Identification With All Humanity Predicts Prosocial and Political Action Intentions During COVID-19

Introduction

The COVID-19 pandemic created a health and economic crisis to everyone, but its impact was especially strong on some vulnerable groups who lacked access to adequate health care and hygiene products, and their livelihood was directly threatened. The immediate problems caused by the pandemic could not be solved individually, but it required broad cooperation. Thinking and acting collectively became crucial to survive and keep societies functioning (Templeton et al., 2020; van Bavel et al., 2020), however, it is not self-evident, when people choose to act on community interests over self-interests (Tse et al., 2021). In this article, we investigate people's motivations for community-oriented actions and their willingness to stand up against authorities for vulnerable groups. It is important to find ways how people can be motivated to step in solidarity with each other, and to challenge political decisions that are not serving their communities.

During the pandemic, both benevolent support and political action became valid modes of engagement. Some people engaged in prosocial behaviors: they joined initiatives of charitable organizations, offering donations or voluntary work to improve the situation of specific groups (for example, the elderly, poor people, and people who have lost their jobs). At the same time, people also engaged in political actions in connection to the pandemic, as they tried to put pressure on authorities to represent the best interests of the community (and at a later stage of the pandemic, to protest against restrictions). In the Hungarian context, several petitions were initiated by NGOs during the second wave of the pandemic to call for help for people threatened by the economic consequences of the pandemic, and to urge authorities to provide financial and professional support for people infected by COVID who could not receive proper treatment due to the overload of hospitals (Balogh, 2020).

Even though the two types of action, prosocial, and political action can be clearly distinguished, they also share some commonalities. Benevolent prosocial behavior usually happens to improve the situation of a specific group, but it does not necessarily question the social structure as a whole. In contrast, political action aims to change an unequal system, so status relations between groups would change (Thomas and McGarty, 2018; Louis et al., 2019). However, as both behaviors can improve the situation of a low status group, their motivations can overlap (Fattori et al., 2015; Kende et al., 2017). For example, both types of action were connected to identification with relevant and meaningful groups, and empathy toward them (see for example Selvanathan et al., 2018; Tropp and Barlow, 2018). Therefore, prosocial and political action can be distinct but equally adequate reactions to the same societal crisis (Lantos et al., 2020).

When people experience that their fellow ingroup members face grievances, based on their level of identification with the ingroup, they will experience the grievance as their own and they will feel empathy, so they become willing to act on their behalf. The social identity theory claims that people manage their self-esteem based on their group memberships by engaging in ingroup favoring behavior (Turner et al., 1979). Collective action for social change is such a behavior, where people assess their group's position as illegitimate, so they want to improve the status of the ingroup (Reicher, 1996; van Zomeren et al., 2008). Identification with a relevant ingroup can be a mobilizing factor also when it comes to prosocial behaviors on behalf of specific groups (Dovidio and Banfield, 2015), or engaging in solidarity and helping behavior during a crisis (Drury and Tekin Guven, 2020; Yzerbyt and Phalet, 2020; Bowe et al., 2021). In summary, social identities are basic predictors of both prosocial behavior and political action for social change.

We feel the most empathy toward ingroup members and we act in solidarity with them (Stürmer et al., 2006). It follows that we feel less empathy and offer less support for members of other groups. As groups represent different levels of inclusiveness, the treatment of ingroup and outgroup members depend on the perception of group boundaries and definitions of belonging. In case of a global crisis, such as the COVID-19 pandemic, inclusive identities can be essential for coordinated action and for pro-social behavior to combat a crisis that affects all of humanity, but especially members of vulnerable groups.

As each country had to find its way of coping with the pandemic on the national level, national identities became important for how individual citizens dealt with the crisis, and how people criticized authorities for the measures they have taken. Identification with the national ingroup can have positive consequences for the ingroup: when people feel attached to their group, they are easily mobilized based on this feeling of togetherness to act for their fellow group members. For example, national identity predicted pro-environmental collective action intentions among New Zealanders, as pro-environmental attitude was part of their national identity, therefore acting on that represented the interest of the ingroup (Milfont et al., 2020).

