Higher Education for Sustainability: A Critical Review of the Empirical Evidence 2013–2020

Abstract

Higher education for sustainable development (HESD) has grown into a substantial field of research and practice. HESD proposes that higher education will be central in a transition towards more sustainable socio-ecological systems. However, the debates on what should be learned in HESD and how this should be learned have remained conceptually controversial and empirically inconclusive. This review examined the evidence that specific pedagogies and content lead to specific “sustainability outcomes” among graduates. Three hundred and fifty-seven studies published between 2013 and 2020 were analyzed. The reviewed research was case-driven and often undertheorized regarding learning processes and outcomes. Despite its volume, the literature did not provide coherent insights into what should be learned and how. If the project of HESD is to be pursued further, more courage will be needed in creating novel forms of higher education, while more purpose and conceptual precision will be required in future research.

1. Introduction

Profound dysfunctions in our socio-ecological systems currently dominate public and academic discourse. The conclusions of reports on climate change have been alarming, and a previously unanticipated surge in warming is now projected [1]. Meanwhile, as we improve our understanding of the extent of ecosystem degradation, we realize that the ability of ecosystems to meet human dietary needs and provide other services grows increasingly dire [2,3]. At the same time, rapid advances in infotech have led to misgivings regarding its effects on civil liberties and the future of work [4]. Further, economic inequality has grown globally despite an overall rise in prosperity [5]. At the time of writing, the COVID pandemic has made interdependencies in our socio-ecological systems more tangible, but the experience of widespread unpredictability and complexity has been challenging for many. Most societies remain deeply divided about how the necessary transition towards more sustainable socio-ecological systems can be achieved.

However, many seem to agree that education (including higher education) plays a key role in shaping such systems, e.g., [6,7,8]. Indeed, the potential of higher education institutions (HEIs) as transformative agents for sustainability has been increasingly acknowledged in policy (e.g., the Sustainable Development Goals, SDGs) and academic discourse, emphasizing HEIs’ multiple functions (research, education, and societal leadership) [9]. This review focuses on the education function, i.e., the potential of HEIs and curriculum to enable learning towards a more sustainable future (higher education for sustainability/sustainable development, HESD).

In the instrumental narrative of HESD, the role of higher education is to produce innovative solutions and graduates equipped with skills to implement such solutions effectively [10]. By contrast, proponents of emancipatory higher education argue that HEIs are responsible for providing spaces in which students discover, create, and implement new ways of being, knowing, and relating [8]. More radical scholars dispute widespread acceptance of “higher education” and argue that attempts at sustainability reforms are trapped in a superficial improvement agenda [11]. Most concepts of HESD can be placed along a continuum from instrumental optimism to emancipatory concern and radical critique. The main paradigmatic divide is between scholars who assume that cognitive and affective outcomes relevant to sustainability should be taught and assessed, e.g., [12,13], and those who have conceptual and ethical concerns about whether higher education can and should tell students what to believe and do, e.g., [14]. Theorists of the latter perspective suggest that we should focus on fostering a critical disposition among students as a form of values education, which may or may not focus on sustainability [15]. Recent attempts have sought to bridge this gap pragmatically by integrating critical thinking or value competence as general competences underlying more specific sustainability competences [16].

Mirroring these tensions, the empirical research on HESD has struggled to demonstrate that specific curriculum experiences (including process and content) lead to specific sustainability outcomes. Recent reviews have contributed discussions on the use of transformative learning theory [17], mindfulness [18], and concepts of action competence in HESD design and assessment [19]. Evans and Ferreira [20] reviewed the effectiveness of pedagogical strategies in teacher education relevant to HESD.

However, the last systematic review of the empirical evidence that particular pedagogies and curriculum content may lead to specific “sustainability outcomes” dates from 2012 [21]. In their review, Barth and Rieckmann [21] diagnosed a lack of robust explanatory studies. Accordingly, I aimed to critically examine whether we have since made progress in generating evidence corresponding to the following research hypotheses (Figure 1):

Hypothesis 1:

Being exposed to specific curricula positively predicts specific “sustainability outcomes” among graduates.

Hypothesis 2:

Using specific pedagogies within curricula positively predicts specific “sustainability outcomes” among graduates.

Hypothesis 3:

Being exposed to or working with specific content in curricula positively predicts “sustainability outcomes” among graduates.

Hypothesis 4:

Confounding factors—affecting the relationships between curriculum, pedagogy, content, and sustainability outcomes—can be identified and appropriately controlled.

In order to address these hypotheses, it was also necessary to examine more closely how “sustainability outcomes”, i.e., sustainability and associated learning gains, were conceptualized in the reviewed studies.

The review is structured as follows: firstly, I outline the review methods. Then I characterize the identified literature, particularly the theoretical foundations, research design, and conceptualization of outcomes. The above hypotheses are then discussed before concluding with the implications of the review for the field of HESD.

2. Methods

This critical review considered qualitative, quantitative, and mixed-methods studies for review and summarized these studies using qualitative and quantitative approaches [22]. It applies procedures proposed by Petticrew and Roberts [23] and was guided by reviews in related fields [24,25,26,27,28,29].

2.1. Search Strategy

Title-abstract-keyword searches were carried out in the Scopus, Web of Science, and ERIC databases on 15 October 2020 for articles published from the beginning of 2013 until the day of the search, using the database-specific syntax of the text string: (higher education OR university OR universities OR tertiary education OR college) AND (education for sustainability OR education for sustainable development OR sustainability learning). The search yielded a combined raw sample of 8350 records, of which 6573 were retained for screening after duplicate removal. The CADIMA tool was used for merging reference lists, removing duplicates, and study screening [30].

