The evolution of SDG-related third sector and public administration literature: an analysis and call for more SDG-related research

Abstract

The seventeen Sustainable Development Goals (SDGs) were adopted in 2015 and achieving them by 2030 is crucial for human development. However, progress on the goals currently remains short of the requirements. As the third and public sectors play a crucial role in achieving the goals, this study analyzes how the SDG-related third sector and public administration literature has evolved over the last thirty years. I use a state-of-the-art method to map articles to the SDGs. In contrast to previous studies that have found an increase in publications that directly mention the SDGs, I find a decline in the proportion of articles that relate to the SDGs without necessarily mentioning them directly. I also analyze how the SDG-relatedness of an article corresponds to its citation count. While I find mixed results across SDGs and data sources, the relationship between SDG-relatedness and citation count is significantly more positive for work published after the adoption of the SDGs. While the association between SDG-relatedness and citation count is now positive for the third sector literature published after 2015, it is still negative for the public administration literature.

Keywords:

• Sustainable development goals
• third sector
• public sector
• public administration
• nonprofit studies
• bibliometric analysis

Introduction

In 2015, the United Nations General Assembly adopted the Sustainable Development Goals (SDGs). The seventeen goals aim to steer humanity toward sustainable development, with the objective of achieving them by 2030. We are now midway through the timeline established to meet these goals, and while some progress has been made, it remains far short of the requirements (Gehringer 2020), with the last two years even showing no progress (Sachs et al. 2022). To achieve the goals, collaboration between the public, private, and third sectors is crucial (Hecht et al. 2012; Saxena et al. 2021). Next to directly contributing to the SDGs via service provision, the third sector plays a key role in achieving the SDGs by serving as a broker between sectors and representing the voice of civil society (Arhin 2016). So although the third and public sectors play a crucial role in achieving the SDGs, studies on how these sectors contribute to the goals are lacking (Pizzi et al. 2020).

This study aims to examine how and to what extent research about the third and public sectors addresses the SDGs. Specifically, I test whether research about the SDGs increased or decreased over the years and which SDGs have received the greatest amount of research attention. I use query systems that have been developed to detect SDGs in text and a novel state-of-the-art ensemble model developed by Wulff, Meier, and Mata (2023) to analyze thousands of research articles published in the nonprofit studies literature and the public administration literature. The results show a decrease in SDG-related research across both bodies of work. While this is a source of concern, the analyses also show that the degree of SDG relatedness of an article published since the adoption of the SDGs is positively associated with its number of citations for the nonprofit studies literature.

I first highlight the role that the third and the public sectors play in achieving the SDGs in the next section and then provide a brief literature review on bibliometric studies about the SDGs. This is followed by a methods section where I introduce the data and the analyses. The results of these analyses are presented in the results section and set into a broader context in the discussion and concluding section.

The role of the third and public sectors in achieving the SDGs

Because the SDGs are global goals, they require global action (Sachs et al. 2019). As development challenges are increasingly interlinked, persistent, and complex, robust partnerships across the private, public, and third sectors are needed to achieve the SDGs (Saxena et al. 2021). This objective is also recognized through SDG 17 (Partnership for the Goals) (von Schnurbein 2020). According to Saxena et al. (2021), partnerships between the third sector and the other two sectors are the least established. This is concerning as one of the third sector’s roles in achieving the SDGs is to represent a diverse range of organizational and civic interests (Hassan, Lee, and Mokhtar 2020). According to Saxena et al. (2021), the public sector’s role is to ensure that the SDGs are considered at the implementation level as well as to set up initiatives to monitor progress on the goals. These roles are also highlighted by Callmer and Bradley (2021) and Hess and Gentry (2019). The private sector’s efforts to contribute to the SDGs have largely been operationalized through practices centered on corporate social responsibility (CSR) (Saxena et al. 2021). But greater integration with the actions of the other two sectors is needed since CSR alone will likely not be sufficient (Mintzberg 2019).