At the same time, based on social identity theory, ingroup- serving behavior is frequently accompanied by outgroup derogation and discrimination. High national identification was consistently related to prejudice and discriminatory tendencies, and even hostile collective action toward outgroups (see for example: Feather, 1994; Esses et al., 2001; Druckman, 2006; Pehrson et al., 2009; Smeekes et al., 2011; Spiegler et al., 2021). Furthermore, high national identification correlates with the so-called glorification of the ingroup, which is an overly positive, and less critical perception of the nation (Roccas et al., 2006). As a consequence, a high national identification can make people less sensitive to the disadvantages of vulnerable outgroups, and we argue that it can also be associated with less critique of pandemic- related measures of governments.

National identification was found as a more limited predictor of helping behavior in the context of the pandemic, compared to identification with all humanity (Barragan et al., 2021). At the same time, an international survey highlighted the potential of national identity in predicting cooperation in terms of public health behaviors during COVID (van Bavel and Boggio, 2020; van Bavel et al., 2022). These ambiguities could be a result of the different contents of national identity, which can be more or less inclusive. For example, Dutch national identification predicted supportive political action intention on behalf of immigrants (Verkuyten and Martinovic, 2015), whereas Hungarian national identification prevented people from supportive collective action for minorities (Kende et al., 2018). A survey study in the context of the pandemic highlighted that national identification in the UK predicted helping behavior toward proximal others, but not toward distal others (Vignoles et al., 2021). An inclusive national identity may have the potential to mobilize for wide cooperation among citizens of a country, while an exclusive national identity can be a barrier to solidarity.

However, identities are not fixed and stable. People can re-categorize members of out-groups as members of a superordinate ingroup category (Gaertner et al., 1999, 2016). Recategorization makes people more emphatic, more cooperative and more supportive of newly included group members (Dovidio and Banfield, 2015). In line with this, we can define communities we belong in an inclusive way. However, the level of identification is not stable either, it can change over time and be influenced by contextual factors. For example, a study showed that self-esteem influenced collective narcissism (an uncritical and biased form of national identification) with time (Golec de Zavala et al., 2020).

All of humanity represents one of the most inclusive social categories. The concept of identification with all humanity (McFarland et al., 2012) suggests that people fully expand the circle of people they care about, in contrast to other identities that restrict the boundaries of the ingroup. Identification with all humanity means belonging to a global community where there is no division between in- and outgroups potentially, motivating prosocial behavior toward all people regardless of other group memberships.

Indeed, identification with all humanity was found to be a powerful predictor of prosocial behavior toward a number of outgroups. It was found to be related to higher empathy (Hamer et al., 2019) and it correlated with higher supportive behavior (both behavioral intention and actual donation) for disadvantaged groups in general (Reese et al., 2015), more welcoming attitudes toward asylum seekers (Nickerson and Louis, 2008), higher support for refugees during the refugee crisis (Bassett and Cleveland, 2019), higher humanitarian helping for victims of natural disasters, poverty, and wars (Sparkman and Hamer, 2020). In addition, there is evidence that global identity made people more cooperative in social dilemma situations (Buchan et al., 2011).

Even in the context of the pandemic, identification with all humanity was found to be a strong predictor of helping strangers, in contrast to national identity, that was a significant but weaker predictor (Barragan et al., 2021). Moreover, identification with all humanity gained a special relevance in the humanitarian crisis of the pandemic, similarly to the environmental crisis, where the problem affects the human community as a whole (Reese at al., 2020). In line with the social identity model of pro-environmental action (Fritsche et al., 2018), in a theoretical paper, human identification as a social identity was highlighted as relevant factor to influence people's willingness for prosocial and political action during the global challenge of the pandemic (Reese et al, 2020). Although identification with all humanity was mostly connected to prosocial behavior in previous research, a recent study highlighted that it has also a connection to collective action for social change (Römpke et al., 2019). This study demonstrated that identification with all humanity did not only make people sensitive for the disadvantage of specific groups, but also for global problems, for example, the environmental crisis. In this experimental study, authors also showed how the quantity of international contact led to higher human identification which led to more pro-environmental collective action and higher global responsibility. Again, this research supported the notion that the level of identification is dependent on specific factors, and a change in identification can lead to a change in behavior.

