In higher education (HE) developments in climate urgency (UoE, 2019) and ICT force educational institutions to reconsider their traditional ways of education. In order to make a contribution to address climate change, HE institutions need to take measures to lower the environmental impact of their organisation (Klein-Banai & Theis, 2013) as well as to prepare students for a rapidly changing world (Pendleton-Jullian, 2019). Technological developments in ICT provide HE with possibilities to create a virtual educational space adjacent to the physical space of the campus. The current COVID-19 pandemic seems to accelerate these developments. In future, creating a mix of on-campus and online learning (so-called blended learning) could become the norm in HE. This transition towards blended forms of learning provides an opportunity for both improving educational quality and to address climate change in a very concrete way by significantly reducing student travel to campus (Versteijlen et al., 2017). Students commuting to campus is a large contributor to carbon emissions and because visiting a virtual space is place-independent, it may lead to a decrease of this commute (Caird et al., 2015; Versteijlen et al., 2017). Taking this environmental impact into consideration along with the capabilities of online learning, an educational design limiting on-campus learning on one or two days per week, thereby reducing student commute, supplemented with online ICT-supported learning, seems to be an obvious next step for HE. Obviously, such design should not compromise educational quality or sustainability-oriented learning. Hence, in this study design propositions are developed to optimise the chance the blended design meets these prerequisites.

Defining blended learning as a mix of online and on-campus learning is imprecise (Bliuc et al., 2007), because the extent in which online learning is used to substitute on-campus learning is not explicitly mentioned. According to Allen & Seaman (2003), you may only speak of a blended course if 30-79 per cent is delivered online. In addition, a blended educational design is more than just adding ICT enhancements to on-campus courses. It requires a fundamental redesign of the pedagogical approach, because blended learning changes or extends the mode of interaction with fellow-students, lecturers and content (Bliuc et al., 2007). If education is limited to one or two days per week, the learning activities before, during and after on-campus learning should be interconnected and enhancing each other (Garrison & Vaughan, 2008).

To prepare students for a rapidly changing world they need to acquire competences to apply new knowledge, while being aware of the environmental consequences (Pendleton-Jullian, 2019; Wals, 2019). Incorporating concepts of ecologies of learning and practice in an educational design probably is the best way forward to create such learning environment (Wals, 2019). According to Jackson (2019), in an ecology of learning and practice a student is considered as a whole person (thinking, feeling, acting with needs, desires, purposes), creating an ecology that is an enabler “to perceive and interact with their environment in order to accomplish the things that matter to them” (Jackson, 2019, p. 87). Blended learning extends this learning environment of students with a virtual space, affecting the learning process of an individual student, relationships with fellow-students and lecturers, access to places, and learning activities.

This research aims at finding propositions for the design of sustainable blended education in HE. Sustainable blended education is a thoughtful fusion of online and on-campus learning that applies concepts of ecological learning and practice. ‘Thoughtful’, because the fusion is characterized by using a pedagogical and didactical approach to integrate online and on-campus learning as well as by having a low-carbon organisation by limiting on-campus education on one or two days per week.

Introducing blended learning meets the key characteristics of a complex social intervention as formulated by Pawson et al. (2004), that is, evolving with changing circumstances because of the dependence on the educational situation regarding actions of students and lecturers, study subject, digital learning environment, and more. Also, there is no theoretical prescription how to design blended learning, so design and implementation involve negotiation and feedback at each stage. For short, there is no universal truth or general solution how to design and implement blended learning. A review using the realist research synthesis approach recognizes and addresses this complexity. Therefore, the aim of a realist review is explanatory: “what works for whom, in what circumstances, in what respects, and how?” (Pawson et al., 2004, p. 5). Using this method, design propositions are systematically developed aiming at optimising the chance to deliver a successful blended learning configuration. This systematic approach consists of four stages: Stage 1 The initial design propositions were extracted from influential theories about blended and ecological learning. These works were chosen because they differ in perspective how to design (blended) learning. Laurillard (2013) specifies with the Conversational Framework the interactions between the student and lecturer and also fellow-students. These interactions change while studying online or on-campus. Garrison and Vaughan (2008) introduces the Community of Inquiry framework and specifies the roles of lecturer and students collaborating in a blended setting. Ellis and Goodyear (2013) take a holistic view and address the learning environment as an ecology of learning in which online and on-campus learning is integrated and ecologically balanced. Barnett and Jackson (2019) add ecological learning and practice principles to the propositions. Stage 2 Several databases were searched for articles which contain empirical findings about relevant interventions with regard to the initial design propositions. A systematic, step-by-step approach was conducted to ensure transparency and rigour in searching these databases. The leading questions for the search were based on CIMO-logic, i.e. “Context to which it applies, Intervention proposed, Mechanism by which the desired Outcome is achieved” (Filius et al., 2018). Stage 3 Data extraction forms based on CIMO logic were developed and populated with the extracted data of 38 studies. The findings from the different studies were compared and the initial design propositions were refined in the light of the found evidence. Stage 4 Findings of stage 3 were discussed with educational experts, resulting in refining the conclusions and recommendations.