While the third sector acts as a partnership broker (Arhin 2016; Gehringer 2020), there are also more direct ways it contributes to the SDGs. For example, with activities such as mass protests and campaigning, civil society actors in general and nongovernmental organizations (NGOs) specifically can directly influence intergovernmental policy making by pressuring governments (Rietig 2011). However, Sénit (2020) found the influence of civil society on global SDG-related policy making only to be moderate. Third sector organizations also play a prominent role in developmental aid (Appe 2022) and in the area of climate change (Johnson, Coma, and Castonguay 2023). More generally, as third sector organizations offer a wide range of services that the private and public sector cannot or do not provide to the community, they play a crucial role in achieving the SDGs by providing services that benefit society (Nordin, Khatibi, and Azam 2022). The ways third sector organizations can contribute to the SDGs are nicely summarized by Kagan and Dodge (2023). Although they focus on climate change, the statement can be expanded to the SDGs more broadly. Kagan and Dodge (2023, 2) note that third sector organizations “play a critical role in the climate crisis by advocating for policy change, influencing corporate and individual behavior, and supporting vulnerable communities.”

Similarly, Hege and Demailly (2017) identified four roles for NGOs in the context of SDG implementation: holding the government accountable, holding the private sector accountable, communicating the SDGs, and implementing projects. So, although the third sector has great promise in contributing to the SDGs, the scholarly literature on how the third sector can contribute might be lagging. This is exemplified by the study of Kagan and Dodge (2023) who found that the response to the climate crisis by the third sector scholarly field has been slow. According to these authors, only a limited body of literature examined the relationship between the third sector and the climate crisis. This study aims to test whether this sluggish pace of scholarly response extends to the broader scope of the SDGs.

Bibliometric studies about the SDGs

A series of bibliometric analyses and expert surveys have shed light on how the SDGs are represented in current academic literature. Most closely related to this article’s field of study, Mio, Panfilo, and Blundo (2020) and Pizzi et al. (2020) investigated the representation of SDGs in business and management research. Mio, Panfilo, and Blundo (2020) found that the most frequently mentioned SDGs in business-related papers were SDG 9 (Industry, Innovation, and Infrastructure), SDG 3 (Good Health and Well-being), and SDG 6 (Clean Water and Sanitation), alongside SDG 12 (Responsible Consumption and Production) and SDG 13 (Climate Action). In contrast, SDG 5 (Gender Equality), SDG 10 (Reduced Inequalities), and SDGs 14–17 (Life Below Water; Life on Land; Peace, Justice and Strong Institutions; and Partnerships to Achieve the Goals) were the least mentioned, appearing in only 14% of articles.

While Pizzi et al. (2020) do not directly map articles to specific SDGs, they identified four major research themes: technological innovation, firms’ contributions in developing countries, non-financial reporting, and education for SDGs. Broadening the scope, several studies have found that SDG 3 (Good Health and Well-being), SDG 13 (Climate Action), and SDG 6 (Clean Water and Sanitation) are recurrently among the most represented SDGs in the literature (Londoño-Pineda and Cano 2022; Meschede 2020; Sweileh 2020; Yamaguchi et al. 2023; Yeh et al. 2022). SDG 7 (Affordable and Clean Energy) and SDG 2 (Zero Hunger) also frequently occur, especially in the bibliometric analyses of Yamaguchi et al. (2023) and Londoño-Pineda and Cano (2022). By contrast, research on SDG 5 (Gender Equality), SDG 10 (Reduced Inequalities), SDG 14 (Life Below Water), and SDG 16 (Peace, Justice, and Strong Institutions) were found to be underrepresented across the board (Londoño-Pineda and Cano 2022; Meschede 2020; Mio, Panfilo, and Blundo 2020; Sweileh 2020; Yamaguchi et al. 2023; Yeh et al. 2022).

As presented above, this scarcity of literature is particularly notable in business and management research (Mio, Panfilo, and Blundo 2020). While these studies are quite consistent in their findings, discrepancies appear when contrasting more global research with specific geographic contexts. For example, Trane et al. (2023) found that in the context of the European Union (EU), environmentally related SDGs received more frequent attention than social issues, which contradicts the findings of Körfgen et al. (2018) for articles published by Austrian universities. Expert opinions collected by Salvia et al. (2019) also offer an interesting contrast to the results that are based on bibliometric data, identifying SDG 8 (Decent Work and Economic Growth), SDG 14 (Life Below Water), and SDG 16 (Peace, Justice, and Strong Institutions) as least represented among the surveyed experts. In general, studies observed an exponential growth in the volume of SDG-related publications over time (Pizzi et al. 2020; Sweileh 2020; Yeh et al. 2022), indicating a rising global interest in research pertaining to the SDGs.