In line with the logic of the above-mentioned study, we argue that collective, political action is a relevant concept not only for understanding intergroup conflicts, but also when it comes to responding to a humanitarian crisis, such as the COVID-19 pandemic. Therefore, we conducted a study with the assumption that identification with all humanity would not only motivate people for prosocial behavior, but also for collective/political action, challenging the authorities for social change. In addition, our study had a repeated measures design which allowed us to test if relationships between human vs. national identification, empathy and action intentions change with time, and within individual respondents.

In the Hungarian context, national identity is highly politicized and exclusive, as the Hungarian government uses national identity as a central reference to address their supporters, and to define their opponents, depicted as being outside the nation and threatening the national interest (Kim, 2021). Furthermore, the government used national ingroup and outgroup divisions in their rhetoric about the pandemic. For example, when the first cases of infection were identified in Hungary, the narrative of the government emphasized that the virus was “brought” to the country by foreigners (Vásárhelyi, 2020). They also communicated that the opposition, or critical actors of the government were against the government's vaccination program, therefore against Hungarian people in general (Verseck, 2021).

The Hungarian government introduced restrictions quickly in the first wave of the pandemic in the spring of 2020. They declared a state of emergency, they introduced a curfew (allowing going out for only specific reasons) and introduced different time shifts for the elderly in the shops. These measures led to a low number of cases in the first wave, at a time when the situation was critical in Western and Southern Europe. However, the strategy of the government completely changed by the second wave, in the autumn of 2020, where the previous caution and preventive measures were kept only partially (for example, wearing masks remained obligatory). The number of infections grew rapidly, which made the situation harder for the healthcare system to handle. Finally, the restrictions were applied again from November 2020. While in many European countries, political protests were more often against the strict regulations and lockdowns in response to the growing number of infections (Baker, 2021), in Hungary, criticism about the government was more about the insufficient support of the healthcare system and those suffering the economic and health consequences of the virus.

Additionally, the Hungarian government was also criticized by the Hungarian opposition parties and international observers alike (including politicians in the European Union) for using the state of emergency for centralizing their power even more (Guasti, 2020; Löblová et al., 2021). The COVID- related measures of the government included forbidding public demonstrations, which made it more difficult for citizens to express dissent. At the same time, other forms of political participation (such as signing online petitions or contacting politicians) and pandemic-related donation and volunteerism increased after the outburst of the coronavirus according to survey on a Hungarian representative sample (Mikecz and Oross, 2020). An international survey study also supported that solidarity increased among Hungarians during the first wave of the pandemic (Voicu et al., 2021). Furthermore, the higher their perceived burden caused by the pandemic was, the higher intention Hungarian people showed to engage in political activism (Bartusevičius et al., 2021). However, another piece of research that investigated solidarity among Hungarians longitudinally, found that solidarity shrank by the second and third wave of the pandemic (Sik and Zakariás, 2021). In summary, Hungarian citizens who generally show low political participation and solidarity (Karp and Banducci, 2007; Simon, 2014), seemed to be mobilized by the shocking outburst of the virus, but this elevated participation decreased over time.

In the politically polarized Hungarian context, we expected that a low identification with the nation may also mean a counter-identification with the government, expressing an oppositional political stand. Political action intention during the COVID-19 pandemic in Hungary could be the result of criticism and discontent of the government. We expected that national identity would be related to accepting government actions during the pandemic, but even in the absence of actual support of the government, it may indicate some loyalty and trust in the way the Hungarian government and authorities dealt with the COVID situation. In sum, we expected that national identification may prevent people from being critical with the government, or put differently, we expected that low national identification (or counter-identification with the nation) would lead to political action intention challenging the authorities for the interest of the community. Except for our expectation with Hungarian national identity, a politicized and context-dependent variable, our expectations related to human identity was more general, and less specific of the Hungarian context.

