Introduction

Climate change is one of the principal challenges facing humanity. Heat retention due to greenhouse gasses -byproducts of the use of fossil fuels- has led to a progressive increase in the temperature of the planet with varying consequences, such as forest fires, droughts, excessive rain, and rising sea levels (^{Gates, 2021}). An important number of actions that contribute to climate change are related to daily habits such as using a car rather than walking or using public transit, the wasteful use of electricity, and the consumption of red meat (^{Poore & Nemecek, 2018}; ^{Sandström et al., 2018}).

Collective action can be understood as behavior undertaken by individuals who desire to achieve a common goal; it is thought that the best way to achieve the results that these initiatives seek is through organized participation (^{Brunsting & Postmes, 2002}). One theory that effectively explains collective action is the Social Identity Model for Collective Action or SIMCA. This model argues that collective action is made up of four psychological motivators: moral conviction, emotions, social identity, and perceived group efficacy (^{Thomas et al., 2020}; ^{Van Zomeren, 2013}). As regards environmental issues, there is evidence that variables from the SIMCA model predict action aimed at combating climate change (^{Bamberg et al., 2015}; ^{Van Zomeren et al., 2019}). Therefore, it is important to connect collective action to the environment; given the complexity of the problem, it will require collective action to stabilize our climate (^{Brechin, 2016}; ^{Higham et al., 2019}).

The use of social media for political means -also known as cyberactivism- can be understood as the use of digital platforms as ways to gain information related to different social causes, as well as participate in them (^{García-Galera et al., 2014}). Social media also presents content that mobilizes sizable groups of people to achieve common goals (^{Valenzuela, 2013}). This type of activism, frequently used in social media such as Twitter and Facebook, has changed the way in which collective action is coordinated and promoted (^{Greijdanus et al., 2020}). Ample research on the connection between social media use and collective action exists (^{Cmeciu & Coman 2016}; ^{Gerbaudo, 2016}). These studies have found that these types of digital activism, such as sharing information and expressing political views, are associated with participation in collective action, such as participating in a protest or signing an online petition (^{Alberici & Milesi, 2018}; ^{Barth et al., 2015}; ^{Halpern et al., 2017}). Specifically, digital collective action can be understood as Internet-based proposals related to a call to authorities for attention, as well as online cooperation with organizations in order to achieve different goals (^{Barth et al., 2015}; ^{Landmann & Rohmann, 2020}).

Regarding the use of social media in the fight against climate change, these platforms have been identified as informing users of the magnitude of the problem as well as promoting the development of environmental initiatives and the intent to maintain them over time (^{Anderson, 2017}; ^{Ballew et al., 2015}; ^{Liu, 2016}; ^{Şen & Şen, 2016}; ^{Wamuyu, 2018}).

Studies exist that analyze the SIMCA model variables as effective mediators between social media use and collective action (^{Chan, 2017}; ^{Odag et al., 2016}), as well as research whose objectives are to analyze the incentives behind participating in climate change collective action using the SIMCA model (^{Brügger et al., 2020}). However, after a systematic review of databases such as scopus and ebsco, we have not identified any study to date that examines the SIMCA variables as mediators of the effect of social media use and collective action against climate change.

In view of the circumstances surrounding climate change today, research that connects collective action to climate change is necessary (^{Bamberg et al., 2015}). Given the importance of social media as an element associated with collective action, we believe that it is important to contribute to this area of knowledge. In addition, despite Peru’s high level of vulnerability to climate change (^{Ministerio del Ambiente, 2019}), we have not identified any study to date that examines climate change collective action. We believe it is important to endeavor to fill this gap.

Given studies that demonstrate the impact of the use of social media on the SIMCA variables (^{Brügger et al., 2020}; ^{Chan, 2017}; ^{Odag et al., 2016}; ^{Ruiz-Dodobara et al., 2021}) and their connection to collective action (^{Bamberg et al., 2015}; ^{Van Zomeren et al., 2019}), this study aims to analyze if the SIMCA variables can explain the relationship between the use of social media and collective action against climate change among a group of Peruvian university students. In this work, we conceptualize climate change collective action in two ways: daily habits that favor protecting the environment, and online collective action.

Due to the prevalence of their use of social media in Peru, this study focuses its analysis on university students (^{Instituto Nacional de Estadística e Informática, 2020}). In addition, many studies indicate that young people use social media to organize, draw attention to issues, encourage participation in collective action, and the political use of these platforms tends to be a good predictor of participation in collective action (^{Hoffmann & Lutz, 2019}; ^{Lee et al., 2017}; ^{Ruiz-Dodobara et al., 2021}).

