US‐based and international environmental nongovernmental organizations use social media, but few have large audiences online

Social media plays an outsized role in information dissemination, issue mobilization, and public influence. For environmental nongovernmental organizations (“eNGOs”), social media plays a critical role in fundraising and inspiring engaged audiences using rich media and punchy taglines. Yet, there is little to no accounting of whether and how eNGOs have adopted social media, limiting analyses of how eNGOs are leveraging social media to develop effective communication strategies and campaigns. We analyzed the social media presence of over 100 eNGOs. We evaluated which platforms are used by eNGOs and differences between them in terms of their visibility on social media. 88%–97% of the eNGOs in our sample use four out of the five leading social media platforms. There were marked differences across eNGOs and platforms in terms of public visibility, represented by the number of followers. eNGOs focusing on advocacy, aquatic environments, and species conservation had significantly larger numbers of followers. Our findings indicate that conservation can make more headway on newer, more visually‐oriented platforms that also have stronger youth traction. We use our analyses to identify which eNGOs have much larger followings on social media than expected. These leaders can provide lessons on how eNGOs can enhance their social media impact.


| INTRODUCTION
Social media is currently the leading source of news information globally (Allcott & Gentzkow, 2017;Auxier & Anderson, 2021;Shearer, 2018;Watson, 2023).Worldwide, around 50% of individuals have at least one social media account, and, on average, users spend nearly 2.5 h per day on social media (Hirose, 2022).Environmental nongovernmental organizations (henceforth, "eNGOs") now rely on social media to communicate and engage with the public (Halpin et al., 2021;Nulman & Özkula, 2016).Social media platforms are a prominent tool for raising awareness about biodiversity conservation issues, fundraising, activating public campaigns, and broadening the base of supporters for pro-environmental causes (Barrios-O'Neill, 2021; Bergman et al., 2022;Chang et al., 2022a).We set out to describe the landscape of social media usage by eNGOs and to identify examples of effective communication practices in conservation.
Social media research in conservation has shown that these platforms can provide powerful insight on a variety of human-nature interactions.Past research has used social media data, particularly Twitter, to examine threats toward species such as the illicit wildlife trade or belief in anthropogenic climate change (Cody et al., 2015;Jang & Hart, 2015;Leas et al., 2016), reveal information on people's sentiment and the volume of discourse toward species (Chang et al., 2022a;Hammond et al., 2022), track public responses to environmental disasters (Kirilenko et al., 2015;Moore & Obradovich, 2020), segment the public based on their environmental interests (Chang et al., 2022b), evaluate park visitation and human wellbeing (Schwartz et al., 2019;Wood et al., 2013), and mobilize supporters to take action (Coppock et al., 2016;Foos et al., 2021;Kubo et al., 2022).In totality, this body of research speaks to the importance of social media in shaping and reflecting public perceptions of nature and biodiversity conservation.In the broader realm of civic society, researchers have found that social media can enable NGOs to engage in three key activities: raising awareness for issues of concern among the public, mobilizing the public for issue advocacy, and influencing stakeholders such as policymakers to take action (Nah & Saxton, 2013;Nulman & Özkula, 2016).
The dominant media modality of a platform shapes the population of users and best practices for communication.Instagram, for example, largely targets audiences through compelling photographs or videos, while Twitter is largely a platform that allows eNGOs to share textual and visual information to target audiences (Martin & MacDonald, 2020).Neuroeconomic and social science research has found that people process visual information differently than text (Wang et al., 2018); in particular, Casas and Williams (2019) found that imagery was more salient for motivating public action for racial justice in the United States.