2.2. Study Selection

The following inclusion criteria were applied, first at title/abstract-level and then at the full-text level: (1) the study population are students or graduates in higher education (college, university); (2) the study empirically investigates curricular processes and/or content as intervention/exposure; (3) the study empirically investigates outcomes that are claimed to explicitly relate to sustainability. The title/abstract screening excluded 5680 records, leaving 893 articles for full-text screening, which resulted in a total of 357 studies for systematic appraisal and further analysis (Figure 2).

2.3. Systematic Appraisal

For the systematic appraisal, I created a database (n = 357) of the bibliographic data and a full-text repository. I manually extracted and coded the relevant information to characterize the included studies and address the research hypotheses as indicated in

Table 1. Analytical Framework (categories, variables) and Coding Scales.

Category	Variable	Scale/Notes
Research interest and design	Research question(s)/hypotheses/objectives	0 = not specified; 1 = can be implied; 2 = explicitly stated; written summary
	Groups	1 = one group; 2 = two-group
	Repetitions	1 = post only; 2 = pre-post; 3 = repeated measures; 4 = longitudinal
	Research instrument	0 = not replicable; 1 = outlined; 2 = replicable and tested; 3 = established, written summary
	Susceptibility to bias	Sum of scores in groups, repetitions, research instrument: ≥7 low susceptibility; 4–6 moderate susceptibility; ≤3 rather susceptible
Population	Study continent/country	in case of international studies, all countries recorded
	Study sample	Characteristics (e.g., degree or course) recorded
	Sample size	Number of study participants
Intervention	Level of intervention	1 = sub-course; 2 = course level; 3 = program level
	Intervention/pedagogy	0 = not specified; 1 = outlined; 2 = described in detail
	Conceptualization of learning	0 = not specified; 1 = can be implied; 2 = explicitly described; 3 = critically reflected on; written summary, key reference where stated
Outcome	Conceptualization of sustainability	0 = undefined; 1 = mono-dimensional; 2 = multi-dimensional; written summary
	Conceptualization of intended effect	0 = not specified; 1 = can be implied; 2 = explicitly described; 3 = critically reflected on; written summary, key reference where stated
	Summary of findings	written summary

.

In empirical social research, making the research question(s), hypotheses, or objectives explicit is commonly considered important in evaluating the appropriateness of the research design [31]. The studies were coded accordingly (Table 1), together with the authors’ formulation of their research question(s), hypotheses, or objectives. The research design, namely groups, repetitions, and research instrument, were considered as indicators of methodological purposefulness. Under groups, I made a distinction between one-group designs and two-group designs (both randomized and non-randomized). Repetitions were coded on a scale ranging from posttest only to longitudinal. The research instrument, e.g., survey, scales, or interview guide, was briefly described and coded based on replicability and consideration given to validity issues. The scores in groups, repetitions, and research instrument were added up to approximate design robustness (unweighted).

To characterize the study population, the study continent and country were coded while allowing for multiple values in the case of international studies. The characteristics of the study sample were noted verbatim, and the sample size was recorded.

To trace the kind of intervention or exposure analyzed in the reviewed studies, the level of intervention was coded on a range from the student experience of specific activities within a course to an entire study program. Furthermore, the level of detail in describing the intervention/pedagogy was coded to support the interpretation of the study results. The way study authors conceptualized learning was recorded verbatim, including any key reference, and coded by level of explicitness and reflection.

Explanatory evidence on “sustainability outcomes” is at the heart of this review, so the way the authors conceptualized sustainability was of particular interest. Any study-specific definition of sustainability was noted, and if so, how this weighted the commonly recognized environmental, economic and social dimensions of sustainability [32]. The weighting was assessed by distributing a total of 3 points: a study considered to have a purely environmental concept of sustainability was scored as environment = 3: economic = 0: social = 0. A study claiming to put equal weight on all dimensions was scored as environment = 1: economic = 1: social = 1.

To capture what a given intervention was trying to achieve, the conceptualization of the intended effect was coded (Table 1), and a written summary was recorded. The investigated effect dimensions were then coded using an inductive approach for further analysis. When reporting the results, I refer to studies that exemplify specific aspects. The study results corresponding to the hypotheses of this review are summarized in tables. More detailed versions of these tables that include verbatim summaries are available as Supplementary Materials.

2.4. Positionality Statement and Limitations

The first exploratory iteration of this review was conducted to outline the state of the art in HESD as part of a project proposal. The refined full systematic review was independent of any project or institutional commitments. My interdisciplinary profile as a researcher and lecturer is probably typical to many of those involved in the field of HESD. With a background in social sciences in relation to natural resource management, my focus has shifted from agricultural innovation processes to learning in the context of sustainable development projects and higher education. Following this perspective, my understanding of research is open to diverse methodologies and appreciative of meaningful contributions from diverse empirical approaches.

This review comprised publications included in the considered databases, which may limit the scope of the review to the debate in academic English. While conducting this study as a single author may raise reliability concerns, both method and interpretation have been reviewed by colleagues in the field, and an earlier version has been presented and discussed at a conference on educational research. To make the process transparent and invite further debate, I describe the research strategy in detail and provide the list of included studies (Supplementary Materials S6).

