence
• scale validation
• sustainability
• questionnaire development
• psychometric

Introduction

One of the most important challenges in today’s society is how to sustain our planet’s resources, and at the same time consider the socio-economic well-being of a growing population. This huge global challenge was defined in the concept of sustainable development (WCED 1987) and in the Global agenda, Agenda 2030 (UN 2015). At the very core of Agenda 2030 are 17 sustainable development goals, which cover the key systemic barriers for sustainable development (Le Blanc 2015; Vladimirova and Le Blanc 2016; UNESCO 2017). Environmental and sustainability education (ESE) has been developed as a key tool for teaching and learning how to deal with these global challenges (Scott 2009; UNESCO 2017; Leicht, Heiss, and Byun 2018). The underlying idea is that ESE empowers young people with action competence so they are equipped to tackle the wicked problems that characterize sustainable development (Mogensen and Schnack 2010; UNESCO 2017). Thus, in the long run, they would contribute to transforming the world into a more sustainable place (Lotz-Sisitka et al. 2015). The concept of action competence, which has a central place in this discourse, was defined in pivotal publications in the late nineties by Breiting and Mogensen (1999), and Jensen and Schnack (1997) as consisting of three key subconstructs i) knowledge of action possibilities, ii) confidence in one’s own influence and iii) the willingness to act. Competences for sustainable development have become an increasingly common way to describe the desired learning outcome of ESE, for example among young people, both in research (e.g. Cincera and Krajhanzl 2013; Liefländer et al. 2015; Olsson, Gericke, and Chang-Rundgren 2016; Goldman, Pe’er, and Yavetz 2017) and in global policy development (e.g. UNESCO 2017).

Studies that aim to gather empirical evidence for the impact of ESE on learning outcomes (e.g. Negev et al. 2008; Krnel and Naglic 2009; Boeve-de Pauw and Van Petegem 2011, 2018; Cincera and Krajhanzl 2013; Berglund, Gericke, and Chang Rundgren 2014; Boeve‐de Pauw et al. 2015; Liefländer et al. 2015; Olsson, Gericke, and Chang-Rundgren 2016; Goldman, Pe’er, and Yavetz 2017; ) typically use psychometric scales adapted to zoom in on specific predetermined outcomes such as pro-environmental and sustainability knowledge, attitudes or behaviors (see the section Scales in environmental and sustainability education research). Hence, the scales used in this field of research have so far failed to tap into the competencies for building information about action possibilities, developing confidence, and the willingness to act towards solving complex issues of sustainability. A scale, anchored in environmental psychology, that reports on people’s action competence for sustainability should therefore be highly relevant to the research field (Capelo, Santos, and Pedrosa 2012, p. 96). To our knowledge there are no such instruments yet with the psychometric focus or quality to tap into action competence for sustainability. In this study we develop and validate an instrument to measure this crucial outcome of ESE, and as such contribute a possible measure for implementation success in terms of people’s self-perceived action competence for sustainability.

Background

In this background section we will first introduce the concept of action competence and its three main interconnected subconstructs. Second, we will explain the concept by elaborating on the meaning of the sub-concepts action and competence, that build up the action competence concept, and connect them to environmental psychology theory. Finally, we will describe the possibilities to operationalize the concept of action competence and look into the presence of scales covering such environmental and sustainability learning outcomes in this background section.

The concept of action competence

Environmental and sustainability action competence was first described in health and environmental education research (Jensen and Schnack 1997; Breiting and Mogensen 1999; Fontes 2004; Chawla and Flanders Cushing 2007; Mogensen and Schnack 2010; Clark 2016). The concept of action competence has been defined to promote three main categories of interconnected outcomes: knowledge of action possibilities, confidence in onés own influence, and the willingness to act (Breiting and Mogensen 1999). Empowering present and future citizens with the ability to take action for sustainability is proposed as the ideal goal of education for sustainable development (UNESCO 2014; 2017). This means that the definition of action competence should align with how learning is described in ESE literature (Mogensen and Schnack 2010). In some cases, action competence has been understood as an educational approach (e.g. Ellis and Weekes 2008) and in other cases as an educational outcome in terms of intentions to act and competencies for sustainability among individuals and groups (e.g. Chawla and Flanders Cushing 2007; Clark 2016; Bonazzi Piasentin and Roberts 2018). In this paper we lean on the latter and consider action competence as a latent capacity among individuals and groups. However, the concepts action and competence have independent meanings, which we will first discuss before bringing them together in the action competence concept.

The first part, action, is central to the action competence concept (Jensen and Schnack 1997). An action differs from a ‘mere’ behavior in that it is voluntary and targeted at bringing about change or solving a controversial issue (Jensen 2000; Breiting et al. 2009). Typically, there is controversy on how to take action in order to manage the necessary change concerning such a controversial and complex issue or problem (Hungerford and Volk, 1990). An action therefore differs from a behavior in that it is decided on by the agents themselves, which also means that sustainability actions can cover individual actions as well as collective civic actions (Levy and Zint 2013). A collective action is similar to an individual action, but refers to the voluntary behavior of a group of people that is aimed at a common goal (Clark 2016). Sustainability problems or issues are characterized by different, and often contrasting, opinions on how to act towards finding a solution. Sustainability issues are also commonly viewed from a geographical perspective since the problems cover the local to the global sphere (Wals and Corcoran 2006; Wals 2015). Sometimes these kinds of complex sustainability issues are defined as wicked problems and are characterized by the inherent relational conflicts between humans, and between humans and nature (Brown, Harris, and Russel 2010). Furthermore, sustainability issues also cover the perspective of an uncertain future, that is, a future that we do not know much about (WCED 1987; Wals 2015). The focus of the learning outcome that would allow learners to deal with such problems, should therefore be on the competences needed for taking well-informed decisions for action, which is in contrast with the view of people showing the “right” environmental and sustainability behavior (Rudsberg and Öhman 2010).