Given the large impact that social media can have on conservation advocacy, it is critical that practitioners and researchers know how eNGOs engage the public across social media platforms.Knowing the prevalence of eNGO social media adoption across platforms is therefore foundational to these aims.Additionally, documenting the size of digital audiences for eNGOs through time can allow conservationists to evaluate how potential interests and the universe of supporters changes through time.The closest line of work has been hampered by relatively small sample sizes of eNGOs, such as focal studies of individual organizations or 10 eNGOs, or has focused narrowly on a single social media platform such as Facebook or Twitter with an expanded sample of 100-300 NGOs, though the share of these NGOs that pertain to environmental issues is unclear (Barrios-O'Neill, 2021;Comfort & Hester, 2019;Halpin et al., 2021;Nah & Saxton, 2013;Vu et al., 2021).To the authors' knowledge, there has been no work that has examined the presence of eNGOs across multiple social media platforms-including much more visual platforms such as Instagram, YouTube, and TikTok-and the number of followers that these groups have on each platform.Our work therefore provides a baseline for eNGO social media usage that can guide future projects comparing public outreach for conservation.In this study, we report on data collected from the most prominent social media platforms: Facebook, Twitter, Instagram, YouTube, and TikTok.We address the following questions: (1) Which organizations are using social media?Do organizations have accounts across multiple platforms?(2) How do organizations differ in terms of the size of their reach online, proxied by follower count?(3) Which organizations have a disproportionately large audience on social media?We then conclude our article by discussing ways that our work can inform future digital environmental messaging by establishing a baseline for social media usage that could guide future work such as tracking new audiences for conservation advocates, targeted content analyses for focal eNGOs, or evaluating strategies used by eNGOs who are disproportionately effective at reaching the public online.

| METHODS
We analyzed a sample of 130 eNGOs.To create this set of eNGOs, we initially obtained information on 1500 organizations from Charity Navigator using a Python (v 3.8) script querying the Application Programming Interface that included exception handling, though we did not observe any error codes such as 400 (Bad request) or 404 (Resource does not exist).We are therefore confident that our script robustly queried Charity Navigator for nonprofits with an environmental focus.Our script focused on querying information for nonprofits with the following environmental causes, based on Charity Navigator's list of causes: wildlife conservation, animal rights, welfare, and services, botanical gardens, parks, and nature centers, environmental protection and conservation, development and relief services, foreign charity support organizations, and zoos and aquariums.Charity Navigator has extensive data that has been used in past eNGO research (Armsworth et al., 2012), and focuses on nonprofits that are registered with the US Internal Revenue Service.Of these 1500 organizations, we randomly sampled 300 for further manual review, given an objective of surveying 100 or more eNGOs and an expectation, informed by our initial review of the data, that around 1/3 of the organizations would be clearly related to environmental conservation.Of the 300 randomly selected nonprofits, one of the authors (HK) evaluated which groups clearly had a focus on biodiversity conservation, land stewardship, or climate change mitigation or adaptation.Therefore, as an example, an organization focused on gardening, pet rescue, or veterinary medicine for domesticated animals would not be classified as an eNGO under our definition.We also augmented the sample to include large eNGOs known to have a social media presence such as the Wildlife Conservation Society or World Wildlife Fund.The remaining authors reviewed this categorization of the organizations.Ultimately, this yielded 130 eNGOs in our sample.
For each of the 130 nonprofits, we used structured search queries (e.g., site:instagram.com"Trust for Public Lands") between July 10 and August 1, 2021 to determine which organizations had accounts on each of the five social media platforms.We also recorded information about each organization such as whether or not their account was verified-a marker of public visibility and trust-and the number of followers for their account.We focused on the leading platforms based on their Alexa Rank as of July 2021, namely, Facebook, Twitter, Instagram, YouTube, and TikTok.If an eNGO had an account on a platform, we recorded the number of followers that they had as well.We created a categorization of the National Taxonomy of Exempt Entities (henceforth, "NTEE") codes to the following focal issue areas: advocacy, aquatic, conservation, funding, research, and species.We elected to create an "aquatic" group based on previous research that freshwater and marine conservation groups have historically faced funding shortfalls compared to terrestrial conservation groups.Additionally, creating a "terrestrial" group would have resulted in an overly broad and unbalanced set of groups; therefore, we divided groups that operated primarily in terrestrial habitats into finer groupings such as advocacy or funding or research based on their NTEE code.We defined the geographic scope of an organization based on whether or not they had an international focus or tended to work within the United States, which we coded as US domestic groups, based on the description field of each eNGO returned from their Charity Navigator profile data.Due to missing Charity Navigator data for some of the eNGOs, our sample size for all subsequent analyses was 117 eNGOs.To evaluate whether or not our sample was adequate to capture potential trends in social media usage, we contrasted how our analyses varied when we subset our data to 75 eNGOs and bootstrapped the estimation of the regression model presented below 999 times.This analysis demonstrated that our inferences would not have been changed by relying on a smaller sample size.