Overall, based on their difference in identity inclusiveness, we expected that identification with all humanity would be a stronger predictor of empathy (Hypothesis 1) and prosocial action intentions (Hypothesis 2) toward vulnerable groups, compared to national identity. In addition, based on previous research about the similar motivations of prosocial and political actions in a crisis (see for example Kende et al., 2017; Lantos et al., 2020), we expected that identification with all humanity and empathy would not only be related to prosocial but also to political action intentions (Hypothesis 3). Taking the specific Hungarian context into consideration, we expected that lower national identity would be associated with higher political action intentions (Hypothesis 4). In other words, lower identification with the national ingroup was expected to predict higher intentions to challenge political decisions (even when actual self-reported political commitment is controlled for). Furthermore, we expected the above relationships of our model to persist over time, but at the same time, changes in identification, empathy, and behavior intentions within individuals may also occur with time (Hypothesis 5).

Methods

Participants and Data Collection

Data was collected with an online survey at four timepoints in Hungary, the questionnaire was in Hungarian. The study was conducted with the ethical permission of Anonymous University. We used convenience sampling, participants were recruited online, on social media platforms and different media outlets. The first data collection was at the start of the first wave of the pandemic in early April 2020 (N_t1 = 6,140), the second was during the first lockdown in May 2020 (N_t2 = 1,515) when the number of cases was the highest, the third was during the summer when the lockdown was over (N_t3 = 725) and only a few cases were registered, and the fourth was at the start of the second wave of the pandemic in September 2020 (N_t4 = 996).

Demographic data was collected at the first time point, as the same people were contacted for participation in the different waves. The gender ratio of participants was balanced in the sample (47.5% men, 52.1% women, and 1.4% other participants). The majority of respondents (73.3%) had a college or university degree, 22.9% had a secondary school degree, 3.2% vocational education, and 0.6% reported primary school degree. Half of the respondents (50.7%) were from Budapest, 39.2% were from other cities, and 10.1% from villages. Self-reported political orientation measured on a 7-point scale from left to right showed that 50.5% identified as left-wing, 30.6% reported to be in the middle, and 18.9% reported to have a right-wing orientation; 11.3% of the sample were supporters of the government. The only demographic variable we measured in the following waves was gender, but there were very slight differences between samples. In the second wave, the gender ratio was the following: 47.7% male, 52.1% female, 0.2% other; in the third wave 47.6% male, 52% female, and 0.4% other, and in the fourth wave 47.1% male, 52.2% female, and 0.7% other. We have no information, whether education level and settlement type changed in the sample.

We conducted all the statistical analysis using IBM SPSS version 22.0 AMOS (Arbuckle, 2011) and Mplus 8.3 (Muthén and Muthén, 2017). We report all measures and data exclusions related to the research question. We ensure data accessibility in a repository of the osf.io website.

Measures

For each item, we asked participants to use a 5-point Likert scale to express how much they agree or disagree with the given statement (1 = completely disagree, 5 = completely agree). For scale reliabilities and correlations (see Table 1).

Table 1. Scale reliabilities, scale points, means, standard deviations and correlations at the first time point (all scales are 5-point).

To measure identification with all humanity and the national ingroup, we used two items out of the 9 items of the “Identification With All Humanity” (IWAH) scale (McFarland et al., 2012). As we collected data as part of an omnibus research, we had to limit the number of items we include, so we chose two items that fitted our goal the most to measure identification in a crisis situation that reflected the bond and concern for these groups. The specific items were “How much do you identify with (that is, feel a part of, feel love toward, have concern for) each of the following?” and “When they are in need, how much do you want to help…” Participants responded to both questions regarding Hungarians, and regarding all humans everywhere. We used the mean of the two items in both cases for the analysis, as they ha d strong correlation (see Table 1).

Empathy was measured by three items, asking participants to what extent they do empathize with different groups of people in the current pandemic. For each item we used the word empathy and tailored the measure to the context to mention vulnerable groups in relation to the pandemic. The items were the following. “To what extent do you empathize with the following groups of people in the current pandemic?”

“Elderly people and those living with chronic illnesses for whom the disease is more dangerous.” “Poor people with no savings at all to tide them over.” “People who lost their income due to the epidemic.” The scale had good reliability (see Table 1).

Prosocial action intentions were measured by the following two items as typical forms of benevolent behavior measured by previous research (see for example Thomas and McGarty, 2018) tailored to the context of the coronavirus: “I would make donations to help others that suffer from coronavirus.” “I would protect vulnerable groups from coronavirus even at my own expense.”