SIMCA variables and climate change collective action

As regards the variables that make up the SIMCA model, there is evidence that these factors predict involvement in climate change collective action. ^{Brügger et al. (2020)}, in a study of 4 057 young Swiss, found that identification with those who participated in protests and the perception of protests’ efficacy to achieve results and create issue consciousness were significant predictors of the decision to take collective environmental action. In their study of MTurk users, ^{Van Zomeren et al. (2019)} found that variables connected to participation in climate change collective action were moral conviction, anger, identification with an opinion group, and the perception of participative efficacy.

A study of 562 adolescents and young adults in Germany found that the participation in the Fridays for Future environmental protests could be predicted based on identification with other environmental activists, personal value norms, the perception of friends’ pro-environmental activism, and anger at government action on environmental and economic policy (^{Wallis & Loy 2021}). ^{Schmitt et al. (2019)} carried out research on American citizens, finding that identification with climate activists predicted pro-environmental actions. Also, in studies that took place with young people in Germany and Australia, ^{Jugert et al. (2016)} found that the manipulation of collective action had an indirect impact on pro-environmental behavior, through an increase in collective action and personal efficacy. ^{Furlong and Vignoles (2021)} in their study of 203 participants who were associated with the environmental movement “Extinction Rebellion” (XR) in the UK found that the perception of group efficacy was a significant predictor of the intent to participate in XR environmental collective action. At the same time, they found a pathway from identification with humanity to collective action and future intent to participate, via participative efficacy and identification with the “XR” movement. They also found that moral conviction had an impact on collective action, via anger and identification with the XR movement. Results also demonstrated that the perception of injustice predicted action and future intent to participate via anger and identification with the XR movement.

Given these prior studies, we propose the following general hypothesis:

More specifically:

Methodology

Analytical Plan

The Spearman correlation matrix was calculated using the variables described above. We then developed four multiple regression models, which had social media use and the SIMCA model variables as independent variables, with collective action against climate change. Models differed based on the type of identity that was analyzed (identification with victims or activists) and the dependent variable (habits or online collective action against climate change).

For the multiple mediation study, we used four models in which only one variable was the independent variable (social media use), four mediating variables that stem from the SIMCA model, and collective action against climate change as the dependent variable. Similarly to the multiple regression models, the multiple mediation models differ regarding the type of identity that was analyzed as well as the dependent variable (Figures 1 to 4).

Mediational Model 1. IV: Social media use, DV: Habits combatting climate change. Mediational Variables: Anger, moral convictions, identification with victims and group efficacy. The upper side of the direct line between social media use and habits indicates the direct effect and the lower side of the line shows the total effect. The results correspond to non - standarized estimates.

** p < 0.1, *** p < 0.001.

Source: Created by the authors.

Figure 1 

Mediational Model 2. IV: Social media use, DV: Habits combatting climate change. Mediational Variables: Anger, moral convictions, identification with activists and group efficacy. The upper side of the direct line between social media use and habits indicates the direct effect and the lower side of the line depicts the total effect. The results correspond to non - standarized estimates.

** p < 0.1, *** p < 0.001.

Source: Created by the authors.

Figure 2 

Mediational Model 3. IV: Social media use, DV: Online collective action against climate change. Mediational Variables: Anger, moral convictions, identification with victims and group efficacy. The upper side of the direct line between social media use and online collective action indicates the direct effect and the lower side of the line depicts the total effect. The results correspond to non - standarized estimates.

** p < 0.1, *** p < 0.001.

Source: Created by the authors.

Figure 3 

Mediational Model 4. IV: Social media use, DV: Online collective action against climate change. Mediational Variables: Anger, moral convictions, identification with activists and group efficacy. The upper side of the direct line between social media use and online collective action indicates the direct effect and the lower side of the line depicts the total effect. The results correspond to non - standarized estimates.

** p < 0.1, *** p < 0.001.

Source: Created by the authors.

Figure 4 

Finally, we tested the goodness of fit for each multiple mediational model using the Chi squared and Chi ratio indexes, squared degrees of freedom, the CFI, TLI, RMSEA and SRMR. For the multiple mediation study, we used a structural equation model using the AMOS v 26 program (^{Arbuckle, 2019}; ^{Preacher & Hayes, 2004}).

Results

We found statistically significant correlations between social media use and the SIMCA variables of anger (r = .43, p < .001), moral conviction (r = .40, p < .001), identification with victims (r = .26, p < .001), identification with activists (r = .34, p < .001) and group efficacy (r = .15, p < .05). We also found a significant relationship regarding social media use and habits combating climate change (r = .29, p < .001) and online collective action against climate change (r = .50, p < .001). In addition, we found that both variables related to collective action (habits and online collective action) were statistically significant with each of the mediating variables.