We used a mixed-effects linear regression model to evaluate the association between public reach, proxied by the number of followers, and an organization's covariates such as their focal issue area while controlling for factors such as organization size (Equation 1).We included income and financial assets as proxies of organization size, which was a factor that we sought to control for in the regression model.In the regression model, we focused on the four platforms that had high rates of usage: Facebook, Instagram, YouTube, and Twitter.We modeled log-transformed follower count as a function of the social media platform, the geographic scope of an eNGO, the eNGO's focal issue, and the eNGO's financial assets and income.Information on an eNGO's financial assets and income came from rejoining our social media platform dataset with the data from Charity Navigator.Both assets and income were log 10 -transformed.Because these observations were repeated for each eNGO, we also included a random intercept for eNGO.Specifying a random intercept for eNGO also permitted us to account for variation associated with each of the organizations based on their unique online identities.In the equation below, i indexes the eNGOs, and j the social media platforms.
Finally, we examined which eNGOs had more followers than predicted by the model, proxied by a positive random intercept value for the eNGO.An eNGO with a positive random intercept in our model could be considered a positive outlier, positive deviate, or a bright spot for public outreach online.Analyses focused on positive outliers have also been used in other contexts to identify which coral reefs exhibited betterthan-expected health given ecological and social covariates (Cinner et al., 2016), and Post and Geldmann (2018) asserted that positive outliers could provide unique insights for conservation science.For this subset of eNGOs, we generated a plot of their follower count, their model residuals, and their income.

| Do eNGOs use social media?
We observed that across the board, eNGOs are using social media.Almost all of the eNGOs in our sample had accounts on Facebook, Twitter, Instagram, and YouTube (Table 1).In contrast, relatively few eNGOs had accounts on TikTok, which was a much newer platform.
While using social media was ubiquitous in general, there were nevertheless differences in usage.eNGOs T A B L E 1 Proportion of organizations that have an account on each platform.The background color is applied along each column, with darker colors representing a larger value among the values in a column.

Level
The proportion of environmental nongovernmental organizations (eNGOs) that are verified on different social media platforms.
focused on the United States (henceforth, "US-based", or "US" organizations) had lower rates of usage for TikTok and YouTube compared to internationally-focused eNGOs (Table 1).In contrast to international organizations, US-based eNGOs used Facebook, Twitter, and Instagram at similar or higher rates.Among the international eNGOs, those focused on funding and research used social media less.US-based eNGOs used Facebook, Twitter, and Instagram at similar rates, but the research and species-focused US domestic eNGOs used YouTube at higher rates.
To differentiate eNGO usage of social media, we used verification status to differentiate eNGOs.Historically, verification was restricted to highly visible and credible accounts, and was thus a proxy for social media savvy.In our subsequent analyses, we will focus on the four platforms with the highest usage rates: Facebook, Twitter, Instagram, and YouTube.Across these four platforms, only 4% to 57% of eNGO accounts were verified (Figure 1).eNGOs had the lowest rates of verification on YouTube and Instagram.
Comparing different categories of eNGOs, international organizations were verified at much higher rates (Figure 1).Within the international eNGOs, those focused on funding and research had the lowest verification rates.Among the national eNGOs, the research-focused groups had the lowest levels of verification; however, the national organizations had much lower rates of verification on Twitter, Instagram, and YouTube in general.
3.2 | Are some eNGOs more visible on social media than others?
We subsequently sought to examine whether and how eNGOs differed in terms of their public visibility, using follower count across platform as a metric of online reach.The number of followers differed widely across eNGOs and platforms (Figure 2).Table 2 lists the follower counts for the eNGOs with the highest average follower numbers across platforms; these groups generally had many more followers on Facebook and Instagram or Twitter than YouTube.However, to examine which groups had a disproportionately large public reach, we used a mixed effects regression model.
Using a mixed-effects regression model, we evaluated the relationship between public reach, proxied by follower counts, and eNGO covariates such as income, focal issue, and geography.The model indicated that follower counts were the highest on Facebook, followed by Twitter, Instagram, and then YouTube (Table S2).Our regression model exhibited substantial goodness-of-fit, with a marginal R 2 of 0.69 for the predictor variables and a conditional R 2 of 0.83 when the random intercepts for eNGO were also included (Table S4); we also did not observe any substantial violations of model assumptions (Figure S2).International eNGOs tended to have, on average, 12,000 more followers than US-based eNGOs across all platforms (Figure 2; Table S2), and their median income was USD 11,794,791 higher than US-based eNGOs.The eNGOs focused on advocacy, species, conservation, and aquatic systems generally had the largest number of followers (Figure 2).However, drawing on our mixed-effects regression model, organizations focused on funding and research had significantly fewer followers on average (Table S2).Organizations that were verified and had higher income had significantly more followers; intriguingly, the size of an eNGO's financial assets was not clearly correlated in either a positive or negative direction with follower count online (Table S2).

