Introduction

Innovation has been touted as instrumental in fostering economic growth, yielding positive outcomes for society and the environment (Allen et al., 2019; Han et al., 2012). However, innovation has had its fair share of challenges in the same space. For instance, innovation does not always alleviate social concerns such as poverty, disease, and access to quality education. Further, innovation can foment new albeit unintended problems, such as the greenhouse crisis, a knock-on effect from the pervasive use of cars and air conditioners. Therefore, the expectations of innovation in sustainability have prompted the emergence and development of social innovation. Scholars believe social innovation can leverage the advantages of government, market, and society by propagating new ideas and practices to address existing societal problems better and charting a reasonable pathway for a more sustainable tomorrow (Fox et al., 2013; Pol and Ville, 2009).

A scan of the literature reveals that social innovation is an overarching term taken to embrace various initiatives and activities (Ayob et al., 2016), including but not limited to addressing unmet societal needs (Westley et al., 2017), balancing economic growth and environmental protection (Katarzyna et al., 2018), and transforming institutions and society for the better (Milley and Szijarto, 2020; Oeij, van der Torre, Vaas, and Dhondt, 2019) through innovative ideas and implementation (Castro-Arce and Vanclay, 2020; Tejedor et al., 2019). Pol and Ville (2009: p. 884) position social innovation as the implied new idea which has the potential to improve the quality or the quantity of life or both.

According to Drucker (2014), knowledge is critical in innovation because knowledge-based innovation impacts society and helps enterprises obtain good income. Inspired by this view, knowledge, and by extension, higher education is key to social innovation. Indeed, an effective avenue for disseminating and absorbing knowledge stems from the bulwark of knowledge (Blass and Hayward, 2014). For example, in university-community partnerships, the university has an advantage in promoting societal participation (Zermeño and de la Garza, 2020) and thus co-produce knowledge and action toward a more sustainable society (Loh, 2016). This finding suggests that the university or higher education catalyzes social innovation (Groulx et al., 2021). Furthermore, while numerous studies have suggested that universities can play a critical role in the quadruple helix to realize sustainable development, few studies have explored ways to promote and generate social innovation (Benneworth and Cunha, 2015). Hence, the value of higher education to social innovation is still unclear and warrants further attention (Schuch and Šalamon, 2021).

Although some studies have applied systematic reviews and meta-analyses to abstract the development of social innovation, focus on those studies that illustrate the broad strokes are lacking, such as the research collaboration network, shift in research topics, and identifying the influencers in the field to guide future research. In this context, a bibliometric map and content analysis were applied to visualize the structure and evolution of social innovation and explore the development of specific themes, namely, social innovation in higher education. We tracked the related articles on social innovation from the Scopus database using the bibliometric mapping software VOSviewer to visualize the network from the pool of existing studies. In contrast, SciMAT was used to track the evolution of social innovation in the field. We then summarized the progress related to higher education in social innovation. Three research questions were formulated:

(1) What is the structure and state of evolution in social innovation?

(2) Which or who are the most influential units (author, country, university) in social innovation?

(3) What is the level of research maturity of social innovation in higher education?

Methods

Research design

This paper examines the social innovation literature using a bibliometric map to reveal the relevance of information resources. It evaluates the authors, the research network, and the pathway of knowledge and ideas over time. Bibliometric maps spatially represent how disciplines, fields, specializations, individual publications or authors are related (Small, 1999). This method can unravel the development of one field and its network (Noyons et al., 1999; Van Eck and Waltman, 2010). Through bibliometric analysis, research can (1) evaluate the literature and academic contributions, (2) determine the influential units, (3) review and follow up on the research trend, (4) determine popular topics, (5) track the flow of knowledge and diffusion of ideas to elicit a richer understanding of a particular field, and (6) identify new research areas (Cobo et al., 2011; Goksu, 2021). The visualization network typically presents the structure, density, and centrality to reveal research trends. Next, content analysis was applied to delve into the development of higher education in the field, especially in tracing new emerging themes from the latest publications to provide further insights, which might have been ignored by bibliometric analysis.

Meta-data set

The Scopus database has material information, such as titles, exact journal names, and names of the first seven and the last authors. Hence, the information on the eight authors will form an accurate and broad cooperation network of the authors. Using Scopus will ensure the quality and validity of the results.