3. Results

3.1. Study Characteristics and Research Designs

In general, this review shows that the HESD discourse continues to be concentrated in the Global North, particularly North America and Europe (Figure 3). Over a quarter of the studies analyzed originated in the USA, followed by studies from Spain at some distance. Between 2013 and 2016, publication activity on explanatory evidence in HESD was stable at around 30 publications a year, before increasing steadily to 67 publications in 2020. The reviewed studies were published in 130 different journals, books, or conference proceedings in diverse scientific fields. The International Journal of Sustainability in Higher Education, the Journal of Cleaner Production, and Environmental Education Research were the most prominent journals until 2017, but Sustainability has since rapidly become the leading journal (Figure 4).

The sample sizes suggest that most studies continue to be conducted as case studies at the course level: the median sample size was 79, and the distribution was strongly right-skewed (Figure 5). Thirteen studies did not specify the size of the sample. The dominance of case studies is evident also in the research designs, which mostly used one-group (83.8%) setups for posttest-only (51.0%) or pre-post (38.9%) observations. In two-thirds of all studies, authors only outlined how they had used either self-developed or substantially adapted research instruments (Figure 6). Only a few studies were based on repeated or longitudinal measures and applied replicable or established research instruments.

3.2. Intervention/Exposures and Concepts of Learning

Most of the interventions analyzed were done at a course or sub-course level (Figure 7). Studies investigating interventions at the sub-course level provided more detailed descriptions than studies at the course level, while studies investigating the effects of student exposure to the entire higher education program mostly only outlined the character of the exposure. Of all 357 studies, 104 (29.13%) did not problematize concepts and practice of learning in higher education. Another sub-group of 93 studies (26.1%) did not explicitly refer to a concept of learning in HESD, but comprehensively described what they considered a useful approach. Eleven of these studies investigated exposure at the program level, with constructivist and experiential elements featuring prominently in this group, e.g., [33]. Additionally, at the course and sub-course levels, experiential, project-based, and problem/inquiry-based approaches were dominant.

A third group of 135 (37.8%) studies explicitly referred to a learning theory or framework. Within this group, 33 (24.4%) were conducted at the sub-course level, 91 (67.4%) at the course level, and 11 (8.2%) at the program level. Table S1 (Supplementary Materials) lists these studies, categorized into theoretical families, subfamilies, and specific concepts. This analysis shows, at least for studies that clearly declare their theoretical basis, that HESD literature is firmly rooted in the constructivist paradigm. Further, the work of John Dewey and David Kolb has been most influential in shaping the way HESD scholars understand and approach learning. Accordingly, a range of specifications of experiential learning was very prominent in the sample: generic experiential learning, service learning, project-based learning, problem-based learning, and game-based learning. There are very few “anomalous” studies that do not fit into the constructivist family—for example, Mosher and Desrochers’ [34] behavior change training, Michel’s [35] cognitivist approach, and Walsh et al.’s [36] relational learning. Zidny and Eilks [37] explored multiple epistemologies (Indigenous and Western) in their concept of learning.

Finally, just 25 (7%) of all studies engaged in a deeper conversation with the HESD literature and problematized learning in higher education at a more fundamental level. A few studies in this category went beyond the constructivist mainstream and explored, for example, transgression [38], spirituality and consciousness [39], and identity development [40].

3.3. Specification of Outcomes

The ways in which sustainability and associated learning gains were conceptualized in the reviewed studies are summarized below.

3.3.1. Sustainability

The reviewed studies were assessed for their articulation of sustainability and the extent of consideration given to its three dimensions (Figure 8). In 21 (5.9%) of the studies, sustainability was not defined, and HESD was introduced as theoretically mature and backed by policy, e.g., [41]. The questions investigated were, in turn, how to implement HESD in practice, e.g., [42], what content to cover in HESD, e.g., [43], and how relevant competencies develop, e.g., [44].

A substantial share of the studies (20.7%) conceptualized HESD exclusively by reference to the environmental dimension. Some studies did so explicitly by investigating environmental attitudes, e.g., [45]. Piyapong [46] and Robinson [47] investigated the relationship between higher education and pro-environmental agency, measured by self-reported behavior. Several environmentally-oriented studies considered behavior change as an outcome of higher education—for example, Desrochers and Mosher [48] and Wisecup et al. [49] explored ways to make students save energy, Ting and Cheng [50] sought to enhance pro-environmental behavior through guided learning, and Filter et al., (2020), investigated the effect of experiencing nature.

A number of studies (2.5%, Figure 8) asserted that sustainability is multi-dimensional, only to concentrate on environmental aspects in learning content [51] or measured outcomes, e.g., [52]. If studies gave equal weight to sustainability’s social and environmental dimensions, they were placed halfway between the two dimensions (6.2%, Figure 8). Typically, authors argued that natural resource management is part of the socio-ecological fabric, suggesting that change requires learning across dimensions, e.g., [53]. Studies in this category tended to emphasize concepts such as social responsibility, e.g., [54], civic engagement, e.g., [55], and taking leadership, e.g., [56].

The majority of studies (52.4%) introduced sustainability compatibly with the multi-dimensionality of the SDGs. Some studies mentioned multi-dimensionality without critically reflecting on the concept, e.g., [57]. Many studies presented the SDGs as an endorsed framework and did not further problematize sustainability, e.g., [58]. Studies utilizing systems thinking defined sustainability as specific states of the systems of interest (e.g., agroecological systems [59], socio-technical systems [60]). Other studies emphasized epistemologies often considered characteristic of sustainability, namely inter- and transdisciplinarity, e.g., [61]. Few studies resonated clearly with the recent “ontological turn” in sociology: Burns [62] and Macintyre [38] described HESD as creating new ways of being with others and our environment, while similar approaches theoretically remained in the paradigm of transdisciplinarity, e.g., [63]. Rarely was the concept of sustainability reflected on more fundamentally—with the exception of Kopnina [64], who submitted that discussing sustainability can only be fruitful when embedded in questioning hegemonic political economies.