The concept of competence frames the second part of the action competence for sustainability concept. As mentioned previously, the kind of competences needed to empower people with action competence for sustainability, consist of three main subconstructs which relate to i) knowledge of action possibilities, ii) confidence in one’s own influence and iii) the willingness to act (Breiting and Mogensen 1999). The description of action competence as comprising these three subconstructs is in line with the environmental theory of the reasonable person model (RPM) developed by Kaplan and Kaplan (2002, 2009). The theory of RPM is designed around the interrelationship between three informational needs among people i) the need for building mental models, ii) the need for being effective and iii) the need for meaningful actions (Kaplan and Kaplan 2009).

The first need is related to knowledge of action possibilities and involves strategies for future actions (Bandura 2005). It is about strengthening people’s competences to seek information and to build knowledge around the possibilities for using such mental models in actions. This first part of the RPM implies that knowing what is possible and knowing how things work is important (Kaplan and Kaplan 2009).

The second part of the RPM is about the empowerment of people, enhancing their feeling of confidence, competence and clear-headedness. This part implies that competence does not only depend on knowing what is needed and possible. Self-efficacy, that is, the feeling of being competent to deal with complex sustainability issues, is also important (Bandura 2001; Kaplan and Kaplan 2009). The need for being effective therefore corresponds to the second part (confidence in one’s own influence) of the sustainability action competence concept. The complex nature of environmental and sustainability problems induces the need for people to have the ability to reconsider their previous strategies, since new insights may contradict previous knowledge (Almers 2013). If previous strategies turn out to be inadequate, people may get into a state of mind characterized by lack of motivation (Pelletier et al. 1999). Consequently, having confidence that one’s own actions matter is an important component in action competence that may counter such feelings. This makes self-efficacy closely related to locus of control and researchers argue that these may have a common core (Judge et al. 2002). Self-efficacy differs from the latter, however, in that it focuses more on a person’s belief in their competences in specific situations, whereas locus of control concerns a person’s sense of control and general attribution of outcomes of a situation (Bandura 2001; Ajzen 2002). People with a strong level of self-efficacy are more likely to engage in pro-environmental and sustainability actions, and to persist when confronted with difficulties (Meinhold and Malkus 2005). Moreover, this self-efficacy and the confidence that one’s sustainability actions matter can be noted at the personal level as well as the societal and more collective level (Chawla and Flanders Cushing 2007; Levy and Zint 2013; Aguilar 2018).

Participation in actions is at the core of the third part of RPM, that is, the need for meaningful actions, which stems from people’s need and willingness to engage and make a difference (Kaplan and Kaplan 2009). In the action competence concept, which we here define as a latent capacity among individuals, the need for meaningful actions is described as the willingness to act for sustainability. Participating in meaningful actions requires that the individual actions that have an impact on the surrounding world, are accompanied by societal actions at the collective level (Clark 2016). Thus, it is important to frame the action competence concept both at the individual and at the collective level. Moreover, local and global environmental and sustainability citizenship can be developed if people engage in wicked problems. The latter include local and global perspectives, since people might feel that such problems lead to actions for sustainable development that are meaningful and relevant for them (Brown, Harris, and Russel 2010; Levy and Zint 2013). Given the third need, the need for meaningful actions (Kaplan and Kaplan 2009), it is important to frame the action competence concept at both the local and the global level. In previous ESE research, it has been found that the three informational needs of the RPM theory are especially salient in relation to young people and the adolescent age group (Olsson and Gericke 2016). Linking the concept of action competence for sustainability to the environmental psychological RPM is an important first step in the validation of its subconstructs.

Monitoring progress in ESE in terms of impact on learners’ action competence for sustainability cannot be empirically determined without a valid and reliable operationalization of the concept into a research instrument (Waltner, Rieß, and Mischo 2019).

Operationalization of action competence into a research instrument

In line with Mogensen and Schnack’s reasoning (2010) it would be impossible to operationalize a measure of what could be considered as a good or bad level of action competence in absolute measurements. Such a research design would imply that there could be a maximum level beyond which the level of action competence cannot be increased any further. Instead, we propose that action competence can be operationalized in a questionnaire instrument which could be used for relative measurements in comparative studies. Given that the action competence concept implies a conscious and intentional behavior that depends on self-reflectiveness and reactiveness (Bandura 2005; Mogensen and Schnack 2010) a measurement scale of action competence could be used to monitor changes in action competence among groups using research designs such as longitudinal studies, pre- and post-tests and cross-sectional studies (Waltner, Rieß, and Mischo 2019).

The three subconstructs of action competence outlined above could therefore be operationalized into a measurement scale of self-perceived action competence. The important link between the intention to act and real pro-environmental and -sustainability actions has previously been established in the ESE research field by, for example, Bamberg and Möser (2007). They found that self-perceived intention to act is the strongest mediator for actual personal environmental and sustainability actions. Accordingly, an instrument designed to measure people’s self-perceived action competence, would contribute to investigations of the effects of ESE implementation efforts. Our current study aims to contribute with such a reliable and valid instrument to the ESE research field: the Self-Perceived Action Competence for sustainability Questionnaire (SPACS-Q).