The Scopus database was searched from 1966 to July 2021 using the keyword “social innovat*”, and 3633 meta-datasets were extracted as of July 2021, without limiting the year, including articles, book chapters, conference papers, reviews, and books. The search language was limited to English. After filtering using VOSviewer, 2920 publications were deemed valid for analysis. Because of the differences between VOSviewer and SciMAT, the meta-data had two formats containing the same information. The files in a CSV format were for the VOSviewer analysis, while those in RIS format were for the SciMAT analysis. Each format had two files because Scopus allowed up to 2000 results to be downloaded simultaneously. The meta-data set on social innovation publications covered 10 disciplines across 189 countries. The datasets contained the article title, publication year, journal, author, abstract, keywords, and citation count. VOSviewer was used to assess the state of research collaboration of the authors, the leading research topics in a specified time window, and the leading papers in the field (Van Eck and Waltman, 2010) using three forms of bibliometric visualization: network, overlay, and density.

Figure 1 shows the data distribution by publication year and indicates that the first piece on social innovation was published in 1966, followed by weak interest till 2010. On the coverage by discipline, Fig. 2 shows that the notable ones were understandably in the social sciences (25%), business, management and accounting (18%), and economics, econometrics, and finance (10%).

Fig. 1
figure 1

Tally of social innovation research by publication year.

Full size image
Fig. 2
figure 2

Distribution of social innovation research by discipline.

Full size image

On network visualization, the nodes denote performance, comprising the tally of publications, number of citations, and article information. The larger the node’s size, the more effective or active the keyword, author, country, and study. In addition, the distance between the nodes indicates the strength of their cooperation. SciMAT is an open-source tool that performs science mapping analysis within a longitudinal framework, thereby measuring the impact of the element studied (Cobo et al., 2012). Hence, SciMAT was used to explore the trend of social innovation through the evolution map and measure the importance of the topics through the strategic diagram output of SciMAT. Harzing’s Publish or Perish software was used to obtain the H-index of the authors.

In short, VOSviewer was used to explore the co-authorship, author-keyword co-occurrence, and citation analysis. To ensure the results, SciMAT was used to explore the density and centrality of the keywords in each period of interest and explore the evolution of social innovation.

Data analysis

At the data preparation stage, words that share the same meaning were merged. For example, the term “social innovations” was replaced by social innovation and higher education was changed to university. In the analysis, a threshold was used to yield a clearer map. In the co-author analysis, a visualization network was formed to reflect the research collaboration by authors on social innovation. Meanwhile, the author-keyword choice was selected in the co-occurrence analysis to ensure an objective outcome using the author-center method. The strategic diagram was used to show the importance of each topic in the latest period, and the evolution map was formed to track the temporal path of social innovation. Performing a bibliographic coupling analysis revealed the similarity of the articles. Hence, clusters were formed through topic similarity from the studies. Citation analysis revealed the number of publications, citations, and the publications’ Total Link Strength (TLS) value. This method compared the impact of the studies, authors, countries, or institutions and provided the influential units. The analysis results were tabulated, and the visualization outputs were shown as networks. Finally, content analysis was used to augment the bibliographic analysis of the development of social innovation in higher education.

Results

Co-authorship status in social innovation

Authors who published at least one publication on social innovation were included in the co-authorship analysis. Among these 5792 authors, 115 had the largest set of connectedness and were selected for further analysis. Consequently, 15 clusters emerged, and 13 key authors functioning as structural holes were formed. Figure 3 shows the 15 co-authorship clusters are partitioned into four groups. The sky-blue cluster occupied the center of the entire network. Manzini had the most articles and was the most important structural hole that connected the other author networks. Each of the other three sets also had authors acting as structural holes. Rizzo was the key author who connected the red and sky-blue clusters with the light blue cluster and had the most publications in the red cluster. The authors with the most publications in the navy cluster were Castro-Spila and Unceta. Cipolla bridged the sky-blue and purple clusters and the green, light blue, and pink-purple clusters. No prominent author was found in these clusters. Next, Gong connected the sky-blue and grass-green clusters and contributed the most publications in the grass-green cluster. Zhang X. connected to the light purple cluster from the grass-green cluster, while Zhang L. linked the orange and light green clusters. The edge from Zhang X was joined to the pink, lemon green, and light brown clusters.