Finally, political action intentions were measured by two items we created to fit the context, based on similar items used by others to measure collective action intention (see for example van Zomeren et al., 2004): “I would put pressure on politicians to make the best decisions for the community”' and “If the authorities make bad decisions, it is our duty to speak up against them.”

We used political orientation as a control variable that was created from a nominal variable, where we asked participants to pick the political party they would vote for on Sunday if it was election day. We distinguished Fidesz-voters (the reigning government) from all other party-voters and those who would not vote, to create a dummy variable of political orientation.

Results

Statistical Analysis Plan

We planned to test our hypotheses in two-steps. As a first step, we wanted to test our hypotheses by investigating between-subject effects using data of the first data collection, where we had the biggest sample. As a second step, we wanted to replicate our results in a multilevel analysis, where we re-tested the between-subject effects on our longitudinal data, collected in four different time points. Additionally, our longitudinal data enabled us to test within-subject effects, demonstrating whether a change in one variable correlated with the change in another variable in the model, within individual respondents in time. This allowed us to test whether our variables are flexible for change with time.

First, we built a path model in AMOS using data collected in the first time point. We conducted direct comparisons between specific paths to test Hypothesis 1 and Hypothesis 2. To test Hypothesis 3 and 4, we described the strength and directions of the connections between our variables. Our model is presented on Figure 1. In order to test within-subject effects for Hypothesis 5, we conducted multilevel analysis that also allowed us to replicate our between-subject results.

Figure 1. Path model on the two forms of identification (human and national) leading to prosocial and political action intention via empathy (using cross-sectional data at Time 1). ***p < 0.001.

We used four longitudinal measurement points to confirm our proposed path model. We used multilevel SEM model with Mplus 8.3 (Muthén and Muthén, 2017; McNeish and Hamaker, 2020) that combines the advantages of both time-series analysis and SEM models (Asparouhov et al., 2018).

Our model is presented in Figure 2. To capture easily comparable results, we used standardized estimates in our analysis. Following the recommendations of Schuurman et al. (2016) we used within-subject standardization. The process was the following: first we standardized the regression coefficients for each person separately based on their within-subject variances. On the between-person level, standardization was based on the between person variances (McNeish and Hamaker, 2020). In the multilevel model we used Bayesian estimation based on two Markov Chain Monte Carlo chains (MCMC), 50,000 iterations with a thinning rate of 20 in Mplus 8.3 (Muthén and Muthén, 2017). We estimated statistical significance by 95% credibility intervals (CI), meaning that each parameter has a 95% chance of falling into this range. If CI does not contain zero, we can conclude that the estimate is different from zero (McNeish and Hamaker, 2020).

Figure 2. Between subject level associations using a multilevel path model (N = 6,240, data collected in four timepoints). ***p < 0.001.

Descriptive Statistics

We conducted our path analysis among participants at the first time point (N_t1 = 6,140). Participants scored higher than the midpoint in all the tested variables, suggesting that identification with both human and national identities, empathy, prosocial, and political action intentions were high in the sample. All the variables were correlated with a medium to large effect size. There was a negative relationship between national identity and political action intention, all other variables had positive correlations. For scale reliabilities, means and correlations between variables (see Table 1).

As we expected significant relationships between all the variables in the model, we had a saturated model with a perfect fit.

Human identity was connected to empathy (B = 0.191; SE = 0.008; p < 0.001; β = 0.286), and to both prosocial action intentions (B = 0.204; SE = 0.014; p < 0.001; β = 0.185) and political action intentions (B = 0.278; SE = 0.014; p < 0.001; β = 0.262). National identity was positively connected to empathy (B = 0.147; SE = 0.008; p < 0.001; β = 0.224), and prosocial action intentions (B = 0.155; SE = 0.014; p < 0.001; β = 0.141) and was negatively connected to political action intentions (B = −0.286; SE = 0.014; p < 0.001; β = −0.272). For a visual presentation of our model (see Figure 1).

Hypotheses Testing With Path Analysis

We tested our hypotheses if there was a significant difference between paths, on one hand, between the two ways of identification and empathy, and on the other hand, between the two ways of identification and action intentions.