| Which eNGOs have a larger reach than expected?
Identifying which eNGOs have a disproportionately large number of followers is one way to identify successful case studies of eNGOs leveraging social media platforms to engage audiences.Using the mixed effects regression model, we identified eNGOs that outperformed expectations by focusing on the groups that had a positive random intercept.We call these groups "influencers".This produced a total of 12 eNGO influencers (Figure S1; Figure 3), which included groups such as Friends of the Earth, the Surfrider Foundation, the Center for Biological Diversity, and the League of Conservation Voters.
Focusing on these 12 eNGOs, we visualized the relationship between an organization's number of followers and their modeled residuals (Figure 3; Table 3).Across the platforms, the influencers differed in terms of their overall follower count and model residuals.For example, on YouTube, Oceana had both a large number of followers and a large, positive model residual for follower count-both its follower count and model residuals were larger than other eNGOs such as Friends of the Earth or American Bird Conservancy (Figure 3).In contrast, on YouTube, the Rainforest Trust had negative model residuals (around À742), indicating that given its organizational size, focus, and geography, it was predicted to achieve a higher number of followers than its observed value of around 800.On Facebook, the notable influencers were American Bird Conservancy, Environmental Working Group, and the Center for Biological Diversity.Oceana had extensive reach and outperformed the model on YouTube and Instagram.Greenpeace had the largest residuals and large audience sizes on Twitter and Instagram.We examined social media usage by eNGOs as well as differences in their social media visibility.We found that nearly all organizations had accounts on Facebook, Twitter, Instagram, and YouTube, though very few were present on TikTok at the time of analysis.We also found that eNGOs focused on international conservation used social media at higher rates.They also had more status on the platforms, as evidenced by their higher rates of verification.We found that eNGOs differed in terms of their visibility, based on the metric of the number of followers that they had on each platform.Some of this variation could be explained by the different conservation foci of eNGOs.But we were able to identify some organizations that outperformed what would be expected in terms of social media visibility when controlling for their different foci.
The size of an account's audience shapes how many people will directly and indirectly see a particular message.We found that the number of followers ranged from several hundred on more visual platforms such as You-Tube or Instagram to millions on more text-oriented platforms such as Facebook and Twitter.eNGOs focused on advocacy and species had larger number of followers.One surprising finding from our study was that aquatic eNGOs also had significantly more followers than other The observed follower count, modeled residual, and income level for the 12 eNGO influencers, or those organizations that had positive random intercepts.The abbreviations for the groups are described in Table 3.  Armsworth et al. (2012) noted that aquatic eNGOs typically received fewer public contributions.Moreover, work focused on a large set of eNGO followers on Twitter found that terrestrial or climate change discourse was discussed nearly twice as much as aquatic or marine issues (Chang et al., 2022a).Therefore, while aquatic eNGOs in our sample may have large digital audiences, there are still opportunities for aquatic and marine issues to better resonate with the public.

| Who could use this information and how?
Our results are most relevant to communication teams within eNGOs, but are also likely to be of interest to researchers interested in understanding social media use patterns or using social media communication to draw inferences about motivations surrounding conservation.One of the primary goals of communication teams in eNGOs is to communicate with the largest audience as possible (Veríssimo, 2019).Our results can be used to benchmark the visibility of their social media strategies compared to other organizations sharing their mission area and geographic scope (Nah & Saxton, 2013;Nulman & Özkula, 2016).They can use our results to identify where potential gains in terms of social media visibility seem most promising, whether by integrating a new platform or investing more in content development for existing platforms.