Fig. 3
figure 3

Co-authorship network map for social innovation.

Full size image

Six co-authorship clusters were found from the co-authorship analysis conducted to obtain the research collaboration by country on social innovation. Only countries with at least 20 publications were included in the analysis; 34 countries satisfied this criterion, as shown in Fig. 4.

Fig. 4
figure 4

Country co-authorship map for social innovation.

Full size image

Figure 4 shows the overlay visualization map comprised of six clusters with countries having at least 20 studies in the social innovation field. The countries exhibited good cooperation with the clusters to which they were connected. Comparing the node size, the most productive countries on social innovation were the UK, the US, Italy, Germany, and Canada, all developed nations. From the color bar at the bottom right-hand corner of the visualization, the darker the color, the earlier the year of publication. From the color of the nodes, the countries mentioned earlier were considerably mature in social innovation. Research on social innovation has also recently emerged in developing economies, such as the Czech Republic, Malaysia, Poland, Greece, Taiwan, and Mexico.

State of studies citing similar publications

In the bibliographic map, 381 publications that met the threshold for use each had at least 20 citations. Some of the 381 items in the network were not connected. The largest set of connected items comprised 19 publications, as shown in Fig. 5.

Fig. 5
figure 5

Network visualization map of an article’s bibliographic coupling.

Full size image

Five clusters emerged, as shown in Fig. 5. The fact that each cluster was less dispersed and the nodes were closer to each other suggested the work in the related clusters was likely to be cross-cited. In addition, the papers were in the same cluster, and their proximity to each other suggested the authors conducted research on similar topics. The paper Fundamentals for an international typology of social enterprise models by Defourny and Nyssens (2017), which generated four types of social enterprise models by combining the principals of interest and resource mixes, had the most citations (101). The paper Social innovation in question: The theoretical and practical implications of a contested concept by Marques et al. (2018), suggested a distinction between structural social innovation, targeted versions of social innovation, socioeconomic institution social innovation, and instrumental social innovation, was linked to the most publications (15). The paper A mobile-based barrier-free service transportation platform for people with disabilities by Wu et al. (2020) used a social business model to illustrate how ICT was integrated with the transport service providers and public resources to meet the transportation needs of disabled people was an emerging piece. No publication had a prominent bibliographic coupling effect in the network.

Keyword analysis and trends in social innovation

Compared to the single factor of co-occurrence analysis, the author-keyword co-occurrence may draw richer conclusions based on the main author of the field and provide clarity as to the research direction.

Overall, 5821 author–keywords were used in 2920 publications on social innovation. The keyword “social innovation” was excluded because it was used in the query. The co-occurrence analysis conducted on the 28 keywords that met the criteria for at least 20 publications identified four clusters of keywords, providing information on the related topics studied in the field.

The darker the color, the earlier the research, while the lighter the color, the more recent the research. Figure 6 shows that the early research topics in social innovation were innovation, social change, corporate social responsibility, and social economy. The keywords in the green and blue nodes included participatory design, empowerment, open innovation, and creativity. The green nodes included social entrepreneurship, sustainable development, education, and social capital. Lastly, recent studies focused on four topics: sustainability, university, governance, smart city, and co-creation. However, research on the last four topics remains insufficient.

Fig. 6
figure 6

Network visualization map of author-keyword co-occurrence.

Full size image

We also leveraged SciMAT to form an evolution map to identify the burning issue(s) in each window period. The H-index was used to benchmark the importance of each keyword. The stand-out topics for the various periods taken from 1983 to July 2021 were tagged in the evolution map in Fig. 7. Before 2010, innovation attracted research attention, and the influential research type was an experimental study. From 2010 to 2013, the focus on innovation persisted, although new topics, such as healthcare, social entrepreneurship, design thinking, and climate change, also attracted research focus. These latter topics, except for healthcare, became more prominent from 2014 to 2017, with innovation continuing to hold a pole position. Topics such as digitalization, social-ecological, and sustainable energy began to emerge during this period. More specialized themes on social innovation appeared in the last period (post-2017)

Fig. 7
figure 7

Strategic diagram for window period 2018–2021.