Scales in environmental and sustainability education research

Numerous scales have been developed to detect cognitive, affective and/or behavioral learning outcomes of ESE (e.g. Bogner and Wiseman 2006; Goldman, Yavetz, and Pe’er 2006; Dunlap 2008; Nisbet, Zelenski, and Murphy 2009; Biasutti and Frate 2017; Kaiser, Merten, and Wetzel 2018; Gericke et al. 2019). In our current study we will use one such scale to examine the convergent validity of our newly introduced SPACS-Q. Many scales are commonly used in the ESE research field and some of them have been widely spread, for example the two-major environmental values scale, 2-MEV (Bogner and Wiseman 2006), and the new ecological paradigm scale, NEP-scale (Dunlap 2008). The 2-MEV covers ecocentric and anthropocentric attitudes towards the environment, and the NEP-scale has been used as a measure of environmental concern in terms of attitudes and beliefs (Manoli et al. 2019). Since these two widely used scales do not cover the full scope of sustainable development, that is, also the social and economic dimensions along with the environmental dimension, other scales have recently been developed (e.g. Biasutti and Frate 2017; Gericke et al. 2019). The scale developed by Biasutti and Frate (2017) was designed to cover sustainability attitudes among university students. The concept of sustainability consciousness (SC) was developed to cover aspects of sustainability knowingness, attitudes and behavior in the environmental, social and economic dimensions of sustainable development (Gericke et al. 2019). The concept of SC has been defined as follows:

Sustainability consciousness (SC) refers then to the experience or awareness of sustainability phenomena. These include experiences and participations that we commonly associate with ourselves such as thoughts, feelings and actions (Gericke, Boeve-de Pauw, Berglund, and Olsson et al. 2019, 37).

The nature of the SC concept has the possibilities to reflect ESE learning outcomes from a broader perspective than the other previously mentioned scales, and could therefore be hypothesized to be closely related to the action competence concept. However, the SC questionnaire instrument and the other scales above lack the operational focus on action competences for sustainability. Instead they focus on ESE outcomes such as attitudes and pre-determined environmental and sustainability behavior.

There are also examples of scales that are more related to either environmental actions or sustainability competences (e.g. Pelletier et al. 1998; Cincera and Krajhanzl 2013; Alisat and Riemer 2015; Li and Monroe 2018; Waltner, Rieß, and Mischo 2019). The scale of Alisat and Riemer (2015) taps into specific individual and collective environmental actions, that is, to what degree the respondents have participated in such actions. ‘Personally wrote to or called a politician/government official about an environmental issue’ is an example of an individual item and ‘participated in a community event which focused on environmental awareness’ is an example of an item at the collective level in the Alisat and Reimer scale (2015, 19). Hence, the scale does not tap into sustainability competences. Cincera and Krajhanzl (2013) developed an instrument for investigations of the ecological footprint in terms of knowledge and skills in relation to water and energy consumption in households among young people. Even though the authors use the term ‘action competence’ as the measured outcome in their study, the scale and the investigation are not targeted towards the competencies and willingness to act, that characterize sustainability action competence as defined by Mogensen and Schnack (2010). The climate change hope scale for high school students was developed to cover three different aspects, where one of the aspects cover the willingness to take action (Li and Monroe 2018). However, Li and Monroe’s scale (2018) is developed to specifically cover the concept of hope and is therefore not fully applicable to the concept of action competence.

Based on the self-determination theory, Pelletier and colleagues (1998) developed the motivation toward the environment scale, MTES. The instrument deals with behavioral strategies and distinguishes between three main types of motivation that underlie environmental behavior: intrinsic motivation, extrinsic motivation, and amotivation. Intrinsically motivated behavior is the most autonomous. Here, the individual gets engaged for the pleasure of performing the behavior in itself. Extrinsic motivation relates to a variety of instrumental behaviors. Consequently, the individual is not interested in a certain behavior for its own sake, but wants to avoid negative consequences or bring about positive ones. Amotivation reflects a lack of motivation, that is, the individual does not see the sense of performing a certain behavior, or lack the competence to foresee the consequence of one’s behavior. An environmentally amotivated person is therefore unable to see the motives underlying a specific environmental behavior. The scale of Pelletier and colleagues (1998) does not deal with competences to act for sustainability explicitly, but rather explores the underlying reason for (not) acting. Still, its conceptualization of intrinsically motivated behavior for the environment does show likeness to the concept of action. However, like the other questionnaire instruments in the field, it lacks the operational focus on sustainability, and the subconstructs of action competence as described previously in the background section. In a recent publication Waltner, Rieß, and Mischo (2019) present the first steps in the development of a new instrument with the intention to measure students’ sustainability competences. According to the authors, the scale has so far been designed around sustainability related knowledge, beliefs, and behavior. Thus, the instrument draws from more or less the same components and items that build up the previously mentioned scale of SC and therefore fails to explicitly tap into the concept of action competence and its subconstructs. So far, reliability measures and minor validity analyses in terms of face validity have been published (Waltner, Rieß, and Mischo 2019).

The different questionnaire instruments outlined above are all adapted to cover specific outcomes of ESE and are well-functioning scales if they are connected to appropriate research questions and contexts. Only a few cover the broader scope of sustainability. However, according to our literature review no scale has yet been presented with the capacity to operationalize the concept of action competence for sustainability in terms of its three subconstructs as defined in environmental psychology theory: knowledge of action possibilities, confidence in one’s own influence and the willingness to act.

Research aim and questions

Given the definition of action competence and its connection to the environmental psychology theory of RPM, this study will contribute with a more coherent operationalization of the action competence concept relative to the ESE research field. The aim of this study is to contribute with a reliable and valid scale tapping into young people’s self-perceived action competence for sustainability. Our two research questions are:

• RQ1: What is the reliability and validity of the self-perceived action competence for sustainability questionnaire instrument, SPACS-Q?