3.3.2. “Sustainability Knowledge”

In 87.4% of all studies, some form of “sustainability knowledge” was considered an outcome of HESD in the cognitive domain. The basic form of knowledge, recalling or knowing about [65], was most common in the reviewed studies, e.g., [66]. Some studies used the term “awareness” to describe basic knowledge, e.g., [67], and one study suggested that “knowledge and awareness about environmental problems” constitute “environmental knowledge” [68]. Others saw “awareness” as cutting across cognitive and affective domains—thus including being conscious about, sensitive to, and willing to attend to sustainability challenges, e.g., [69]. By reference to Bloom’s taxonomical hierarchy, some studies specified degrees of knowledge, such as Koch et al. [70], distinguishing between situational, conceptual, and procedural knowledge. Others integrated content knowledge and the ability to teach the content, e.g., [71].

To assess “sustainability knowledge”, some of the reviewed studies used direct measurements of student performance, for example, rubrics or simple grading [72]. Another strategy was to analyze the materials produced by students, but not as part of a course, e.g., [73]. Instead of measuring observable characteristics, many studies relied on stated or perceived knowledge: about the SDGs [74], climate change [46], and sustainability in planning [43].

3.3.3. “Sustainability Abilities and Skills”

About 60.5% of studies were explicitly interested in the effect of exposure to HESD on “sustainability abilities and skills”. One essential skill identified in the HESD literature is the ability to think of reality as complex systems, e.g., [75]. For example, some of the analyzed interventions asked students to, variously: model and evaluate systems, identify interactions and feedback loops, or practice decision-making in complex situations [76,77,78]. Changes in students’ systems thinking were documented using tests and rubrics, e.g., [79], or by assessing the complexity of students’ conceptual work, e.g., [80]. Hu et al. [81] adopted a novel approach, measuring brain activity during concept-mapping and processing of complex scenarios. Many authors also suggested that graduates should be able to solve sustainability problems, particularly in inter- and transdisciplinary settings, e.g., [82]. Again, this skill was investigated by direct measurements [83] or student reflections, e.g., [84]. Other studies focused on critical thinking, which was variously conceptualized as a higher-level cognitive ability, e.g., [85], the inclination to question internal and external perceptions, e.g., [86], or a reflexive way of being, e.g., [51]. While some authors thought of critical thinking mostly as an approach or pedagogical process, e.g., [87], others attempted to measure it as a reflection on a certain practice, e.g., [88]. Critical thinking was investigated indirectly by analyzing students’ reflections in journals [89] and directly by asking students about their perceived critical thinking gains, e.g., [90]. Another focus included intra- and interpersonal skills, which were described as: self-efficacy, e.g., [91], decision-making ability [92], communication [93], leadership skills, e.g., [58], or agency, e.g., [94]. Frequently, the boundaries between skills and affective dimensions were diffuse: Ruiz-Palmino et al. [95] explored decision-making and the ethics of students’ decisions; Yusop and Correia [96] proposed that HESD should create agency as well as “motivated and interested” graduates; Thomas et al. [97] considered intentions as indicators of leadership.

3.3.4. “Sustainability Attitudes and Values”

Of the reviewed studies, 46.2% explicitly referred to affective outcomes in their analysis. Most commonly, researchers investigated changes in students’ attitudes, with fewer considering broader concepts such as values. Attitudes were often conceptualized as part of an assumed knowledge-attitude-behavior continuum in relation to sustainability either broadly, e.g., [98], or with an environmental focus, e.g., [99]. Compared to other concepts in the HESD literature, attitudes seem to be more coherently theorized, based on concepts such as the Theory of Planned Behavior [100]. Since attitudes relate to specific objects or behaviors, researchers also seemed to find it easier to develop research instruments, often following the item-response paradigm. A number of studies relied on established scales, particularly in the environmental domain, such as the New Ecological Paradigm, e.g., [101] and the Environmental Attitudes Inventory, e.g., [102]. Others developed a research instrument as part of their study, using open and reflective approaches, e.g., [103], or scales targeted to specific aspects such as animals or water, e.g., [104]. Reflecting conceptual tensions in the HESD literature, some studies sought to demonstrate attitude improvements, e.g., [105], while others considered the existence of a positive environmental attitude a HESD competence [106]. When attempting to measure affective learning more broadly, the reviewed literature referred, for example, to the Theory of Basic Human Values, e.g., [107] and Values-Beliefs-Norms Theory, e.g., [108]. Several studies were vague in conceptualizing “sets of values” [109] or “perceptions” about aspects of sustainability, e.g., [110]. Like the sometimes-diffused boundaries between knowledge, skills, and affect, the boundaries between motivation or intention and behavior were not always clear, e.g., [111,112]. Moreover, many of the studies that used “sustainability competences” as a framework included affective aspects (see below).

3.3.5. “Sustainability Behavior”

Behavioral dimensions were explicitly addressed in 28.0% of all studies. Typically, the studies focused on “green” behavior as tantamount to “sustainable” behavior, e.g., [113]. Some studies assessed individuals’ environmental behavior in general, e.g., [114]. Others were more specific and homed in on consumption and waste, e.g., [115], or energy use, e.g., [116]. Studies investigating “green” behavior often attempted to use direct measurements (e.g., electricity meters). Broader concepts of “sustainability behavior”, including notions of collective action, were less common in the reviewed studies. Such studies looked, for example, into civil society activism and leadership in the workplace, e.g., [117]. Very few studies investigated behavior beyond the environmental dimension—for example, Hatipoglu et al. [118] included social equity and Borges [119] included gender equality as essential aspects of “sustainable” behavior. Studies based on broader concepts of sustainability relied mostly on reported behavior in their measurements.