In line with our first hypothesis, we found a significant difference between the path from human identity to empathy, and the path from national identity to empathy (Δχ2 = 9.92, Δdf = 1, p = 0.002). In line with our expectation in Hypothesis 2, there was also a significant difference between the path from human identity to prosocial action intention, and the path from national identity to prosocial action intention (Δχ2 = 4.63, Δdf = 1, p = 0.03).

Mediation analyses were conducted with the bootstrapping technique suggested by Macho and Ledermann (2011), where we requested 95% confidence intervals using 2000 re-samples. Mediation analysis showed that empathy was a significant mediator between identification with all humanity and prosocial action intention (B = 0.079, p < 0.001, CI [0.067; 0.92]). In line with Hypothesis 3, empathy also mediated the effect of identification with all humanity on political action intention (B = 0.050, p < 0.001, CI [0.041; 0.60]).

Hypothesis 4 predicting that national identification had a negative connection to political action intention, was supported (B = −0.286; SE = 0.014; p < 0.001; β = −0.272). At the same time, national identity had a positive connection to empathy, which had a positive connection to political action intention, resulting in the fact that the direct and indirect effects between national identity and political action had a contrast.

As national identification is rather politicized in Hungary, we wanted to investigate how much role political orientation (if someone reports being pro-government or not) plays in predicting action intentions. Furthermore, our political action measure was specifically about challenging authorities during the pandemic, so it is logical to expect that political orientation matters in this respect. However, we expected that national identification would be a meaningful predictor for action intentions, even when political orientation is controlled for. We conducted our path model again, controlling for political orientation this time, and found that national identification still had a significant effect on both prosocial action intention (B = 0.163; SE = 0.015; p < 0.001; β = 0.149) and political action intention (B = −0.114; SE = 0.014; p < 0.001; β = −0.109). However, the strength of the path between national identity and political action intention decreased. For a visual presentation of the model with the control variable (see the Supplementary Material).

Hypotheses Testing With Multilevel Analysis

To test whether the above relationships would persist over time (Hypothesis 5), we differentiated between possible within subject and between subject differences in variance. We used data from four consecutive measurement points. Within- subject variance was significantly different from zero for both identification with all humanity (var = 0.228, SD = 0.006, p < 0.001, CI [0.217; 239]) and national identity (var = 0.213, SD = 0.005, p < 0.001, CI [0.203; 224]). The hypotheses were tested using the multilevel SEM framework. Results of the structural model are presented on Figures 2, 3.

Figure 3. Within subject level associations using a multilevel path model (N = 6,240, data collected in four timepoints). ***p < 0.001, **p < 0.05.

First, we report our between subject level-findings. Empathy for vulnerable groups was positively associated with both identification with all humanity (β = 0.45; p < 0.001, CI [0.41; 0.49]) and national identity (β = 0.28, p < 0.001, CI [0.24; 0.32]), however the first association was found to be a more substantial, replicating our findings we had in the first timepoint. Prosocial action intention was positively associated with empathy for vulnerable groups (β = 0.38, SD = 0.02, p < 0.001, CI [0.33; 0.43]), with identification with all humanity (β = 0.20, SD = 0.02, p < 0.001, CI [0.16; 0.25]), and with national identity as well (β = 0.14, SD = 0.02, p < 0.001, CI [0.10; 0.18]). There was hardly any difference between the paths connecting the two forms of identification with prosocial action intention, similarly to our cross-sectional analysis. Political action intention, on the other hand, was positively associated with empathy for vulnerable groups (β = 0.21, SD = 0.02, p < 0.001, CI [0.17; 0.26]), and identification with all humanity (β =.30, SD = 0.02, p < .001, CI[.26;.34]) but negatively associated with national identity (β = −0.38, SD = 0.02, p < 0.001, CI [−0.42; −0.35]), demonstrating a similar pattern to the path analysis in the first timepoint. We used mediation analysis on the between-subject level. In line with the one timepoint path analysis, empathy mediated the effect of identification with all humanity on prosocial action intention (B = 0.176, p < 0.001, CI [0.149; 203]), and also on political action intention (B = 0.176, p < 0.001, CI [0.149; 203]).