The current study differs from the above-mentioned bibliometric research in two important ways. First, all the above-mentioned bibliometric investigations base their sample on literature that directly mentions the SDGs in the title, abstract, or keywords. For example, the sample of Pizzi et al. (2020) is based on management-related articles whose titles, abstracts, or keywords contained the words “sustainable development goal*” or “sdg*.” One limitation of this search strategy is that it only identifies a fraction of the articles that are actually related to the SDGs, since published work can be related to the SDGs without explicitly referring to the SDGs (e.g., research on poverty reduction or climate change). The current study avoids this limitation by analyzing whether articles are related to the SDGs even if they do not explicitly mention the goals. This drastically broadens the scope of investigation, since only a small fraction of the putatively relevant scientific literature directly mentions the SDGs. This study is therefore able to paint a more complete picture of how the (nonprofit- and public administration related) literature addresses the SDGs.

Second, the current study differs from the above-mentioned research in how articles are mapped to the SDGs. While I also rely on an automated method to map the studies to the SDGs, in contrast to most previous studies I use a validated state-of-the-art method to do so. This contrasts with articles that developed their own method to map the studies to the SDGs (e.g., Körfgen et al. 2018; Londoño-Pineda and Cano 2022). This is problematic because the analysis of Wulff, Meier, and Mata (2023) showed that these methods can produce biased results (i.e., some systems are more sensitive to specific SDGs relative to others). For example, Sweileh (2020) used the Aurora queries (Vanderfeesten, Otten, and Spielberg 2020) to map studies to the SDGs.¹ While this method was validated (Vanderfeesten, Spielberg, and Gunes 2020), a comparison of different methods to map text to SDGs carried out by Wulff, Meier, and Mata (2023) found that it does not perform well with the kind of text used in the study by Sweileh (2020) (i.e., abstracts). The results of Wulff, Meier, and Mata (2023) also revealed that the Aurora system underestimates SDGs 2, 6, 7, 8, 9, and 10 but overestimates SDG 13. Other systems that can be used to map text to SDGs face similar issues and as outlined in the next section I therefore use the state-of-the-art ensemble model developed by Wulff, Meier, and Mata (2023) to map the articles to the SDGs.

Methods

Data

Articles published between 1990 and 2022 form the basis of this analysis. To identify articles from the nonprofit studies literature, I followed LePere-Schloop and Nesbit (2023) and used the query they developed to distinguish third sector-related literature (reported in the Supplementary Material). Also ­following LePere-Schloop and Nesbit (2023), I used the Web of Science “public administration” subject-category filter to find and download public administration articles. For all these downloaded records, I only retained articles with neither a missing title nor a missing abstract. Table 1 displays the number of articles that were available with title and abstract. This data formed the basis for the analyses described in the next section. Because the subject category “public administration” only contained three articles in 1990, I excluded publications from this year from the analysis for this data source. Similarly, there were no articles remaining for the year 1990 in the nonprofit studies literature data after excluding data that either had a missing title or a missing abstract.

Table 1. Number of articles per source.

Data	Nonprofit studies literature	Public administration literature
Number of articles	27,286	58,155

Analyses

To detect SDGs in the titles and abstracts of the analyzed papers, I used the text2sdg R package developed by Meier, Mata, and Wulff (2021). This package implements query systems and ensemble models that have been developed to detect SDGs in text. A total of six query systems are available through text2sdg. These query systems assign SDGs to a text if a given query produces a hit. Much like a query that can be used to query databases like Scopus, these queries use a combination of keywords. For example, the query system developed by Elsevier (Jayabalasingham et al. 2019) contains a query that assigns SDG 1 (No Poverty) to text that contains the word combination “extreme poverty.” Most of these queries were designed by domain experts and gradually refined to make them more accurate (e.g., Vanderfeesten, Spielberg, and Gunes 2020). Although these queries can be used to detect SDGs in any kind of text, most of them were designed to detect SDGs in academic publications (i.e., abstracts). These queries are already being used in policy-relevant publications. For example, the Elsevier queries are used to rank the SDG-related research output of universities in the Times Higher Education Impact Rankings.