Full size image

A strategic diagram was formed to identify the importance of the topics and research opportunities from 2018–2021. Using the criteria of centrality and density, a 2 × 2 strategic diagram was formed with four quadrants: Quadrant #1 (motor themes that advance and sharpen the field), Quadrant #2 (developed but isolated themes that attracted much attention), Quadrant #3 (emerging or declining themes that suggest research opportunities or risks), and Quadrant #4 (basic and transversal themes that reveal common shared topics in the field) (Cobo et al., 2012; Viedma et al., 2020). From Fig. 8, Quadrant #1 revealed the theme “innovation” to have the highest H-index and centrality value, followed by “healthcare delivery”, which had high density and centrality. Quadrant #2 showed that the research topics “energy management”, “corporate social innovation”, “information processing”, and “consumer behavior” were well-developed but standalone topics. Among them, the keyword “energy management” had the highest H-index and density. Quadrant #3 indicated that the themes “circular economy”, “stakeholder management”, “technological innovation”, and “social network” were either emerging or declining. Lastly, the keywords “design thinking”, “social impact”, “sustainable food”, and “entrepreneurship” in Quadrant#4 were basic and transversal themes. The keyword “digitalization” also had high centrality and zero density, while “community base” was diametric.

Fig. 8
figure 8

Social innovation research by period (pre 2010, 2010–2013, 2014–2017, post 2017).

Full size image

Finally, Figs. 7 and 8 show that the number of publications, H-index, and citations were used to deepen the understanding of the dynamic evolution within social innovation. As shown in Table 1, the theme “innovation” was embedded in the social innovation research journey, followed by “design thinking”. This finding suggests that the two themes were inseparable from social innovation. In 2010–2017, “social entrepreneurship” was the basic and transversal theme and earned high citations, while “climate change” attracted some research traction. However, both themes disappeared after 2017, suggesting that the themes diminished in research importance. After 2014, “digitalization” became a common shared theme while “organization management” became a motor theme (strong centrality and high density), as evidenced by their higher publication rates. After 2017, studies on “smart city” and “health care delivery” entered the field. The themes “information processing”, “energy management”, “corporate social innovation”, “consumer behavior”, and “community-based” were well-developed but not well connected with earlier themes. Indeed, topics such as “technological innovation”, “social network”, “stakeholder management”, and “circular economy” were emerging and “social impact”, “entrepreneurship”, and “sustainable food” became common considerations in social innovation research.

Table 1 Principal research themes related to social innovation.
Full size table

Influential author

From the citation analysis of the authors with at least five publications, the top 15 authors with the highest TLS were found (see Table 2). Nijnik, Westley, Moulaert, Pisani, and Secco had the highest TLS scores. Meanwhile, Swyngedouw, Moulaert, and Westley had the most citations and publications. The latter group of authors also had the highest H-index, indicating the strength of their peer influence. Hence, the influential authors are Swyngedouw, Moulaert, and Westley, with Swyngedouw and Westley hailing from universities in the UK.

Table 2 Leading lights in social innovation.
Full size table

Most influential publication

Among the 2920 papers on social innovation found through citation analysis, 381 were cited at least 20 times, and 56 were connected to other research on social innovation. From Fig. 9, the visualization network presents the papers that act as structural holes. Table 3 shows the top 10 influential publications.

Fig. 9
figure 9

Overlay of cited articles.

Full size image
Table 3 Seminal work on social innovation.
Full size table

Figure 9 shows the poor network connection, linked by a chain of other publications. The red cluster occupied a special position in the publication network. In that cluster, the paper Social innovation in question: The theoretical and practical implications of a contested concept (Marques et al., 2018), which sought to clarify the definition and types of social innovation, was the structural hole linking 16 other publications. To the right of the network were two other key linkage papers. One paper Social innovation, an answer to contemporary societal challenges? Locating the concept in theory and practice (Grimm et al., 2013) sought to clarify the concept of social innovation and argued that a clearer definition would enable governments to develop more effective policy tools. The other paper, Social innovation: buzz word or enduring term? (Pol and Ville, 2009) distinguished social innovation from business innovation and identified a subset of social innovation.

From the number of citations in Table 3, the most cited paper was Governance Innovation and the citizen: The Janus face of governance-beyond-the-state focused on political governance, the fifth dimension of social innovation, published in Urban Studies, with Swyngedouw as the first author. Pol and Ville (2009)’s paper Social innovation: buzzword or enduring term? Published in the Journal of Socio-Economics was one of the top 10 publications and was a structural hole in the network.