We then examined four possible regression models using habits to combat climate change and online collective action against climate change as dependent variables. In all models we controlled for sex and age. We did not observe multicollinearity between the variables in any of the models.

In Model 1 (Table 1), one can see that only social media use and group efficacy were significant predictors of habits to combat climate change. This model explains (R²) 18% of the variance in the habits to combat climate change. While in Model 2, only group efficacy predicted the dependent variable. Both models explained (R²) 18% of the variance of the habits against climate change.

Regarding Model 3, one can see that all variables except identification with victims were significant predictors of online collective action against climate change (Table 1). This model explains (R²) 43% of the variance in digital collective action.

Finally, Model 4 demonstrates that, with the exception of moral conviction, all variables are significant predictors of online collective action against climate change. This model explains (R²) 47% of the variance of this variable.

Mediational analysis

As with the regression analysis, in the mediational analysis we controlled for both sex and age. First, in Model 1 (β = .056, SE = .039, 95% CI [.017, .157] y 2 (β = .061, SE = .040, 95% CI [.020, .182]) we find that group efficacy totally mediates the relationship between social media use and habits against climate change ( Figures 1 and 2). Both Model 1 (χ2 = 1 486.34; χ2/GL = 2.71; CFI = 0.93; TLI = 0.90; RMSEA = 0.05; SRMR = 0.07) and Model 2 (χ2 = 1 691.12; χ2/GL = 2.74; CFI = 0.94; TLI = 0.92; RMSEA = 0.05; SRMR = 0.08) present acceptable goodness of fit measures.

As regards Model 3, we find that moral convictions (β = .128, SE = .278, 95% CI [.030, .219]; anger (β = .128, SE = .085, 95% CI [.048, .204]) and group efficacy (β = .062, SE = .049, 95% CI [.021, .124]) partially mediate the relationship between social media use media and online collective action. Model 3 also has an acceptable goodness of fit (χ2 = 1 694.14; χ2/GL = 2.74; CFI = 0.93; TLI = 0.916; RMSEA = 0.08; srMr = 0.07).

On the other hand, we observed in Model 4 that anger (β = .097, SE = .060, 95% CI [.001, .186]), identification with activists, (β = .184, SE = .062, 95% CI [.079, .324]) and group efficacy (β = .056, SE = .035, 95% CI [.007, .131]) partially mediate the relationship between social media use and online collective action. It is important to highlight that this model presents the best goodness of fit of all models (χ2 = 1 678.37; χ2/GL = 2.57; CFI = 0.93; TLI = 0.92; RMSEA= 0.06; SRMR = 0.07) (Figure 4).

Discussion and conclusion

This study found that social media use, with the exception of Model 2 which had a marginal level of significance (p = 0.052), was a significant predictor of climate change collective action both as regards habits and online collective action (Table 1). This finding coincides with many studies that find a relationship between online activism and collective action linked to political issues (^{Alberici & Milesi, 2018}; ^{Cmeciu, 2016}; ^{Gerbaudo, 2016}). The possible explanation for this link lies in the exposure to political stances in social media such as Twitter and Facebook, which can lead to involvement in public collective action (^{Halpern et al., 2017}). According to ^{Wamuyu (2018)} the use of social media as a way to take care of the environment could impact community involvement and intentions to perform pro-environmental online activism. In addition, our findings coincide with other studies that identify a relationship between the use of social media for political activism among young people and participation in collective action (^{Hoffmann & Lutz, 2019}; ^{Lee et al., 2017}; ^{Valenzuela et al., 2013}).

Regarding moral convictions, we found that they were only significant predictors of digital collective action against climate change when the regression model (Model 3) included identification with those impacted by climate change (Table 1). These results align with those of ^{Van Zomeren et al. (2019)}, who found that moral conviction predicted participation in climate change collective action. This in turn is related to mediational Model 3, in which moral convictions mediates social media use and online climate change collective action. This suggests that an online discussion that promotes reflection would activate moral norms, which could lead to participation in collective action (^{Alberici & Milesi, 2018}), in this case against climate change.

As regards anger derived from the perception of injustice, we found that this was a good predictor of digital collective action (Table 1). Results coincide with ^{Van Zomeren et al.’s (2019)} findings, where they found that anger predicted collective action in defense of the environment. The explanation for the mobilizing impact of anger is due to frustration many feel in the face of government environmental policy, such as described in the studies of ^{Furlong and Vinogles (2021)} y ^{Wallis and Loy (2021)}. These studies find that this emotion plays a significant role in participation in collective action.