• RQ2: In relation to an established instrument in the ESE research field (the sustainability consciousness questionnaire instrument), what is the convergent validity of the self-perceived action competence for sustainability questionnaire instrument, SPACS-Q?

Method

In this section we will elaborate on the development of the SPACS-Q and on the analytical approach for establishing the reliability and validity of the questionnaire instrument. Finally, we will describe the sample of participants for the present study.

The self-perceived action competence questionnaire

The items of the SPACS-Q were developed in early 2017. This was the starting point of a longitudinal research project aiming to investigate effects of whole school approaches to ESE in a Swedish municipality. The SPACS-Q was developed to monitor the students’ development of self-perceived action competence throughout the project. Hence, this study was conducted in a school context. The questionnaire instrument is built on the three main latent categories of interconnected outcomes of action competence connected to the environmental psychological theory of RPM: i) knowledge of action possibilities (KAP), ii) confidence in onés own influence (COI) and iii) the willingness to act (WTA) (Breiting and Mogensen 1999; Kaplan and Kaplan 2009). The theoretical model of the self-perceived action competence for sustainability (SPACS) concept is shown in Figure 1. The response option used in the questionnaire is a five-point Likert-type scale with answer options ‘strongly disagree’, ‘disagree’, ‘agree nor disagree’, ‘agree’ and ‘strongly agree’. The idea is to let people themselves report to which degree they agree to items covering the concept of action competence for sustainability. Hence, the SPACS-Q provides information on the respondents’ self-perceived action competence.

Figure 1. Theoretical model of the self-perceived action competence for sustainability (SPACS) concept with its sub-constructs knowledge of action possibilities (KAP), confidence in one’s own influence (COI) and willingness to act (WTA).

The first subconstruct of items, KAP, considers the part of action competence associated with having the knowledge of action possibilities and being able to deal with interrelated, complex and controversial problems (Breiting et al. 2009; Kaplan and Kaplan 2009). Gaining such knowledge should contribute to a clearer envisaging of future conditions. The second subconstruct of items, COI, considers aspects of self-efficacy and confidence in one’s own influence, that is, the need of being effective, the belief that a given action will produce the desired effect, and the notion that people will engage in actions if they are confident that their actions matter (Breiting et al. 2009; Kaplan and Kaplan 2009). The third subconstruct of items, WTA, deals with people’s willingness and commitment to take action, the willingness to act. It typically involves the voluntary type of motivation in alignment with the norms and values of the actors (Jensen and Schnack 1997; Kaplan and Kaplan 2009).

Given the three different parts of SPACS in Figure 1, the questionnaire instrument, SPACS-Q, is built around 12 items in total. There are four items in each of the three sub-scales KAP (covering knowledge of action possibilities), COI (covering confidence in one’s own influence) and WTA (covering the willingness to act) (see Table 1). We considered the final twelve items to cover the concept of self-perceived action competence. From a statistical point of view four items building up each latent subconstruct (factor) suffices (Field 2013; Muthén and Muthén 2015).

Table 1. The items and sub-scales of the self-perceived action competence for sustainability questionnaire, SPACS-Q.

Sub scale/Item
Knowledge of action possibilities
1	I can see different points of view on issues when people think differently.
2	I know how one should take action at school in order to contribute to sustainable development.
3	I know how one should take action at home in order to contribute to sustainable development.
4	I know how one should take action together with others in order to contribute to sustainable societal development.
Confidence in onés own influence
5	I believe I can influence global sustainable development through my actions.
6	I believe I can influence sustainable development in my community.
7	I believe I have good opportunities to participate in influencing our shared future.
8	I believe what each person does matters for sustainable development.
Willingness to act
9	I want to take action for sustainable development in my community.
10	I want to take action for global sustainable development.
11	I want to engage in changing society towards sustainable development.
12	I want schoolwork to be about how we can shape a sustainable future together.

Three of the researchers in our team had several meetings where we discussed meaning and phrasing of the items for each latent subconstruct. When developing the items, we considered the personal and collective level of sustainability action competence (Chawla and Flanders Cushing 2007; Clark 2016; Bonazzi Piasentin and Roberts 2018) as well as geographical, local and global, perspectives (Wals and Corcoran 2006; Wals 2015) in each of the three sub-scales. We did this to make sure that the items included the different facets of action competence for sustainability. Items related to the personal level of sustainability action competence are the item numbers 1, 2 and 3 of the KAP-scale, 5, 6 and 8 of the COI-scale and 9 and 10 of the WTA-scale (see Table 1). The first item actually deals with the competence to see the perspective of others, which is considered as an important competence for sustainable development (Rudsberg and Öhman 2010). Hence, the item reflects the individual knowledge to view things from many perspectives. Each of the three sub-scales have at least one item targeted at the collective level of sustainability action competence. In the KAP-scale item number 4, the COI-scale item number 5, and the WTA-scale items 11 and 12 include the collective level of sustainability action competence (Table 1). Moreover, we made sure that some of the items also cover the geographical, local and global, perspectives inherent in wicked sustainability problems issues (Brown, Harris, and Russel 2010). The local perspective is included in items 2, 3, 6, and 9 and the more societal and global perspectives are included in items 4, 5, 10, and 11 (see Table 1). Two of the items, 2 and 12, were specifically developed for the school context and are thus related to the fact that this study was conducted within the field of ESE.