Wulff, Meier, and Mata (2023) tested how accurately these queries detect SDGs in text and found that the systems vary in their sensitivity (i.e., true-positive rate) and specificity (i.e., true-negative rate). These analyses also showed that the labeling systems are sensitive to the length of the input text, with longer input text (erroneously) generating more hits. The results also revealed that the query systems show biases for specific SDGs (i.e., some query systems systematically over- or underestimate certain SDGs) (Wulff, Meier, and Mata 2023). To alleviate these shortcomings, Wulff, Meier, and Mata (2023) developed an ensemble model that leverages the predictions from all six labeling systems and takes the text length into account. This model exceeds the labeling performance of all other labeling systems and achieves both a high accuracy and a low false positive rate. That the ensemble model takes text length into account is crucial for this study, as abstracts became longer over the years (reported in the Supplementary Material). Next to expert-labeled data, the ensemble model was also trained on synthetic data sets to reduce the false positive rate. The text2sdg R package allows the user to choose ensemble models that vary in the amount of synthetic data that was used to train the model. The higher the amount of synthetic data, the lower the false positive rate (Wulff, Meier, and Mata 2023). Since the variant of the ensemble model that was not trained on synthetic data suffers from a very high false positive rate, I do not use this model in the analysis. Due to the shortcomings of the individual query-based systems, I only deployed the ensemble models for the analysis. I use models that required a third, equal, or triple the amount of synthetic data relative to the expert-labeled data.

To answer how the amount of research that is related to the SDGs evolved over the years, I calculated the share of articles that relate to at least one SDG for each year between 1991 and 2022. Additionally, I determined the average number of SDGs that articles were assigned during this time frame. These results are visualized for all ensemble models. I also compute the prevalence of each SDG over the years and the overall prevalence combined over all years. I used the ensemble model that was trained on an equal amount of synthetic and expert-labeled data for these two tasks because this model has both a low false positive rate and a high accuracy. It is also the default model in the text2sdg package. To assess whether certain SDGs are likely to occur together in the published articles, I calculated the correlation between the detected SDGs.

To make the trends more visible, I plot the smoothed instead of the raw values for figures that include a temporal dimension (Figures 1 and 2). I used the function “geom_smooth” to do this with the ggplot2 R package (Wickham 2011). This function fits a local polynomial regression of degree two to model the relationship between x and y.

Figure 1. Evolution of SDG prevalence in the nonprofit studies literature.

Figure 2. Evolution of SDG prevalence in the public administration literature.

To assess whether the adoption of the SDGs in 2015 led to an increase in the number of published SDG-related articles, I used the Causalimpact R package (Brodersen et al. 2015). This package fits a Bayesian structural time-series model to the data prior to the intervention (i.e., prior to the adoption of the SDGs). This model is then used to predict the counterfactual after the intervention, i.e., how the dependent variable would have evolved after the intervention if the intervention had never occurred. This counterfactual is then used to test whether the intervention had a significant effect. More details can be found in Brodersen et al. (2015) and in the detailed description of the results of these models in the Supplementary Material.

I relied on Poisson regression models to test whether articles that are related to SDGs were cited more often. I used the results of the ensemble equal model for these analyses. In all models, the number of citations an article received as of 2022 is the dependent variable. All models also control for a linear and a quadratic time trend (year of publication). The independent variables in the first model are dummy variables that specify for each SDG whether a given article relates to this SDG or not. I also use a model to test whether the adoption of the SDGs led to an increase in citations for articles that relate to the SDGs. I used the number of SDGs an article relates to as the independent variable for this model and interacted this variable with a dummy that indicates whether an article was published after 2015. The coefficient of this interaction will tell us whether the association between an article’s degree of SDG relatedness and its citation count is stronger for articles published after 2015. Finally, I also fit a model only for articles published after 2015 with the number of SDGs a given article relates to as the independent variable to see how strong the association is for articles published after the adoption of the SDGs. The formulas of these three Poisson regression models can be found in the Supplementary Material. I only plotted the coefficients of interest in the results section but included the regression tables in the Supplementary Material.