Associations showed a different pattern on the within subject level. Both identification with all humanity (β = 0.08, SD = 0.02, p < 0.001, CI [0.05; 0.12]) and national identity (β = 0.15, SD = 0.02, p < 0.001, CI [0.11; 0.18]) had a positive association with empathy for vulnerable groups, however, the connections were weaker. Prosocial action intention was positively associated with empathy for vulnerable groups (β = 0.14, SD = 0.02, p < 0.001, CI [0.11; 0.17]), with identification with all humanity (β = 0.09, SD = 0.02, p < 0.001, CI [0.05; 0.12]), and with national identity as well (β = 0.10, SD = 0.02, p < 0.001, CI [0.06; 0.14]), but both forms of identification had a rather weak relationship with prosocial action intention. Political action intention was positively associated with empathy for vulnerable groups (β = 0.07, SD = 0.02, p < 0.001, CI [0.03; 0.10]), identification with all humanity (β = 0.06, SD = 0.02, p < 0.001, CI [0.03; 0.10]) and surprisingly, with national identity as well (β = 0.05, SD = 0.02, p < 0.001, CI [0.01; 0.09]). However, these connections were rather weak. We also conducted mediation analyses on the within-subject level, however the associations were considerably weaker. Empathy mediated the effect of identification with all humanity on prosocial action intention (β = 0.014, p < 0.001, CI [0.007; 0.22]) and on political action intention as well (β = 0.005, p < 0.001, CI [0.002; 0.10]).

Discussion

Our expectation that identification with all humanity would lead to higher empathy and higher willingness for prosocial action on behalf of vulnerable groups, compared to national identity, was supported. This correspondence was also reinforced by our longitudinal analysis. In line with the assumption that higher identification with all humanity predicts empathy, and empathy in turn leads to higher prosocial action intention, we found a mediation effect between these variables.

Furthermore, our findings supported that identification with all humanity and empathy were not only connected to prosocial behavioral intention, but political action intention, too. In contrast to identification with all humanity, we expected national identification to have a negative relationship with political action intention, and this was also supported. The patterns we found in our cross-sectional path analysis were confirmed by our multilevel analysis using longitudinal data.

In addition, by means of our longitudinal data, we had a chance to explore if our variables were changeable with time within individuals in our sample. We could see the pattern that both forms of identification led to empathy, which meant that a change in one's identification with one or the other community led to a change in one's empathy toward vulnerable groups as well. At the same time, the change in one's empathy led to a change in one's prosocial action intention as well.

As between-subject effects were much stronger among these variables, we can conclude that there are correspondences between these constructs in general, and this relationship is also influenced to some extent by individual changes in time. Overall, it seems that empathy as an emotion plays a mediating role between identification and behavior: if the level of people's identification with their community changes with time, it can shift their emotion and also their behavior. As identification with all humanity and its relationship with behavior and behavioral intentions was mostly studied in cross-sectional survey studies, our work present a novel approach, where correspondences were tested with a repeated measures design. This strengthens our findings in two ways. On one hand, we were able to repeatedly test our model in multiple time points, which increases the reliability of our findings on the between subject-level. On the other hand, on the within subject-level, we tested and found that these variables show changes with time within individuals, which suggest that they are flexible, therefore they can be also targets of interventions.

Extending previous research, where identification with all humanity was mostly related to prosocial action, our results highlighted that identification with all humanity made people not only sensitive to vulnerable groups, but critical to the government and made them more willing to challenge political decisions. The fact that human identity became a predictor of political action in the context of COVID-19 shows that solidarity could manifest both in prosocial and political action in these circumstances. Our study adds to the line of studies investigating the similarities between the motivators of prosocial and political action tendencies in an extreme situation, where the motivation for social change includes prosocial behavior for vulnerable groups, and also challenging the decisions of authorities (similarly to cases like the refugee crisis, or other humanitarian crisis, where “giving” and “acting” kind of help are both necessary, see Kende et al., 2017; Thomas and McGarty, 2018; Lantos et al., 2020).