To make sure that the questionnaire instrument covers all facets of the concept of action competence for sustainability the content validity of the items and the scales were discussed in an expert group, until consensus of the phrasing in Swedish was agreed upon (Hair et al. 2010). The expert group consisted of two senior researchers and three doctoral candidates from two universities (one university in Sweden and one university in Belgium), all of which with long experience in the ESE research field. Using back-translation (Brislin 1970), the items were translated into English by an English/Swedish language expert and then back translated to Swedish by another English/Swedish language expert to ensure a correct translation of the items. Only a few minor changes were made in this step to improve the language and to make sure the right meaning of the items was conveyed. The English version of the SPACS-Q and its subconstructs are shown in Table 1 and the original Swedish items are presented in the appendix.

Analytical approach

Confirmatory factor analysis (CFA) is a common starting point in the validation of a scale (Hine, Kormos, and Marks 2016). CFA investigates how well the data fits a theoretical model (in our case the theoretical model of SPACS). We used the statistical software package Mplus (Muthén and Muthén 2015) for the CFA. Since the data were categorical the weighted least squares mean and variance (WLSMV) estimator was used with delta parameterization. Since the participants were from different schools we corrected for the nested nature of the data (or hierarchical dependency of the errors, i.e. students were clustered within schools) through the complex command in Mplus (Muthén and Muthén 2015). Multiple fit indices were used to evaluate the model, with the recommended values of ∼.95 for the comparative fit index (CFI) and Tucker-Lewis index (TLI). Values ≤.06 were used for the root mean square error of approximation (RMSEA) (Tabachnick and Fidell 2007). Where necessary, modification indices based on meaningful error co-variances between items were used to further improve the model (Byrne 2012).

In the second step, the reliability of the sub-scales and the SPACS-Q as a whole were tested by calculating Cronbach’s alpha. To get a sense of the distribution of the answers we calculated descriptive statistics of each item in the next step (Field 2013). These calculations included the mean and standard deviation for each item as well as the percentage distribution of the response options for each item. IBM SPSS Statistics version 25 was used for the second and third step and the next, final, step in the analytical approach.

Table 2. Number of participants and girl:boy ratio.

Grade	6	7	8	9	10	11	12	Total
Number of students	85	98	106	95	96	78	56	614
Sexratio	0.95	0.71	1.02	1.16	0.86	0.39	0.61	0.81

In the fourth and final step of the analyses we further validated the SPACS-Q by examining its convergent validity, which means correlations with another related instruments (Mowday, Steers, and Porter 1979; AERA, APA and NCME 2014; Li and Monroe 2018). For this purpose, we used the short version of the sustainability consciousness questionnaire (SCQ-S) that was developed and validated by Gericke and colleagues (2019). The nature of the SC concept offers the possibilities to reflect ESE learning outcomes from a broad perspective, including sustainability knowingness, attitude and behavior. The cognitive part of the SC concept focuses on the knowingness of what is necessary for sustainable development, that is, similar to the component of action possibilities in the SPACS concept. The affective part of the SC concept focuses on people’s attitudes towards sustainability issues and the behavioral part taps into people’s behavior in relation to sustainability issues (Gericke et al. 2019). Even though there is a focus on cognitive, affective and behavioral components, the SCQ-S is focused on pro-sustainability attitudes and pre-determined sustainability behaviors. The SCQ-S therefore differs from the SPACS-Q in that it does not possess (nor does it aim to do so) the full capacity to operationalize the concept of action competence for sustainability into research in terms of all its three main constructs: KAP, COI and WTA. However, based on the theoretical background (outlined in the introduction section) and due to the broad perspectives inherent in the SCQ-S, the questionnaire instrument should be correlated with the SPACS-Q, but still, the two instruments should not fully converge into the same construct and measures. The SCQ-S should therefore be a good instrument to include in the convergent validity analysis by using Pearson’s correlation, which is a standardized measure of the strength of the relationship between the two scales (Field 2013). Investigations of construct validity for the SCQ-S using CFA (the same analytical approach as for the SPACS-Q) were performed previous to the convergent validity analysis. The final SCQ-S model of Gericke and Colleagues (2019) was used as a starting point for the CFA.

Table 4. Descriptive statistics for the SPACS-Q items.

Dimensions and items in the SPACS-Q		Likert scale answers					Estimates
		1	2	3	4	5	M	SD	Factor weight
KAP : Knowledge of Action Possibilities							3.80	0.84
1	I can see different points of view on issues when people think differently.	4.7%	3.7%	24.1%	30.3%	37.1%	3.91	1.09	0.685
2	I know how one should take action at school in order to contribute to sustainable development.	6.2%	8.1%	29.6%	32.7%	23.3%	3.59	1.12	0.783
3	I know how one should take action at home in order to contribute to sustainable development.	3.9%	3.4%	20.7%	37.6%	34.4%	3.95	1.02	0.826
4	I know how one should take action together with others in order to contribute to sustainable societal development.	4.1%	6.0%	28.8%	34.2%	26.9%	3.74	1.05	0.827
COI : Confidence in Own Influence							3.65	1.01
”I believe …”
5	… I can influence global sustainable development through my actions.	10.6%	8.3%	27.5%	26.2%	27.4%	3.51	1.27	0.867
6	… I can influence sustainable development in my community.	8.0%	8.3%	29.5%	27.0%	27.2%	3.57	1.20	0.891
7	… I have good opportunities to participate in influencing our shared future.	8.8%	9.9%	31.1%	24.4%	25.7%	3.48	1.22	0.797
8	… what each person does matters for sustainable development.	4.9%	3.6%	22.0%	22.0%	47.6%	4.04	1.13	0.754
WTA : Willingsness To Act							3.54	1.02
”I want …”
9	… to take action for sustainable development in my community.	7.3%	8.5%	30.3%	25.2%	28.7%	3.59	1.20	0.883
10	… to take action for global sustainable development.	6.4%	6.0%	27.2%	28.7%	31.8%	3.73	1.16	0.917
11	… to engage in changing society towards sustainable development.	8.3%	11.4%	33.4%	23.5%	23.5%	3.42	1.20	0.802
12	… schoolwork to be about how we can shape a sustainable future together.	10.4%	10.1%	31.6%	22.3%	25.6%	3.43	1.26	0.770