Results

Figure 1(A) shows the percentage of articles with at least one SDG in the nonprofit studies literature over time. The ensemble equal model shows a minimal decline in articles with at least one SDG over the years, with less than 50% of the articles being related to at least one SDG. Figure 1(B) reports the mean number of SDGs per article over time. The patterns are very similar to those reported in Figure 1(A), indicating that not many articles relate to more than one SDG. Figure 1(C) shows the percentage of articles with a given SDG over time. The prevalence of the individual SDGs remained relatively stable over the years. Quite a few SDGs reached their highest prevalence in 2022, indicating a recent growth in articles covering these SDGs. Among these are SDG 4 (Quality Education), SDG 5 (Gender Equality), and SDG 17 (Partnership for the Goals).

Just like the third sector, the public sector plays a crucial role in achieving the SDGs. Figure 2 shows the evolution of SDG-related research in this literature. While the amount of SDG-related research is relatively steady until 2010, this year marks the beginning of a slow decline in SDG-related research both in the percentage of articles with at least one SDG (Figure 2(A)) and in the mean number of SDGs per article (Figure 2(B)). Compared to the nonprofit studies data, the mean number of SDGs per article is slightly higher (around 1, ensemble equal model) until 2010. From then on, it slowly converges to about the same number as for the nonprofit studies literature data set. Looking at the evolution of individual SDGs in Figure 2(C), we see that SDG 3 (Good Health and Well-being), SDG 13 (Climate Action), and SDG 17 (Partnerships for the Goals) reached their highest prevalence in 2022. SDG 8 (Decent Work and Economic Growth) is the most prevalent SDG over all years but shows an almost steady decline.

In the Supplementary Material, I report the results of the causal impact analysis that tests whether the adoption of the SDGs in 2015 led to a significant increase in the amount of published SDG-related articles. This method indicated a significant decrease in the mean number of SDGs per article for the public administration literature. A significant negative effect was also found for the number of articles with at least one SDG for articles published after 2015 in the public administration literature. Unsurprisingly given the already reported results, none of the tests indicated an increase in SDG-related research after 2015.

This is in stark contrast to the percentage of articles that directly mention the SDGs (“sdg”, “sdgs”, or “sustainable development goals”) in the title, abstract, or keywords. There was an exponential increase in these articles over the years for both the public administration literature and the nonprofit studies literature (Figure 3). This shows that while there was a decrease in the share of articles that (indirectly) relate to the SDGs, the share of articles that directly mention/study the SDGs increased over time.

Figure 3. Percentage of articles published between 2010 and 2021 that directly mention the SDGs in the title, abstract, or keywords.

Figure 4 shows the prevalence of SDGs for the two data sources aggregated over all years. SDG 8 (Decent Work and Economic Growth) is the most prevalent SDG overall. This is followed by SDG 10 (Reduced Inequalities) for the nonprofit studies literature data and SDG 16 (Peace, Justice, and Strong Institutions) for the public administration literature. These SDGs are the most prevalent overall in both datasets. SDG 14 (Life Below Water) and SDG 15 (Life on Land) belong to the least prevalent SDGs. Looking at the differences in the frequency of specific SDGs, SDG 3 (Good Health and Well-being) is a bit more common in the nonprofit studies literature than in the public administration literature, and SDG 13 (Climate Action) and SDG 16 (Peace, Justice, and Strong Institutions) are more prevalent in the public administration literature.

Figure 4. Overall prevalence of SDGs. Note: The highest and lowest ranks per data source are indicated with numbers on the bars.

Finally, we can also look at the correlation of SDGs, i.e., whether certain SDGs are more likely to occur together in the articles. Figure 5 shows the SDG correlation matrix for the public administration literature data, ordered according to a hierarchical cluster analysis. The dendrogram of the cluster analysis is shown on top of the figure. Figure 5 shows that SDG 7 (Affordable and Clean Energy) and SDG 13 (Climate Action) are most likely to co-occur within a given article, with a correlation of 0.38. This is followed by SDG 6 (Clean Water and Sanitation) and SDG 11 (Sustainable Cities and Communities) (0.32) and SDG 1 (No Poverty) and SDG 10 (Reduced Inequalities) (0.26).

Figure 5. Correlation of SDGs within public administration articles, ordered according to a hierarchical cluster analysis.