Our findings on positive outliers for audience size can lead to future analyses that compare and contrast organizational efficacy at building their social media presence.For example, Oceana and Greenpeace emerged as being among the most effective social media communicators within our sample.One striking aspect of the strategies pursued by these two organizations is their rich engagement in the highly visual platforms of Instagram and YouTube.Indeed, Yeomans (2018) lauded Greenpeace's aggressive social media campaigns across multiple platforms which achieved high public visibility and impact.Leaders in social media communication in conservation like this provide models to other conservation organizations.Steps they have taken may suggest ways other organizations can reach and mobilize large numbers of people with their conservation messaging.
Researchers seeking to understand conservation communication or to use social media to draw inferences about conservation motivations can use our information to stratify organizations for sampling and identify priorities for future applied work.Examining whether and how the 12 eNGO influencers, identified through our mixed effects regression model as the positive outlier eNGOs, differ in their messaging strategy can inform social media campaigns.For instance, practitioners and researchers could track trends in public interest in conservation across time and different platforms to evaluate the impact of messaging campaigns or focal incidents of interest (Correia et al., 2021;Ladle et al., 2016).For particular issues of interest, more focused analyses such as content analyses of social media posts using computational approaches could offer deeper insight on how to engage the public for a given awareness or advocacy goal.Historically, there have been concerns that social media may be limited beyond US audiences, which would make social media campaigns of questionable efficacy outside of the US; however, our results showed that international eNGOs had on average larger digital public audiences, gesturing to the powerful role that social media can play globally for conservation engagement.

| Priorities for future work
The ecosystem of social media platforms is highly dynamic, and the relevance of social media platforms can fluctuate markedly through time (Chang et al., 2023).eNGOs must be nimble and responsive to societal trends in social media use.They need to shift their communication strategies within social media platforms, and build out presence on new platforms to continue to engage the public online.More agile communication practices would benefit conservation communications in general, given the fleeting nature of public attention to striking biodiversity events (Jari c et al., 2023).While it is beyond the scope of work in this manuscript, future work could identify and contrast social media messaging practices by eNGOs that promote positive pro-conservation outcomes (Bergman et al., 2022).
Among the sampled eNGOs, usage of Facebook was substantially higher than Instagram, Twitter, and You-Tube.In contrast, the number of groups using TikTok was so low at the time of data collection that we could not include this platform in statistical analyses.Yet, TikTok was by some accounts one of the fastest growing social media platforms in recent years (Statista, 2023).Moreover, eNGOs were verified on Instagram, YouTube, and especially TikTok at much lower rates than Facebook and Twitter.The dynamic nature and increasing significance of social media platforms as a communication and engagement tool also implies the importance of tracking changes in public engagement through time on dominant platforms (Ladle et al., 2016;Veríssimo, 2019).Such changes are indicative of how conservation messaging strategies develop, are disseminated, and whether or not they are impactful at reaching a diverse set of stakeholders, given the broad reach of social media platforms globally, which numbered in the hundreds of millions to billions of users at the time of our analysis (SI, Bergman et al., 2022;Chang et al., 2022b).Future projects can also track how eNGOs reach new audiences through time, which would necessitate tracking followers through time, identifying new followers at regular time intervals, and evaluating how different types of social media content may reduce user engagement or attract new audiences to engage in conservation.
Image-and video-based platforms are increasingly popular and are a leading channel for youth engagement (Hirose, 2022).Platforms such as YouTube have an extremely high engagement by the public and are critically important in shaping worldviews (Hosseinmardi et al., 2021).In fact, Freund et al. (2021) noted that YouTube was a leading source of web traffic for orangutan conservation NGOs.Thus, to connect with younger stakeholders, it is important that biodiversity conservation advocates and researchers track and cultivate a presence on visual platforms.
Our work demonstrates that social media is an important and widely used venue for communications by eNGOs.Nevertheless, there were differences across platforms and categories of eNGOs in terms of social media usage and visibility.We demonstrate one method to identify influencer eNGOs.These influencers have valuable lessons to teach the conservation community as a whole about digital public engagement.Future research can build on the work here to track changes through time and expand on the scope of features that shape digital conservation communications and engagement.As social media continues to grow in importance as a source of information and venue for communication, achieving pro-conservation success at scale requires continued and effective social media approaches.
The distribution of followers for environmental NGOs across social media platforms, separated by organizations that are international or domestic to the USA.
The top 20 organizations based on average follower count across platforms.
T A B L E 2 Abbreviations for the 12 positive outlier eNGOs. of groups; this result runs counter to patterns from charitable giving.In the context of charitable giving, types