The sample

In collaboration with a Swedish urban municipality, 614 young people from four schools participated voluntarily in the study by filling in the SPACS-Q and the SCQ-S. Depending on how it fitted into their weekly schedule the teachers in each of the four schools chose 2-4 classes in each grade to participate. Students between the sixth and twelfth grade participated, which means that the age was between 12 and 19 years. The distribution of students in each grade and the sex-ratio (girls:boys) can be found in Table 2.

Ethical guidelines and regulations in Sweden associated with this type of research were followed (see Hermerén et al. 2011). About a week before the data collection, the schools received an information letter about the aim of the study and the data collection. Students from sixth to twelfth grade could be expected to decide on their participation when the questionnaire items concerned matters that are a natural part of school education (e.g. sustainable development). If the students did not want to participate they simply did not submit the online questionnaires.

Results

The results are presented in accordance with the previously described analytical approach. Hence, we will first present the result of the confirmatory factor analysis, followed by the reliability measures and the descriptive statistics for the items of the SPACS-Q. Finally, the results related to the convergence validity analysis connected to the second research question are presented.

Construct validity

Given that we were specifically aiming to confirm the hypothesized SPACS model, we used structural equation modelling (SEM) to perform confirmatory factor analysis. All 12 items of the SPACS-Q were inserted in the two-order model as shown in Figure 2. Based on the theory, the model consists of a second-order latent construct, SPACS, which is composed by three first-order latent constructs: KAP, COI and WTA. The fit indices showed the model to have excellent fit to our data with RMSEA=.058, CFI=.989 and TLI=.986. Therefore, no modification indices were used to improve the model any further.

Figure 2. The factor structure of the self-perceived action competence for sustainability (SPACS). KAP = knowledge of action possibilities, COI = confidence in one’s own influence and WTA = willingness to act.

Reliability results

We here report on reliability measures in terms of Cronbach’s α for the questionnaire instrument as a whole and for the three different sub-scales KAP, COI and WTA. The α-values in Table 3 indicate an instrument with good reliability in all its constructs (Field 2013). In the validation process we also used the short version of the sustainability consciousness questionnaire, the SCQ-S. This questionnaire instrument is also built up around three main components: sustainability knowingness, sustainability attitudes and sustainability behavior (see Gericke et al. 2019). The α-values of the SCQ-S and its sub-scales are also reported on in Table 3 and indicate an instrument with good reliability (Field 2013).

Table 3. Cronbach's α for the SPACS-Q and the SCQ-S and its sub-scales.

Scale	SPACS-Q				SCQ-S
	SPACS	KAP	COI	WTA	SC	K	A	B
Items	12	4	4	4	27	9	9	9
Cronbach’s α	.90	.80	.86	.87	.87	.82	.82	.78

Descriptive statistics of the self-perceived action competence scale

We calculated the percentage distribution of the answers for the response options of each item to illuminate how the Likert-scale for the SPACS-Q was used by the respondents. The results are shown in Table 4. In addition, we also calculated the mean value and standard deviation for each item in the SPACS-Q. The mean value is between the neutral middle Likert-scale option (3) and four for all items except for item number 8 where the mean value is slightly above four (4.04).

The factor weight for each item is also presented in Table 4. The factor weights are standardized and were estimated in the CFA. Hair et al. (2010) suggests a value above 0.55 to be relevant for observed indicators in CFA. Given this rule of thumb, the factor loadings presented in Table 4 indicate that all the items are important and contribute to the operationalization of the theoretical latent construct.

Convergent validity

In the final step of our analysis we used the SCQ-S as a comparison for further validation of the SPACS-Q to answer the second research question. Hence, we also performed CFA for the SCQ-S, since it was used for this comparison with our new SPACS-Q instrument. For this CFA we used the exact same model as Gericke and colleagues (2019) in their validation of the SCQ-S. The third order latent construct, SC, is composed by three second order latent constructs: sustainability knowingness, attitudes and behaviors. Each of these three latent constructs are composed by three first order latent variables consisting of the environmental, social and economic dimensions of sustainable development. The full theoretical model can be seen in Gericke et al. (2019). As in the original factor structure, we used two modification indices to improve the model. One included the correlation between two environmental behavior items and the other a correlation between two social behavior items (see Gericke et al. 2019, 43). The fit indices for SCQ-S showed the theoretical model to have good fit to our data with RMSEA=.025, CFI=.953 and TLI=.960. Therefore, no modification indices were used to improve the model further (other than the original ones).

The relation between the two questionnaire instruments SPACS-Q and SCQ-S were analyzed using Pearson’s correlation tests. A correlation coefficient above .3 indicates a meaningful correlation between two (sub)concepts (Field 2013). A very high correlation on the other hand, above .8, indicates that the instruments actually could be a measure of the same construct (Field 2013, 685-686). The correlation matrix is shown in Table 5, where Pearson’s r is shown above the diagonal. The part marked in grey show the correlations between the two questionnaire instruments and their sub-scales. The matrix shows that SPACS-Q and SCQ-S and their respective sub-scales are significantly correlated (all values above .3). The correlation coefficient, r, is between .380-.666, which means that even though the two instruments are related, they are not measuring the exact same construct. The highest correlation is between the two questionnaire instruments SPACS-Q and SCQ-S as a whole, r = .666 and between the SPACS-Q and sustainability behavior, r = .636. This means that less than half, 44.4% (r² = .444) of the SPACS-Q is predictable by the SCQ-S and 40,4% (r² = .404) of the SPAC-Q is predictable by sustainability behavior.