Figure 6 shows the same analysis for the nonprofit studies literature. In these articles, SDG 8 (Decent Work and Economic Growth) and SDG 9 (Industry, Innovation and Infrastructure) show the highest correlation (0.27), followed by SDG 6 (Clean Water and Sanitation) and SDG 11 (Sustainable Cities and Communities) which are again relatively highly correlated (0.26), and SDG 8 (Decent Work and Economic Growth) and SDG 10 (Reduced Inequalities), which are slightly more correlated in the nonprofit studies literature than in the public administration literature (0.21 vs. 0.19).

Figure 6. Correlation of SDGs within nonprofit studies articles, ordered according to a hierarchical cluster analysis.

Comparing the clusters that are formed by specifying five clusters, we see that three clusters are stable across the two sets of literature. Namely, in both bodies of literature SDG 7 (Affordable and Clean Energy) and SDG 13 (Climate Action) form a cluster, SDG 2 (Zero Hunger), SDG 14 (Life Below Water), and SDG 15 (Life on Land) form a cluster, and SDG 6 (Clean Water and Sanitation), SDG 11 (Sustainable Cities and Communities), and SDG 12 (Responsible Consumption and Production) form a cluster. Indeed, the only difference in the clusters across the two sets of literature is that SDG 8 (Decent Work and Economic Growth), SDG 9 (Industry, Innovation, and Infrastructure), and SDG 17 (Partnerships for the Goals) form a cluster in the public administration literature while this cluster additionally contains SDG 4 (Quality Education) and SDG 16 (Peace, Justice, and Strong Institutions) in the nonprofit studies literature.

Do SDG-related articles get cited more often?

The results so far show a decrease in the amount of SDG-related research in the nonprofit and public administration literature over the years. However, perhaps SDG-related articles received greater attention in the form of being cited more often than their non-SDG-related counterparts. As described in the methods section, I therefore also analyzed whether articles that are related to (specific) SDGs are cited more often than their counterparts. Figure 7(A) shows the coefficients of a Poisson model that has the number of times an article has been cited as the dependent variable and whether a given article relates to a given SDG as independent variables. The model also controls for a linear and quadratic time trend (year of publication). Figure 7(A) shows that for most SDGs, articles related to the specific SDG were cited less frequently. This is confirmed by an analysis reported in the Supplementary Material where I used the number of SDGs an article is related to as the independent variable in the Poisson models instead of the individual SDGs. However, some SDGs are associated with an article being cited more often. This positive association is more pronounced in the nonprofit studies data, where this is the case for seven SDGs. SDG 17 (Partnerships for the Goals), SDG 7 (Affordable and Clean Energy), SDG 12 (Responsible Consumption and Production), SDG 9 (Industry, Innovation, and Infrastructure), and SDG 13 (Climate Action) show the strongest positive association among these SDGs. SDG 13 and SDG 17 are positively associated with the number of citations for both data sets, whereas SDG 15 (Life on Land), SDG 11 (Sustainable Cities and Communities), SDG 6 (Clean Water and Sanitation), SDG 4 (Quality Education), SDG 3 (Good Health and Well-being), and SDG 1 (No Poverty) are negatively associated with citations across both data sets.

Figure 7. Association between SDG-relatedness of an article and its citation count. Notes: (A) Effect of an article being related to a specific SDG on citation count. The horizontal axis is truncated for visibility. Regression tables are shown in the Supplementary Material. (B) Effect of SDG relatedness of an article on citation count. (C) Effect of SDG relatedness of an article on citation count for articles published after 2015.

It could be that the adoption of the SDGs in 2015 led to an increase in the number of citations of SDG-related articles. I tested this by interacting the number of SDGs an article is related to with a dummy that indicates whether the article was published after 2015 in the Poisson regression. The interaction is significantly positive for both data sources, indicating that the association between the number of SDGs an article is related to and the number of times it is cited is significantly more positive for articles published after 2015 (Figure 7(B)). Indeed, there is now a significant positive association between the number of SDGs an article relates to and the number of times the article has been cited for articles published after 2015 in the nonprofit studies literature (Figure 7(C)). It would be interesting to test whether SDG-related articles published before 2015 also experienced an increase in citations after 2015, but since I do not have yearly citation data, I unfortunately was not able to test that proposition.