Most influential journal

Using citation analysis to identify the leading journals on social innovation, we noted that 34 of the 1439 journals published at least 10 studies on social innovation. Table 4 lists the top 15 journals by TLS value.

Table 4 Top 15 influential journals on social innovation.
Full size table

Among these top-ranked journals, Sustainability (Switzerland) and Technological Forecasting and Social Change led on TLS value. In contrast, Urban Studies and Journal of Cleaner Production led on citation count, and Sustainability and Journal of Social Entrepreneurship had the most articles. Thus, the journal Sustainability, with the highest TLS value and containing the most articles, was the front-runner.

Most influential country

Research on social innovation has been undertaken in 189 countries and 4773 institutions of higher learning. Thus, we applied the criteria of article count, citation count, and TLS value to identify the leading universities and countries involved in social innovation research. Figure 10 shows the top 15 most influential countries on social innovation. Table 5 lists the top 15 institutions.

Fig. 10
figure 10

Breakdown of social innovation research by country and publication year.

Full size image
Table 5 Countries leading in social innovation research.
Full size table

The right side of Fig. 10 shows the ranking of the countries based on publications. The top three countries are the UK, the US, and Italy. The left side of Fig. 10 shows that research from the UK and the US had progressed steadily, while research from Italy peaked in 2017 and declined in 2018 and 2019. Overall, a leap in research output on social innovation by country in 2020.

Table 5 shows the distribution of social innovation by article, citation, and TLS value. The UK had the most articles, citations, and TLS, making the UK the most influential country in terms of social innovation research. However, the TLS of Spain was high despite its publications and citation counts being relatively less.

State of social innovation in higher education

Comparing the results of the author-keyword occurrence (Fig. 6) and the themes that emerged in the evolution of social innovation (Table 1), we observed that most of the keywords were the same or similar except the keywords “university” and “education”. This observation suggested that “university” was an emerging topic. Content analysis was performed to confirm this observation. Using the search string in “social innovation” AND “university” OR “higher education”, as well as limiting the language to “English” and paper type to “article”, 158 publications were extracted. Next, we scanned the title and abstract of the articles. As a result, 31 articles not related to “social innovation in higher education” and 1 article with no abstract were removed.

Among the 126 publications available for analysis, the criteria of publication year showed that only eight8 articles were published before 2012, followed by a handful of publications each subsequent year and surging in 2019. This trend was consistent with the result of the author-keyword occurrence, which suggests that scholars were paying greater attention to the role of higher education in social innovation. Case study was the main research method; eight articles were published after 2019, the other 116 articles were qualitative, and only 1 article used mixed methods research. As shown in Table 6, unlike the qualitative studies that looked at multi-stakeholders impact, 6 of the 9 quantitative articles focused on student groups. In short, quantitative studies sampled students. Finally, 67 survey samples from developed countries, including Spain, the United States, the United Kingdom, Italy, and Canada, contributed the most publications, with 12, 11, 9, 9, and 8 articles, respectively. Another 40 surveys were from developing countries, led by South Africa, Brazil, Malaysia, Colombia, and China, with 5, 5, 4, 4, and 3 publications, respectively. Only five surveys were collected cooperatively between developed and developing countries. A further five surveys focused on the collaboration between developing countries. The remaining nine articles were theoretical pieces. Moreover, 35 and 39 articles from developing and developed countries were published from 2017 to 2021. These findings suggested research traction on the role of higher education in social innovation.

Table 6 Quantitative articles on social innovation in higher education.
Full size table

Next, we sieved each abstract based on two aspects. First, we looked at the relationship between social innovation and higher education. Second, we looked at the ways that universities promoted social innovation. On the first aspect, few articles argued that social innovation offered creative approaches to universities. For example, using the case of Universiti Malaysia Terengganu, Hasan et al. (2017) mentioned that young universities needed to focus on social innovation and sustainability related courses for success. Colasanti et al. (2018) asserted that civic crowdfunding was an effective social innovative pathway for universities to improve their structure and service offerings. The dearth of publications on this aspect hinted at a long way to realizing the full potential of the relationship of social innovation with higher education.