3.3.6. “Sustainability Competences”

The idea of “sustainability competence” has grown prominent in the HESD discourse, loosely described as what students should know, be able, and sometimes be willing to do in relation to sustainability as a result of higher education. About 15.4% of the studies investigated such outcomes. The common usage of “competence” refers to “the ability to do something well” or “a skill that you need in a particular job or for a particular task” [120]. Some authors apply this meaning [121,122], while others use knowledge and skill elements of integrated competence frameworks, e.g., [123].

Many studies take an approach similar to Wiek et al. [13], and use “competences” as an umbrella concept to frame knowledge, skills, and attitudes they consider necessary for performing tasks, e.g., [41,124,125,126,127]. Several authors further defined and attempted to measure specific “positive behaviors” as a competence [44,57,109,128,129]. Some authors more carefully investigated the development or understanding of values as a competence, e.g., [130]. To assess changes in “competence”, most of the studies relied on self-evaluation or open reflection on perceived changes [131,132,133]. The use of closed perceptional questions may be considered particularly prone to social desirability biases, e.g., [127,134].

3.3.7. Sustainability: Other Outcomes

Considering that HESD is aiming to rethink higher education to strengthen its contribution to a necessary sustainability transition, the proposed outcomes of higher education were rather conservative: very few (7%) of the studies conceptualized outcomes beyond the categories elaborated above. These studies often referred to relational aspects, associated with particular ways of being. For example, Gedžūne [135] built her study on the concept of care and identification with nature, and Haigh [39] sought to establish a connection through tree planting. Another group of studies centered on social aspects of community and opportunities to cross institutional boundaries [38,53,136]. These studies used explorative and qualitative instruments, reflecting their open conceptual frameworks.

3.4. Summary of Empirical Evidence

3.4.1. Hypothesis 1: Being Exposed to Specific Curricula Positively Predicts Specific Sustainability Outcomes among Graduates

Of the 17 studies investigating the development of specific knowledge due to the completion of programs focusing on sustainability or environmental issues, 10 established a positive association (

Table 2. Does Exposure to specific curricula predict sustainability outcomes? Number of studies, direction of prediction, and susceptibility to bias.

Sustainability Outcome	Positive Prediction	No Prediction	Negative Prediction
Knowledge	◉ ● ○ ○ ○ ○ ○ ○ ○ ○	● ● ● ○ ○ ○ ○
(Cap) abilities, skills	◉ ◎ ○ ○ ○ ○ ○	● ○ ○
Attitudes, values	◉ ● ● ○ ○ ○ ○ ○	◉ ◉ ● ● ○ ○	● ● ●
Behavior	● ● ○	◉ ● ● ○ ○	◉
Competences		○ ○	○

Studies with: ◉ low susceptibility to bias, ≥2 groups; ◎ low susceptibility to bias, 1 group; ● moderate susceptibility to bias, ≥2 groups; ○ moderate susceptibility to bias, 1 group. Table S2 (Supplementary Materials) lists the studies and summarizes the results.

; Table S2, Supplementary Materials, summarizes study results). Others found that students’ sustainability knowledge did not substantially increase when it was already comprehensive at the beginning of a program, e.g., [70]. In the study by Amador et al. [137], there was no correlation between the intended learning objectives of a study program and the learning outcomes reported by students [137]. There was also a wide gap between the perceived sustainability knowledge and the actual knowledge ascertained by examination [138].

A similar picture emerged regarding the possible effects of higher education on skills and abilities. Some studies indicated that completing specific programs of study enhanced the ability to think systemically and critically, regardless of whether that was an intended learning goal of the program or not [92,139]. The study experience might also promote collaborative skills, e.g., [140]. Another study suggested that the higher education experience shifted the perception of sustainability from a matter of personal concern to a technical task [40].

The picture is much less clear at the level of values and attitudes. Some studies found that students attached greater importance to sustainability than the general population, which might relate to students having more universalistic values [141,142]. Such values may also increase throughout the study experience, e.g., [101]. However, many studies could not detect a clear association between sustainability attitudes and values and different fields of study or academic progress, e.g., [143,144]. Some studies showed that students in fact developed a more anthropocentric and instrumentalist perspective during their studies [145,146]. Students in the humanities might have stronger sustainability attitudes than students from programs directly concerned with environmental issues [147].

The evidence of a possible association of higher education and sustainability behavior is weak: while some studies found that students from environmentally focused programs reported more environmentally responsible behavior and activism, e.g., [117], others asserted that sustainability behavior did not change due to exposure to sustainability-related higher education, e.g., [148]. The year of study did not predict sustainability behavior [149], and environmental advocacy actually decreased throughout experiencing higher education, e.g., [55]. In another study, graduates, although knowledgeable, did not show sustainability leadership in the workplace [97].

Few studies have investigated a possible association between being exposed to specific curricula and developing sustainability competences. Trencher et al. [124] found that practice and research-oriented programs were more successful in developing such competences. However, Muñoz-Rodríguez et al. [150] suggested that such competences may decrease over the course of attending a program. Valderrama-Hernández et al. [139] found that reported changes in competences did not correspond to the intended outcomes of any course in the investigated study programs.