Discussion and conclusion

Because the third sector and the public sector play an important role in achieving the SDGs, this study set out to examine the prevalence and evolution of SDG-related research in the nonprofit studies and public administration literature. The results of the analyses show that third sector and public administration research that is related to the SDGs remains scarce and has even declined over the past few years. While it might be surprising or even worrying that such research is of modest prevalence despite the importance of the SDGs, it is in line with recent reviews which have found that very few top business journals published articles about the SDGs (Mio, Panfilo, and Blundo 2020; Pizzi et al. 2020). The results are also consistent with a recent review that found limited third sector research that addresses climate change (Kagan and Dodge 2023).

The results showed that SDG 8 (Decent Work and Economic Growth) is most common in this literature, followed by SDG 10 (Reduced Inequalities) for the nonprofit studies literature and SDG 16 (Peace, Justice, and Strong Institutions) for the public administration literature. The high prevalence of SDG 8 aligns with the findings of Mio, Panfilo, and Blundo (2020) and Yamaguchi et al. (2023), who also found this goal to be highly represented in the literature. The strong prevalence of SDG 3 (Good Health and Well-being) is also in line with previous literature (Londoño-Pineda and Cano 2022; Meschede 2020; Sweileh 2020; Yamaguchi et al. 2023; Yeh et al. 2022). Although these same studies also found SDG 6 (Clean Water and Sanitation) and SDG 13 (Climate Action) to be among the most represented goals, that was not the case in this research. That SDG 14 (Life Below Water) and SDG 15 (Life on Land) were found to be among the least prevalent SDGs is consistent with previous studies indicating that especially SDG 14 is underrepresented in the literature (Londoño-Pineda and Cano 2022; Meschede 2020; Mio, Panfilo, and Blundo 2020; Sweileh 2020; Yamaguchi et al. 2023; Yeh et al. 2022). While the strong presence of SDG 8 contradicts previous studies to some extent (Salvia et al. 2019), it confirms the findings of Yamaguchi et al. (2023) who showed that keywords relating to SDG 8 (Decent Work and Economic Growth) and SDG 9 (Industry, Innovation, and Infrastructure) had the greatest relevance among the analyzed keywords and were present in multiple clusters. Previous studies found that SDG 5 (Gender Equality), SDG 10 (Reduced Inequalities), and SDG 16 (Peace, Justice and Strong Institutions) have been the focus of limited attention (Londoño-Pineda and Cano 2022; Meschede 2020; Mio, Panfilo, and Blundo 2020; Sweileh 2020; Yamaguchi et al. 2023; Yeh et al. 2022). This finding contrasts with the results of this study, where these SDGs are among the most prevalent ones.

The correlation between the SDGs revealed that some goals are more likely to co-occur in the articles than others. In fact, the clusters revealed by the correlation matrix were highly similar across the two bodies of literature. The correlation of SDGs corresponds to the literature on synergies between the SDGs (i.e., positive correlations between pairs of SDG indicators) (Kroll, Warchold, and Pradhan 2019; Pradhan et al. 2017). For example, SDG 1 (No Poverty) and SDG 10 (Reduced Inequalities) belong to the SDGs with the strongest synergies in the article by Pradhan et al. (2017), which is in line with the positive correlation of these SDGs in the ­current study.

While some of the results reported here are in line with the outcomes of previous research, there also are some notable differences. Two factors can explain these varying results. First, in contrast to previous bibliometric studies about the SDGs, I analyzed all articles in the population of interest, and not only those that directly refer to the SDGs. This paints a more comprehensive picture of SDG-related research, as articles can relate to the SDGs without directly mentioning the goals. Indeed, this is the case for about half of the articles analyzed in this study (Figures 1(A) and 2(A)), while only a fraction of them directly mentions the goals (Figure 3). Second, the difference in results probably relates to the subject area of the analyzed articles. For someone who is familiar with the nonprofit studies and public administration literature, it is not surprising to see that SDG 8 (Decent Work and Economic Growth), SDG 10 (Reduced Inequalities), and SDG 16 (Peace, Justice, and Strong Institutions) are the most represented goals, given that these are central themes in this literature (Ma and Konrath 2018). Indeed, SDG 8 also belonged to the three most represented goals in a bibliometric study about supply-chain practices and business strategies (Agrawal et al. 2022).