We also found that anger mediated the relationship between social media use and digital collective action against climate change (Mediation Models 3 and 4). These findings could be due to the fact that much of the content that is shared regarding environmental issues on social media entails descriptions of events and/or practices that impact climate negatively (^{Cody et al., 2015}; ^{Mavrodieva et al., 2019}); therefore, these events activate emotions such as anger, contributing to an impact on participation in collective action. At the same time, we must acknowledge that social media can create polarization, as these networks tend to show users information that is consistent with their opinions as well as individuals who think like them (^{Törnberg, 2018}; ^{Yarchi et al., 2021}). Therefore, the use of social media could radicalize postures that were initially indignation in the face of weak support for climate protections. In turn, this indignation could unleash a wave of emotions such as anger and the subsequent desire to participate in climate change collective action.

It is important to highlight that anger did not mediate the relationship between the social media use and habits to combat climate change (Mediation Models 1 and 2). Anger, when associated with the consumption of information on social media, does not result in daily routines such as saving energy and reducing the consumption of meat; however, it does have an impact regarding the intent to participate in collective action online (Mediation Models 3 and 4). Thus, as regards habits for protecting the environment, we do not find emotional involvement to be a significant predictor of action; the opposite is true when individuals express their frustration with politicians (^{Wallis & Loy, 2021}), experience indignation due to climate change, or participate in protests (^{Furlong & Vignoles, 2021}).

In relation to identity, we found that identification with activists (Regression Model 4) was the only significant predictor of collective online action. This finding aligns with those of other studies that find identification with individuals who protest against climate change predicts participation in collective action (^{Brügger et al., 2020}; ^{Van Zomeren et al., 2019}; ^{Wallis & Loy, 2021}). At the same time, we found that this type of identity was a significant mediator of social media use and collective digital action against climate change (Mediational Model 4). On social media, people often associate themselves with individuals and/or groups who have similar interests and political positions (^{Yarchi et al., 2021}).

We also found that digital spaces to exchange information, debate, etc. could have an impact on the political identity of users and might even result in the intent to participate in collective action (^{Alberici & Milesi, 2018}). As regards this study, since climate change is intricately linked to the political world, the use of social media should reinforce political identification with activists, which might in turn impact participation in collective action.

Identification with those impacted by climate change did not predict either habits or online collective action (Table 1), nor did it mediate the relationship between social media use and both types of climate change collective action (Mediation models 1 and 3). These findings are interesting as information about victims of climate change are often shared online (^{Mavrodieva et al., 2019}), which could foster empathy. In addition, other studies have found that in the face of disasters, those who identify with the out-group are often those mobilized to help (^{Vezzali et al., 2015}). ^{Drury et al. (2019)} argues that in the presence of a common disaster, individuals experience what they call a perception of a “shared destiny,” which impacts identification with victims of the event and fosters helping others. Based on our findings, despite the fact that climate change can be conceptualized as an event that will have a disastrous impact on all of humanity, this is not related to the perception of a shared destiny that would lead individuals to closely identify with those who are most impacted by climate change.

Regarding group efficacy, we find that in all regression models this variable is a good predictor of habits and online collective action against climate change. These findings are consistent with prior studies (^{Brügger et al, 2020}; ^{Furlong & Vignoles, 2021}; ^{Jugert et al., 2016}; ^{Van Zomeren et al., 2019};). We also find that in all mediation models efficacy mediates the relationship between social media use and both dependent variables.

As we’ve mentioned previously, on social media individuals tend to connect with those individuals or groups with whom they have a shared affinity (Yarchi, 2020); if social media users observe that users who are similar to them use these platforms to achieve success with certain initiatives (such as habits or digital collective action) against climate change, with a base in vicarious learning (^{Bandura, 2002}), we can infer that this information would impact the efficacy of these individuals and lead them to perform collective action.

Another possible explanation is that the high number of online discussion leads to collective efficacy (^{Alberici & Milesi 2013}) and makes it more likely that individuals will participate in digital or inperson political activities (^{Vaccari et al., 2015}). Therefore, higher use of social media as a way to debate and exchange information regarding climate change may impact the belief in individual efficacy in participating in political processes. In turn, this higher efficacy may lead individuals to engage in collective action.

In conclusion, the principal contribution of this study is that it explains the relationship between the use of social media and collective action through various SIMCA model variables, such as group efficacy in the case of habits for combating climate change, as well as nearly all the SIMCA variables regarding online collective action. We found that the mediators and predictors of both dependent variables were different, which indicates that a differential focus at the moment of promoting both types of collective action will be necessary. We also found that the use of identification with climate change victims might not be an effective strategy for the promotion of environmental action, as compared to identification with individuals who actively take action against climate change. Future studies should continue to investigate this area.

This study is not without limitations. It was based on a cross-sectional design using a sample of university students, which limits our conclusions in terms of causality as well as to this specific population.