As stated before, it is not clearly prescribed how to design blended learning, so each studied intervention has its own design. In order to keep the results comprehensible the design of blended learning is generalized into three types, based on the division of learning activities performed online and in-class. These types are: 1. flipped design: learning acquisition performed online and practice assignments in-class, 2. blended design: knowledge acquisition and practice more or less equally divided between online and in-class learning, 3. collaborative blended design: knowledge acquisition and practice equally divided and supplemented with a collaboration project. This study resulted in the development of six design propositions: 1. Facilitation of the learning process of a student aims at self-regulation of learning and practice 2. Fostering of an affective learning climate 3. Facilitation of (a)synchronous interaction and discussion between fellow-students, and with lecturer 4. Transforming learning through acquisition and inquiry into an active process based on existing knowledge 5. Creating personal or social value with scaffolded and theory-based practice, compatible with learning preferences of students 6. Collaboration for constructing a shared (locally realised) outcome through participation and negotiation with fellow-students in a technologically-enhanced learning environment Comparing the programme theories (stage 1) with the results of the data extraction of the 38 studies (stage 3) there are some differences. Exclusively mentioned in the programme theories are: goal orientation in the learning process of students, creating socially relevant value with the specified tasks and locally realised outcomes in a collaboration project. The 38 studies add to the programme theories: personal learning styles and advantages of synchronous and asynchronous interaction and discussion. In the last stage recommendations were formulated for each design proposition, based on the mechanisms they probably evoke according to the studied experiences with the three types of blended learning.

Allen, I. E., & Seaman, J. (2003). Sizing the Opportunity: The Quality and Extent of Online Education in the United States, 2002 and 2003. Sloan Consortium (NJ1). Barnett, R., & Jackson, N. (2019). Ecologies for learning and practice: Emerging ideas, sightings and possibilities. Routledge. Bliuc, A., Goodyear, P., & Ellis, R. A. (2007). Research focus and methodological choices in studies into students' experiences of blended learning in higher education. The Internet and Higher Education, 10(4), 231-244. Caird, S., Lane, A., Swithenby, E., Roy, R., & Potter, S. (2015). Design of higher education teaching models and carbon impacts. . International Journal of Sustainability in Higher Education, 16(1), 96-111. Ellis, R., & Goodyear, P. (2013). Students' experiences of e-learning in higher education: The ecology of sustainable innovation. Routledge. Filius, R. M., de Kleijn, R. A., Uijl, S. G., Prins, F., van Rijen, H. V., & Grobbee, D. E. (2018). Promoting deep learning through online feedback in SPOCs. Frontline Learning Research, 6(2), 92. Garrison, D. R., & Vaughan, N. D. (2008). Blended learning in higher education: Framework, principles, and guidelines. John Wiley & Sons. Jackson, N. J. (2019). Ecologies for learning and practice in higher education ecosystems. In Barnett, R., & Jackson, N. (Ed.). Ecologies for learning and practice: Emerging ideas, sightings, and possibilities (pp. 81-96). Routledge. Klein-Banai, C., & Theis, T. L. (2013). Quantitative analysis of factors affecting greenhouse gas emissions at institutions of higher education. Journal of Cleaner Production, 48, 29-38. doi://dx.doi.org/10.1016/j.jclepro.2011.06.004 Laurillard, D. (2013). Teaching as a design science: Building pedagogical patterns for learning and technology. Routledge. Pawson, R., Greenhalgh, T., Harvey, G., & Walshe, K. (2004). Realist synthesis: An introduction. Manchester: ESRC Research Methods Programme, University of Manchester. Pendleton-Jullian, A. (2019). Education and innovation ecotones. Ecologies for learning and practice: Emerging ideas, sightings, and possibilities. Routledge. UoE.(2019). Environment & climate emergency working group white paper. Retrieved from https://www.exeter.ac.uk/media/universityofexeter/campusservices/sustainability/climateemergency/documents/Full-EnvClimE-White-Paper-11_11_19.pdf Versteijlen, M., Salgado, F. P., Groesbeek, M. J., & Counotte, A. (2017). Pros and cons of online education as a measure to reduce carbon emissions in higher education in the Netherlands. Current Opinion in Environmental Sustainability, 28, 80-89. Wals, A. E. (2019). Sustainability-oriented ecologies of learning. In Barnett, R., & Jackson, N. (Ed.). Ecologies for learning and practice: Emerging ideas, sightings, and possibilities (pp. 61-78). Routledge.