This study also revealed an important difference between the publication trend of research that directly mentions the SDGs and research that is related to the goals but does not mention them explicitly. While the exponential increase in the former articles is consistent with the findings of past studies (Pizzi et al. 2020; Sweileh 2020; Yeh et al. 2022), the relative decline in the latter contradicts this pattern. Thus, while there seems to be an increase in research that directly mentions the SDGs, there appears to be a decrease in the share of published articles that relate to the goals without directly mentioning them. This is worrying because it is these articles that build a broad knowledge source on how we can achieve the SDGs. Of course, articles that directly mention the goals and therefore likely have them as their main focus of investigation are at least as important. By, for example, synthesizing the results of studies that relate to the SDGs without explicitly mentioning them, these articles provide important contributions. However, without the former publications, there is not a lot that this work can build on.

While the volume of literature related to SDGs does not necessarily reflect the progress made toward meeting the goals, it may act as an early signal of future progress. This is because science plays a critical part in promoting evidence-based strategies and in conducting research that supports the attainment of the SDGs (Allen, Metternicht, and Wiedmann 2021; Messerli et al. 2019). Accordingly, the high prevalence of nonprofit studies and public administration articles that are related to SDG 8 (Decent Work and Economic Growth), SDG 10 (Reduced Inequalities), and SDG 16 (Peace, Justice, and Strong Institutions) is a promising result. Especially when considering that current progress on these goals is moderate at best, with a trend that is not particularly encouraging (United Nations 2022). By contrast, the reported decrease in the share of SDG-related articles might be an early warning sign of future stagnation on the goals, especially in civil society and public sector-related areas.

An overly optimistic interpretation would be to attribute the decline in SDG-related research to a saturation of work on this topic. However, both my results and findings from the literature show that more research is urgently needed in this area (Kagan and Dodge 2023; Pizzi et al. 2020). I therefore hope that my results might serve as a call for researchers to produce more research that relates to the SDGs. Science, being inherently slow, could be another explanation for why we have not (yet) witnessed an increase in SDG-related research. However, the myriad of published COVID-19-related research over the last three years casts doubt on this explanation. We are then left with a more sobering but probable explanation, namely that the SDGs are not yet important enough for researchers in these fields. It will probably take institutional changes and incentives to change this situation. Popular university rankings such as the Times Higher Education rankings that now assess universities on their SDG-related research are an example of such incentives. These incentives can also come from the research community itself. For example, my results show that for the nonprofit studies data, the degree of SDG relatedness is positively associated with an article’s citation success for publications appearing after the adoption of the SDGs. This result could be a silver lining showing that while SDG-related research production remains scarce, at least scholars increasingly cite such existing work. I invite researchers to capitalize on this development by producing more SDG-related research. I hope this study inspires researchers to take a more in-depth look at how third sector and public administration research addresses the SDGs, as has been done for research published in business journals (Mio, Panfilo, and Blundo 2020; Pizzi et al. 2020).

One of the limitations of my results is that the automated method of detecting SDGs in the articles might have produced false positives or false negatives. While I cannot rule this out, I safeguarded against these possibilities by using a state-of-the-art tool to label the articles. Another limitation is that I did not analyze the ways in which the identified articles study/relate to the SDGs. As demonstrated by the article by Kagan and Dodge (2023), there is great merit in doing this, and I hope that future research will expand on such SDG-specific literature reviews. Finally, the relative decrease in SDG-related research is associated with a relative increase in other topics. For example, as mentioned by a reviewer, climate change has been replaced by biodiversity risks in policy agendas (WEF 2023). Although the method I used also contains biodiversity-related queries, it might not have detected newly emerging SDG-related topics. Future research could also use unsupervised machine-learning methods to uncover common themes among SDG-related articles. A recent example of such an approach is the study by Ligorio, Venturelli, and Caputo (2022), where latent Dirichlet allocation (LDA) revealed eight different trends in sustainable cities research which the authors then connected to the targets of SDG 11. As such, I hope that the results of this study, as well as the new methods to detect SDGs in text (Meier, Mata, and Wulff 2021; Wulff, Meier, and Mata 2023), inspire many researchers to study the SDGs and its related literature more intensively.

Disclosure statement

No potential conflict of interest was reported by the author.

Notes

1 The Aurora query system is a set of queries developed by the Aurora Universities Network’s “Societal Impact and Relevance of Research" working group. These queries were developed with the goal of identifying SDG-related academic publications included in the Scopus database. See the methods section for more details on how these queries work.