3.4.2. Hypothesis 2: Using Specific Pedagogies within Curricula Positively Predicts Specific Sustainability Outcomes among Graduates

Many studies investigated the possible effects of specific pedagogies on knowledge and abilities and skills (

Table 3. Does using specific pedagogies predict sustainability outcomes? Number of studies, direction of prediction, and susceptibility to bias.

Sustainability Outcome	Positive Prediction	No Prediction	Negative Prediction
Knowledge	◉ ● ● ● ● ● ● ● ● ● ● ○ ○ ○ ○ ○ ○ ○ ○ ○ ○ ○ ○ ○ ○ ○ ○ ○ ○ ○ ○ ○ ○ ○ ○ ○ ○ ○ ○ ○ ○	○
(Cap)abilities; skills	◉ ◉ ● ● ● ● ● ○ ○ ○ ○ ○ ○ ○ ○ ○ ○○ ○ ○ ○ ○ ○ ○ ○ ○ ○○ ○ ○ ○ ○ ○ ○ ○ ○ ○○ ○ ○ ○	● ○
Attitudes; values	◉ ● ● ● ● ● ● ○ ○ ○ ○ ○ ○ ○ ○ ○ ○ ○ ○ ○ ○ ○ ○	○
Behavior	◉ ○ ○ ○ ○ ○ ○	◉ ○
Competences	● ● ● ○ ○ ○ ○ ○ ○ ○	●
Other	○	○

Studies with: ◉ low susceptibility to bias, ≥2 groups; ◎ low susceptibility to bias, 1 group; ● moderate susceptibility to bias, ≥2 groups; ○ moderate susceptibility to bias, 1 group. Table S3 (Supplementary Materials) lists the studies and summarizes the results.

; Table S3, Supplementary Materials, summarizes study results). However, few studies purposively compared the effect of different pedagogies while keeping intended outcomes constant, e.g., [151]. Most studies were case-driven and sought to confirm the effect of a specific course or course element—and many studies did so successfully, e.g., [152]. Whether these pedagogies were more effective than other approaches could not be concluded from the reviewed studies. Furthermore, most investigated cases used pedagogies fitting into the family of experiential learning approaches. The question of whether research-based, project-based, or service-learning were not only effective but also reasonably efficient for faculty was not addressed. Finally, only very few authors reflected on the question of why a pedagogy would be specific to “sustainability knowledge” or “sustainability abilities and skills”, e.g., [35].

As discussed above, attitudes and values are of particular concern in the field of HESD. However, the explanatory evidence on the related effects of specific pedagogies remains vague. First, only a few studies were purposively designed to compare the effect of different approaches e.g., [153]. Several studies showed that holistic, multifaceted pedagogies did affect attitudes and values, potentially leading to clarification and shifts e.g., [94,107]. Others could not detect any effect of instructional practice on affective dimensions, e.g., [154], or found that “traditional” formats affected attitudes as well [155]. In sum, it seems that pedagogies inviting students to engage more deeply with a phenomenon, and particularly affected people, can lead to increased care and differentiated perspectives among participants. While we could argue that care and differentiated perspectives are relevant in shaping sustainability concerns, the literature did not establish why these attributes would not be equally relevant to other concerns. Educating students to act sustainably and to promote sustainability is a central motivation underlying HESD, yet the empirical evidence relating specific pedagogies to this endeavor is limited (Table 3 and Table S3). Several studies underlined the importance of crossing boundaries, for example, through studying abroad, e.g., [156] or taking learning outdoors or into communities [157,158].

The results regarding “sustainability competences” were similar: again, there was a lack of research keeping the intended outcome constant while varying the pedagogical approach. Several studies showed that engaging approaches like experiential and service-learning successfully build competences, e.g., [41]. While open learning journeys often characterized such approaches, competence development might in fact benefit from more structured approaches [159]. Overall, “sustainability competences” seemed challenging to capture meaningfully since several authors broke down the umbrella concept into different aspects or found that some aspects were not well covered, e.g., [160].

Very few studies did not fit in the categorization of outcomes above—Whiteside et al. [161], for example, showed that participating in a family wellbeing program did positively impact student wellbeing.

3.4.3. Hypothesis 3: Being Exposed to or Working with Specific Content within Curricula Positively Predicts Sustainability Outcomes among Graduates

We can confidently conclude from the reviewed literature that being exposed to or working with sustainability in course content builds corresponding knowledge (

Table 4. Does exposure to or working with specific content predict sustainability outcomes? Number of studies, direction of prediction, and susceptibility to bias.

Sustainability Outcome	Positive Prediction	No Prediction	Negative Prediction
Knowledge	● ● ○ ○ ○ ○ ○ ○ ○ ○ ○ ○ ○ ○ ○ ○ ○ ○ ○ ○ ○ ○ ○ ○	◎ ○
(Cap)abilities; skills	● ○ ○ ○	◎ ●
Attitudes; values	● ○ ○ ○ ○ ○ ○ ○ ○ ○ ○ ○ ○ ○	◎ ○ ○ ○ ○	● ○
Behavior	● ● ● ○ ○ ○ ○	◉ ○ ○ ○ ○ ○	◉
Competences	◎ ○

Studies with: >◉ low susceptibility to bias, ≥2 groups; ◎ low susceptibility to bias, 1 group; ● moderate susceptibility to bias, ≥2 groups; ○ moderate susceptibility to bias, 1 group. Table S4 (online only) lists the studies and summarizes the results.

, Table S4, Supplementary Materials, summarizes study results). However, the evidence of a possible relation between content exposure and relevant abilities and skills development is very limited.