Table 5. Pearson's correlations (r) between the SPACS-Q and the SCQ-S

	KAP	COI	WTA	SPACS-Q	K	A	B	SCQ-S
KAP	1	.485*	.495*	.763*	.391*	.399*	.510*	.539*
COI		1	.656*	.867*	.380*	.435*	.501*	.521*
WTA			1	.872*	.450*	.486*	.585*	.611*
SPACS-Q				1	.489*	.529*	.636*	.666*
K					1	.610*	.444*	.833*
A						1	.398*	.807*
B							1	.796*
SCQ-S								1

Note: *Pearson’s correlation (r) is significant at the 0.01 level (2-tailed). The correlations between the two instruments and its subscales are marked in grey.

Discussion and conclusions

Based on the theoretical foundation described and the validation process of the SPACS-Q, we argue that the operationalization of a reliable and valid instrument has been provided in accordance with theoretical and psychometric standards. The instrument could therefore be of use for the ESE research community where monitoring development of peoplés action competence in terms of self-perceived action competence are of interest. Moreover, the instrument could also be of use in investigations of effects at the student level of environmental and sustainability education interventions, for example in terms of pre- and post-tests. However, in line with Mogensen and Schnack (2010), the instrument should not be used to determine what is good or bad action competence according to a predetermined absolute level. As discussed in the introduction, such an idea of measuring good or bad action competence implies that there is a maximum level beyond which it is not possible to enhance more action competence. However, the work of Bamberg and Möser (2007) showed that self-perceived actions are a strong mediator for real actions. Using the same reasoning for the concept of action competence, consisting of the subconstructs KAP, COI and WTA, our conclusion is that it is possible to monitor changes in action competence among people in terms of self-perceived action competence as was called for by Waltner, Reiß and Mischo (2019).

Our first research question was targeted at the measures of reliability and validity of the SPACS-Q. For the specific age group in this study, the reliability and validity measures in the results section suggest that the theory behind the SPACS is captured by the SPACS-Q instrument. The latent construct is built up by items covering the three subconstructs KAP, COI and WTA. These three constructs have been defined in previous research that describes the concept of action competence (Jensen and Schnack 1997: Breiting and Mogensen 1999; Mogensen and Schnack 2010). It is important to apply reliability measures of the different sub-scales (factors) if such sub-scales exist (Field 2013, 709-710). In our case we calculated Cronbach’s alpha for our three sub-scales KAP, COI, and WTA as well as for the questionnaire instrument as a whole (see Table 3). The α-values between .8 and .9 indicate a questionnaire instrument and sub-scales of high internal consistency well above the general rule of thumb of .7 for this kind of psychometric test (Field 2013, 709-710). The high α-value of the questionnaire instrument as a whole is what could be expected when the scale includes interconnected constructs (Field 2013). The descriptive statistics in Table 4 show that the respondents use the full scale of response options. Given the fit indices of the construct validity presented in the results section, the SPAQS-Q shows what could be categorized as excellent model fit (Tabachnick and Fidell 2007) and can thus be considered a valid instrument for measuring people’s self-perceived action competence.

Our second research question was targeted at measuring convergent validity. The question raised was to what extent the scales that formed the SCQ-S and the SPACS-Q converged? The SC concept (Gericke et al. 2019) is a concept that theoretically overlaps to some degree with the concept of action competence (see background section). In line with this theoretically driven presumption, our empirical results confirmed that the two operationalizing questionnaire instruments of the concepts of SC and SPACS are related. The correlations between the two instruments were above .3 for all sub-scales, which means that they are significantly correlated (see Table 5). Moreover, there was a high correlation (.636) between the SPACS-Q and sustainability behavior, which means that if a person scores high in the SPACS-Q it is likely that the person will also score high in the sub-scale of self-reported sustainability behavior in the SCQ-S. The highest correlation (.666) was found between the two questionnaire instruments as a whole. The implication for this is that if people score high in the SCQ-S (or in the aspect of sustainability behavior) it is likely that they will have a high score in the SPACS-Q. However, even though the two questionnaire instruments are related, they are not two different measures of the same concept (Field 2013, pp. 685-686). As mentioned in the results section, this means that less than half (44.4%) of the SPACS-Q is predictable by the SCQ-S. Consequently, we can conclude that the SPACS-Q and the SCQ-S are related, but do not converge into measuring the exact same construct (see Li and Monroe 2018). The SPACS-Q is designed to directly tap into the concept of action competence for sustainability in terms of the three subconstructs KAP, COI and WTA. The SCQ-S, on the other hand, is designed to measure sustainability outcomes in terms of knowledge, attitudes, and behaviors, and the SC concept has previously been reported to reflect aspects of action competence (e.g. Olsson et al. 2019), but the SCQ-S instrument measure other aspects of the concept. A theoretical explanation for this difference is that the SPACS-Q measures an individual’s latent capacity for voluntary and targeted action to bringing about change. Such latent capacity can be hypothesized to precede and be differentiated from an individual’s self-reported behavior, which in turn is a sub-concept measured for example in the SCQ-S. The SPACS-Q therefore fills a gap in ESE research by the operationalization of action competence for sustainability in a different way than scales measuring pro-sustainability behavior.