Most articles viewed universities as engines of social innovation, with three specific topics of note. First, not only would the outcome of social innovation and universities benefit by embedding social innovation and social impact in the syllabus (Alden Rivers et al., 2015; Unceta, Guerra, and Barandiaran, 2021), students as social entrepreneurs will also be promoted, thereby increasing their confidence (Smith and Woodworth, 2012). Moreover, recent studies suggested that leveraging technologies (Massive Open Online Courses) and hackathons for university courses could drive social innovation more effectively (Calvo et al., 2019; Dabral et al., 2021). Second, some scholars argued for community-university interactions as a creative approach to research in social innovation. For instance, Groulx et al. (2021) noted that social innovation could catalyze community-university partnerships and alleviate resource constraints(Petersen and Kruss, 2021). Thus, using resources, such as incubators, open labs, living labs, and public collaboration labs was not only an effective way to gather human capital for the sustainable development of urban, rural, and marginalized areas (Kennedy et al., 2020), but could also generate collective knowledge through the exchange of experience and best practices (Gómez Zermeño and Alemán de la Garza, 2021; Prendiville, 2018; Zermeño and de la Garza, 2020). Third, several studies have posited that multi-disciplinary and cross-sectoral collaborations were needed to foster social innovation (Hayllar et al., 2018; Van Rensburg et al., 2019). Helix partnerships among civil society, business, government, and academia had to be formed (Bellandi et al., 2021; Zielke et al., 2021), notably for the Global South (Parthasarathy et al., 2021).

Conclusion and discussion

The Scopus database yielded 2920 publications for the bibliographic analysis, to explore the structure and evolution, and to identify the leading units of social innovation. Next, the structure of social innovation was revealed using co-authorship and coupling analyses, author-keyword occurrence analysis, strategic diagram, and evolution map. Finally, 126 publications were filtered and reviewed to augment the difference between the results and to elicit the nascent sub-topics of social innovation in higher education.

Structure of social innovation

Co-authorship analysis suggested weak research collaboration. Only 115 authors from 5792 candidates were linked to each other research-wise, and the linkages among the clusters they had formed were neither closed nor extensive. Manzini had the most articles and was the node that connected all clusters, making him the prime researcher in social innovation. Thus, Manzini’s work may lead to other key articles in the field. The bibliometric map also suggested that the other clusters were four standalone groups, each with a centroid linked to the primary cluster. This finding could be attributed to the overarching and cross-disciplinary nature of social innovation research.

Co-authorship analysis by country and year of publication revealed that social innovation research also involved inter-country cooperation. The result suggested that research partners from developed countries had experience in this domain, but an appetite for this research domain from developing country researchers was growing. The most productive countries in this regard were the UK, the US, Italy, Germany, and Canada. The cooperation amongst these countries was good. The Czech Republic, Malaysia, Poland, Greece, Taiwan, and Mexico were the new emerging countries with a thrust on social innovation research. Two reasons accounted for the more mature social innovation play in the developed countries than in the developing ones. First, social innovation needed both economic and knowledge bases. Economically poorer countries may favor innovations that generate economic value. Thus, developed countries would be relatively more assertive on social innovation. Second, the population in the developed countries are better educated and have higher literacy rates, which would naturally translate to more social innovation ideas (and more academic publications on social innovation). Nonetheless, anecdotal evidence suggests social innovation activities in developing countries had a recent resurgence consistent with the earlier analysis.

From the bibliographic coupling analysis, the publication Fundamentals for an international typology of social enterprise models (Defourny and Nyssens, 2017) had the most citations while Social innovation in question: The theoretical and practical implications of a contested concept (Marques et al., 2018) had the most links focused on the concepts of social innovation. In addition, citation count analysis suggested that 6 out of the top 10 of cited publications remain focused on new research agendas, pointing to the need for greater empirical verification to assess social innovation topics that are in vogue.

Evolution of social innovation

The analysis of the author-keyword co-occurrence suggested that innovation, social entrepreneurship, and sustainability were the most obvious keywords in social innovation. This result weighed in on the view found in many social entrepreneurship studies, that is, social entrepreneurship was a process involving the innovative use and allocation of resources to realize improved social value (Mair and Marti, 2006; Roslan et al., 2020; Saebi, Foss, and Linder, 2019). The keyword “sustainable development” stood out. This result concurred with the findings suggested by Eichler and Schwarz (2019), i.e., most social innovation cases were linked to at least one sustainability goal. From examining the year of publication, the keywords “sustainability”, “university”, and “governance” had emerged, suggesting renewed research interest on those topics.