The field of HESD has paid more attention to the possible effects of exposure on attitudes and values: several studies found a positive relationship, for example, regarding attitudes towards the environment [162]—these changes were sometimes only slight or short-term [163,164]. Other studies could not detect a change in attitude [60] or showed that students’ attitudes changed towards anthropocentrism throughout their studies [145]. A similar picture emerges regarding the possible effect of sustainability content exposure on behavior: some studies did report a change towards, for example, pro-environmental behavior [48]. Others could not confirm such a trend [165] or showed that studying majors usually associated with sustainability challenges (e.g., engineering) was a negative predictor of “sustainability behavior” [148]. Regarding competences, a positive effect was suggested (Table 4 and Table S4).

3.4.4. Hypothesis 4: Confounding Factors—Affecting the Relationships between Curriculum, Pedagogy, Content, and Sustainability Outcomes—Can Be Identified and Controlled

Based on the reviewed literature, we cannot assume that specific factors generally and consistently affect the relationship between curriculum content and processes (

Table 5. Do confounding factors affect sustainability outcomes? Number of studies, direction of prediction, and susceptibility to bias.

Sustainability Outcome	Positive Prediction	No Prediction	Negative Prediction
Knowledge	Being male ○	Educational level of parents ○ Pre-university focus of education ○ Gender ○	Being female ○
(Cap)abilities; skills	Privileged social group ◉	Gender ◉
Attitudes; values	Being female ◎ ● ● ○ Being older ● Regional origin ●	Educational level of parents ○ Pre-university focus of education ○ Gender ○ ○
Behavior	Being female ● ○ Regional origin ○	Educational level of parents ○ Pre-university focus of education ○
Competences	Being younger ○ Being female ○

Studies with: ◉ low susceptibility to bias, ≥2 groups; ◎ low susceptibility to bias, 1 group; ● moderate susceptibility to bias, ≥2 groups; ○ moderate susceptibility to bias, 1 group. Table S5 (online only) lists the studies and summarizes the results.

; Table S5, Supplementary Materials, summarizes study results). While some studies found significant effects of gender on knowledge, attitudes, and values in relation to learning and sustainability [100], others could not confirm this, e.g., [92]. The socio-economic background of students might have an effect, but parents’ educational level did not e.g., [128]. Meanwhile, the evidence on age effects is inconclusive, e.g., [163,166].

4. Discussion

The purpose of this review was to critically assess whether the field of HESD has made progress in generating evidence that specific pedagogies and content lead to specific sustainability outcomes. I begin the discussion by revisiting the conceptual and empirical patterns of what should be learned and how it should be learned in HESD according to the reviewed literature. I conclude by discussing the way HESD has been researched.

4.1. What Should Be Learned in HESD

By definition, learning gains related to sustainability are at the heart of the literature on HESD. Prominent voices in the discourse consider the “marriage (of education and sustainability) not only a necessity but a “win-win” for both fields” [167]. Indeed, this “marriage” is prominently reflected in SDG 4, and education is considered a catalyst for all the other SDGs [32]. The SDGs also defined sustainability to comprise social, economic, and environmental dimensions [168]. We might therefore expect that a scholarly field explicitly concerned with measuring educational gains related to sustainability would be particularly rigorous in defining the concept and its operationalization in research instruments. Yet, unfortunately, this review indicates that the empirical literature in HESD has largely adopted the “shallow consensus” on sustainability that Wals and Jickling [169] cautioned against.

In many studies, sustainability was treated as a fixed ontology that did not require explicit articulation or critical reflection. Others remained firmly rooted in environmental and ecological systems discourses. Very few studies proposed that sustainability might be a transient step in the continuous effort of reimagining ways of being in an uncertain world. Mostly, sustainability in the reviewed literature was not introduced as an educational prompt for epistemological, ontological, and ethical debate but as an educational end. Therefore, we might question whether sustainability as referred to in HESD can create sufficient dissent and alterity in higher education to encourage the educational “anomalies” and experimentation necessary to transform higher education [170]. This review provides good reasons to be doubtful.

The least controversial learning outcome associated with sustainability was knowledge, and most studies agreed that conditions like awareness or understanding are prerequisites of other sustainability learning gains. Higher education focusing on sustainability from sub-course to curriculum level seems to successfully build knowledge, although this relationship becomes less clear the broader the scope of a study.

The HESD literature also seems to agree that certain abilities and skills should be particularly associated with sustainability, such as systems thinking, critical thinking, problem-solving, and intra- and interpersonal skills. They are arguably so relevant to sustainability concerns that they should be called “sustainability abilities and skills”. I assume, however, that most educators regardless of their concern with sustainability would aim to enable students to think critically, for example. Accordingly, it seems more sensible to consider these abilities and skills outcomes of meaningful higher education in general. Further, it is surprising that HESD pays relatively little attention to abilities and skills that would empower students to turn their critical or systemic thinking into disruptive actions and strategic societal change if so desired. In that respect, the content of HESD was rather well-behaved.

The HESD discourse appears to have placed greater emphasis on affective learning outcomes such as changes in attitudes and values. Indeed, following the argument of Shephard [171], the central attribute of HESD could be considered the pursuit of affective learning outcomes. Accordingly, the reviewed literature often investigated (environmental) attitudes typically defined as an affect directed at a specific object or behavior [172]. To understand affective learning more broadly, the reviewed studies investigated values, meaning beliefs about desirable end states or modes of conduct transcending a specific situation [173]. Studies addressing values were particularly vague regarding conceptualization and operationalization, however—and I believe that at this point the “HESD elephant” entered the room: educators want to see specific sustainability attitudes, values, and behavior as educational outcomes, but would not openly teach or train for them. The currently dominant response to this contradiction is to combine “cognitive, affective, volitional and motivational elements” [16] into the collective framework of “sustainability competences”. Competence can be observed only as performance. When we attempt to assess affective performance, we necessarily need to predefine our expectations to this performance. This has raised concerns that higher education should not “tell students what they should be willing to do”, but educate their graduates to think “deeply, independently and critically” [174]. Empirically, the reviewed literature struggled to meaningfully assess competence gains, relying on either limited assessment of selected competences or student self-assessment of perceived gains. When combined with broad concepts of sustainability, dealing with different aspects of competences collectively results in widespread ambiguity and a large number of vaguely defined dimensions. Although a “shallow consensus” is created, the controversy and debate necessary for rethinking higher education are effectively limited.