This study also shows some limitations. The SPACS-Q was used with one cohort of students in a Swedish context, which means that we do not know if the instrument would perform as well in other countries as it did in Sweden. The user of the SPACS-Q should therefore perform CFA to ensure construct validity for the SPACS-Q in their own research context, or investigate if the measures of the questionnaire instrument are invariant between different contexts (Hair et al. 2010). We used the instrument with students aged 13-19 (in Sweden this corresponds to the end of grade six till the end of grade twelve), which seems to be an ideal age group according to our results. We think that the use of the SPACS-Q with younger age groups should be done with caution. Too young students could have problems with interpreting the meaning of the items in a coherent way, which consequently risks the Cronbach’s alpha values of the SPACS-Q and the sub-scales to drop to a non-reliable level (Field 2013).

The data for the present study was collected in a school context. We, however, think that the questionnaire instrument could also be of use in other contexts, such as when monitoring the development of action competence for sustainability in organizations or the impact of policies on the action competence for sustainability of citizens in various contexts. There are only two items related to the specific school context, item number 2 (in the KAP-scale) and 12 (in the WTA-scale) (see Table 1 or Table 4), and these could be removed in studies outside the school context. Three items (indicators) is regarded as the lowest number of items needed to build up a latent factor (Hair et al. 2010). Hence, if items 2 and 12 would be removed from their respective sub-scales, enough items would remain to build up each latent subconstruct when using the SPACS-Q.

To summarize, given our analytical approach, the results and the limitations, we claim that the SPACS-Q so far stands the test of contributing a novel, theory-driven, empirically reliable and valid questionnaire instrument to the research field. The SPACS-Q puts specific attention to action competence for sustainability as an outcome of ESE, which is new in this field. Hence, the instrument is not focused on specific pro-environmental or sustainability behavior, which is covered by other scales. Instead, it measures a latent capacity to act before the action takes place. This claim is supported by our results, which provide empirical evidence that SPACS is a moderate predictor for self-reported behavior (see correlation between SPACS-Q and B in Table 5).

Implications

In Agenda 2030 the 17 sustainable development goals summarize the global plan for sustainable development (UN 2015). Education is highlighted as a key ingredient for achieving Agenda 2030. Furthermore, one of the goals (number 4) is the goal for reorienting education towards sustainable development in order to equip learners with competences to contribute to the transformation of society towards sustainable development (UNESCO 2017). It could therefore be meaningful to use SPACS-Q in the monitoring of reorientation processes and initiatives of current educational systems in terms of students’ self-perceived action competence for sustainability. Such evaluations could be an important piece in the process of the empowerment of young people to take action as active citizens in their society. We therefore invite and encourage fellow researchers to use the SPACS-Q in investigations that include monitoring the self-perceived action competence for sustainability concept among people.

Acknowledgements

We are grateful to Teresa Berglund at Karlstad university for sharing her expertise and for her important input during the development of this article.

Disclosure statement

No potential conflict of interest was reported by the author.

Additional information

Funding

This work was supported by The Swedish Institute for Educational Research [grant numbers: 2017-00065]

Notes on contributors

Daniel Olsson

Daniel Olsson is a PhD in biology at the Institution of Environmental and Life Sciences, Karlstad university Sweden. His research interest includes young people’s perceptions of environmental and sustainability issues and how education can be improved to empower action competence for sustainability among people.

Niklas Gericke

Niklas Gericke is Professor in Science Education, and Director of the SMEER (Science, Mathematics and Engineering Education Research) research Centre at Karlstad University in Sweden and visiting professor at NTNU in Trondheim, Norway. His main research interests are biology education and sustainability education from conceptual, teaching as well as implementation perspectives.

Wanda Sass

Wanda Sass is pursuing a doctoral degree in Training and Educational Sciences at the University of Antwerp, Belgium. Her research interests include motivation toward the environment, the concept of action competence, and action competence in sustainable development, focusing on early adolescents and their teachers.

Jelle Boeve-de Pauw

Jelle Boeve-de Pauw is a postdoc fellow and project leader at the University of Antwerp, faculty of Social Sciences, research unit Edubron, Belium. He is also head of the Expertise centre on Urban education at the Karel de Grote University College in Antwerp, Belgium. His main research interests are education for sustainable development, science education and urban education. 

Appendix

Swedish version of SPACS-Q

Du kan markera Dina svar på en skala från Instämmer inte alls till Instämmer helt. Om du varken instämmer eller inte instämmer markerar du mittenalternativet.

1. Jag kan sätta mig in i olika synsätt i frågor där människor tycker olika.

2. Jag vet hur man ska agera i skolan för att bidra till en hållbar utveckling.

3. Jag vet hur man ska agera hemma för att bidra till en hållbar utveckling.

4. Jag vet hur man ska agera tillsammans med andra för att bidra till en hållbar utveckling i samhället.

Jag tror att…

5. …jag kan påverka en global hållbar utveckling genom mina handlingar.

6. …jag kan påverka en hållbar utveckling i samhället där jag bor.

7. …jag tror att jag har stora möjligheter att vara med och påverka vår gemensamma framtid.

8. …vad var och en gör spelar roll för en hållbar utveckling.

Jag vill…

9. … agera för en hållbar utveckling i samhället där jag bor.

10. … agera för en global hållbar utveckling.

11. … engagera mig i arbetet med att förändra samhället för en mer hållbar utveckling.

12. … att skolarbetet skall handla om hur vi gemensamt kan forma en hållbar framtid.