The strategic diagram informed that innovation was the most effective motor theme, energy management was a well-developed theme, which concerned environmental sustainability. The most important emerging theme was social networks, which pertained to social sustainability. The basic and universal theme is social impact, i.e., more articles were focused on the effects of social innovation.

The evolution map formed by SciMAT also offered some insights. Except for two themes, innovation and design thinking which are embedded into the history of social innovation, most of the other themes attracted attention for no longer than three years. In other words, the themes within the social innovation research space could be faddish. The results were consistent with past studies. Thus, the concept of social innovation carried multiple connotations (Ayob et al., 2016) and shifted with time (Godin, 2012).

Leading authors in social innovation

Though citation analysis identified the leading authors, countries, and journals, choosing the most influential author was challenging because the outcome was criteria dependent. Nevertheless, social network theory posits that the individual functioning as a structural hole had a unique advantage in the network. From this perspective, Manzini, who occupies the core of the co-authorship network was the most influential. However, the other tests yielded different outcomes. For instance, Swyngedouw, Moulaert, and Westley were listed as the top three influential authors based on the TLS value, citation count, and H-index. Citation analysis suggested that the UK had the most citations and highest TLS value, which are a testament to the UK being the birthplace of social innovation. The results from the co-authorship analysis that developed countries, especially European countries have more social innovation activity and research indicate that the findings are consistent with the practice.

Finally, the journal Sustainability (Switzerland) had the highest TLS value and most publications on social innovation. The journal Technological Forecasting and Social Change ranked second on TLS value, but not on citation count and publication volume. Thus, diligently following the works of such authors or exercising keen readership in the respective journals is a better bet in identifying cutting edge work in social innovation.

Research maturity of social innovation in higher education

On the number of publications, few articles were pertinent to the topic “social innovation in higher education”. Echoing Schuch and Šalamon (2021), the relationship between social innovation and higher education remained unclear. Few articles examined the effects of social innovation on higher education. Finally, the results by publication year indicated that the number of articles in developing countries was catching up to the developed countries, suggesting that the Global South has accepted the value of universities in promoting knowledge democracy through social innovation. However, collaboration between the Global North and Global South has been limited, which could indicate a lack of knowledge transfer from the Global North to the Global South (Thomas and Pugh, 2020).

The extant studies on social innovation in higher education present three insights. The first insight is on curriculum transformation. Some scholars believed that student capacity could be improved by embedding social innovation into the curriculum. However, very few studies have explored the role of advanced technology in social innovation education, pointing to a research direction of this topic in the digitalization context. The second insight concerns university-community partnerships. Multi-prong approaches, especially incubator labs, have been examined and were deemed effective in leveraging and upskilling human capital. The third insight considers the role of multi-stakeholders to assess whether helix partnerships are the pathway to foster social innovation. However, no study has explored how higher education could be supported, such as providing clear guidance on impact measurement, financial support, appropriate infrastructure, and cooperation platform. Thus, there is research imperative to seek out the antecedents of social innovation, especially from the perspective of higher education.

Limitations and research directions

As with most bibliometric mapping and content analyses, this study has limitations. First, the meta-data were extracted solely from Scopus. Although Scopus contains the greatest number of publications globally, and we had endeavored to include as many forms of publication as possible to expand the scope of the study, the extant studies are not universal. Data could be collected from multiple databases and languages to fully cognate social innovation and its development in higher education. Second, this study had combined the strengths of network visualization of VOSviewer and strategic diagram, and the evolution map of SciMAT to highlight the emergent areas in social innovation research and the corresponding pathfinders and to elicit new research avenues. However, no bibliometric mapping software that supplies a plethora of analytical tools is available. This work recognizes leading scholars through the TLS and H-index but could not confirm if they remained active nor identified emerging pathfinders. Similarly, we identified leading countries through the number of publications but might have overlooked countries with strong practice but weak publication. Finally, this work illustrated the trajectory of social innovation from the perspective of changing research themes. Other perspectives exist, such as the impact factor of a publication in real terms, or how a university leads its peers. Both perspectives could be leveraged to broaden the knowledge in social innovation and higher education.