In the face of these unresolved issues, I suggest that HESD research and practice should focus on greater conceptual clarity and more realistic, humble expectations of what higher education is currently able to achieve. This, I believe, would provide a firmer foundation for rethinking learning and advancing the field of HESD, rather than further conflating different learning gains in seemingly holistic concepts that mainly serve to stifle meaningful debate.

4.2. How It Should Be Learned

Given that the reviewed literature sought to establish explanatory evidence on the association of specific interventions or exposures with specific learning gains, we might expect that the studies engaged in a lively debate questioning established learning concepts and practices in higher education. This seems essential to any transformative re-thinking of higher education, whether within current institutional settings or beyond. After all, the common diagnosis of the reviewed studies was that higher education had to evolve if it was to contribute to sustainable development more meaningfully.

However, the review showed that many cases undertheorized learning in the design and reflection of interventions or exposures in HESD. Besides, and similar to the learning content discussed earlier, the pedagogical approaches were “well-behaved”. Rather than reflect on assumptions underlying higher education and possible “hidden curricula” more fundamentally, the cases looked for ways to upgrade established learning formats mostly with experiential elements. The anomalies from this general trend in HESD were few, and the reviewed research was notably apolitical. On a more positive note, the large number of documented cases underlines how deeply many lecturers in higher education care about their students’ educational experience. In this educational work, confounding factors (e.g., gender, socio-economic background) do not seem to play a consistent and significant role. We may conclude that meaningful learning is a universal experience that cannot be predicted by factors understood from an essentialist perspective.

The reviewed literature struggled to establish that specific pedagogies are more effective than others in achieving specific learning gains since interventions were rarely varied while keeping the intended outcome constant. In general, the literature provided evidence that higher education effectively builds knowledge and a range of skills. Such effects were not equally evident for affective and behavioral aspects and “competences”, which seem less accessible to conventional methods of measuring. We could also conclude that higher education in its current form is not performing well in enabling such learning gains. A fundamental question was barely addressed in the reviewed literature: why should a specific pedagogy be considered particularly or exclusively relevant to learning in the context of sustainability, in contrast to other complex societal concerns?

In conclusion, the findings on how students should learn in HESD offer a number of ways forward for the field. Firstly, higher education could focus on what it appears to be doing well: establishing knowledge and building skills. To reinforce its contribution to the required sustainability transition, HESD could further explore pedagogies associated with widely desired learning gains, such as system thinking and critical thinking. Second, if HESD remains committed to developing and reflecting affective learning gains, more ambition and courage are needed to imagine and experiment with new educational formats. Sustainability may be a critical and timely impetus for such invigorated discourse, but the empirical evidence reviewed here did not dispel the conceptual and ethical unease around sustainability concern as an educational end in itself.

4.3. How HESD Has Been Researched

The volume of research on HESD has increased steadily in recent years, with the majority of studies originating in the Global North. For the most part, the studies were reflections on specific courses conducted by the implementing lecturers themselves rather than deliberately designed comparative studies. As in other research fields, case studies in HESD have to deal with tensions between contextual relevance and transferability and abstraction of the findings, e.g., [175]. Unfortunately, most of the reviewed studies avoided taking a clear position on this issue. Many reflected on a specific case that was not contextualized in detail, while simultaneously seeking to establish universal causalities. Moreover, a substantial share of the studies did not fulfill the quality criteria for case studies in HESD proposed by Kyburz-Graber [176], who demanded a sound theoretical basis and complete documentation of the research instrument and process. Consequently, the findings of these studies neither provide generalizable guidance on effective HESD pedagogies nor serve as examples of critical co-creation of knowledge that could inspire a transformation of higher education elsewhere. In short, there has been limited methodological progress in the empirical HESD research over the period of this review.

Conceptual challenges in HESD regarding what should be learned and how it should be learned seem to contribute to the observed weaknesses in research instrumentation. The casual way in which “sustainability” was defined, if at all, across the reviewed studies demonstrates the importance of a precise definition in future HESD research. This is particularly important where “sustainability” is intended to include social and economic dimensions. Otherwise, HESD research risks perpetuating the “shallow consensus” on sustainability. The need for precision and clarity applies equally to the way learning gains are conceptualized and measured in HESD, notably regarding learning gains in affective domains. Asking students about their perceived gains in often highly abstract dimensions may be a pragmatic research approach, but it raises many questions about the validity and utility of the results. The collective concept of “sustainability competences” seems to further exacerbate such issues rather than helping to overcome them.

In conclusion, I suggest that for empirical inquiry into HESD to move forward, it will need to build momentum through a deliberate investigation into novel and theoretically sound learning approaches. Such research could be co-creative, in-depth, contextualized or comparative, larger-scale, and longitudinal. In all cases, it will need to be conceptually and instrumentally precise, inviting diversity rather than avoiding controversy through the use of conceptual umbrellas.