In line with the theoretical consideration that national identity is less inclusive compared to human identity, it was a weaker but significant predictor of prosocial behavior on behalf of vulnerable groups. We can assume that the listed groups were perceived as ingroup based on both the national and human identity. In the context of the pandemic, identification with a wider and narrower community were both efficient in mobilizing people to support groups in need. The fact that both national and human identity predicted prosocial action intention, indicated that even if the pandemic was a global phenomenon, people were willing to act in their local communities, where both national and human identities were useful basis for solidarity, connecting people in a crisis.

However, in contrast to human identity, national identity played a negative role in political action intention: the ones with low national identity were willing to challenge political decisions. These findings are in line with the fact that the Hungarian government used a rhetoric favoring the interest of Hungarian people and used a hostile rhetoric toward foreigners and minorities even in relation to the pandemic (Vásárhelyi, 2020). While both identification with all humanity and national identity united people in caring for others in a crisis, the two forms of identification divided them in questioning governmental measures. Those high in human identification and low in national identification were motivated to stand up for community interests. National identity is clearly politicized and related to political orientation, but it seems that human identity can also become politicized when it comes to dissent behavior for the sake of the community. Despite the fact that national identity had a negative relationship with political action intention, national identity also had an indirect and positive route to political action intention, via empathy. It seems that two processes occur at the same time: commitment to the national ingroup prevent people from dissent behavior to the national government, and at the same time, national identity was related to empathy toward others, and this empathy was related to the protection of community interest, even if this meant a dissent to the national government.

Overall, our results suggested that both types of identification correlated with community-serving behavior in general, and these connections were also sensitive to individual changes in time. Empathy as an emotion that links identification with behavior seemed to be the most changeable, so a shift in identification can translate to a change in empathy. In turn, this change of empathy may influence prosocial behavioral intention. Our findings supported that identification, emotion, and behavior were strongly connected in general, but they were not fixed, and a change with time is also possible within individuals.

However, it is important to note that we tested relationships in a theoretical model, but our data did not allow us to test for causal connections between variables. Experimental studies need to be conducted to test, how the level of specific types of identification lead to emotional and behavioral outcomes. We hope that our model supported by our data would inspire future studies to build experimental designs where causality between variables can be directly tested. At the same time, an advantage of our survey design is its ecological validity: we could directly ask about people's reaction during the crisis of the pandemic.

Another limitation of our study is that the scope of the measured emotion and behavioral intention was not fully consistent, and the relationships between our variables may be weaker due to this fact. In the items measuring empathy, we named specific target groups, but when we asked about prosocial action intentions, we referred to “vulnerable groups” and “those in need” in general. In our political behavior intention measure, we did not name vulnerable groups, either, but referred to the “community interest” in general, so this variable is less directly linked to the protection for vulnerable groups, but more to a general dissent to authorities in relation to the pandemic. We would expect even stronger connections between the variables in the model with more consistently worded measurements.

We also consider the fact that we used only two items (item 5 and item 9) of the IWAH scale as a limitation of our study. Previous studies differentiated between two factors of the original scale, such as global self-definition and global self-investment (Reese et al., 2015), while others supported the same factor structure, but named the factors as bond and concern (Hamer et al., 2021). Reese et al. (2015) have found that global self-investment was a stronger predictor of prosocial behavioral intentions than global self-definition, and McFarland and Hornsby (2015) also found that global self-investment was an important predictor of humanitarian action intentions. In line with this, we chose two items that primarily expressed self-investment and concern for all of humanity. However, item 5 was the only one that loaded on both factors in previous studies, so it was omitted from the final subscales, as it expressed self-definition and self-investment at the same time. For future research, it is advised to use either the whole IWAH scale, or to use only the items of global self-investment subscale that is the most direct proxy of supportive behavioral intentions.

It is also important to highlight that we measured behavioral intentions, but not behavior directly that is another limitation of the study, even if using intention measures for collective action is common in the literature. At the same time, there is strong empirical evidence that action intentions are good proxies for actual behavior (for example De Weerd and Klandermans, 1999; Armitage and Christian, 2003), in line with the theory of reasoned action (Ajzen and Fishbein, 1977) and the theory of planned behavior (Ajzen, 1991), stating that behavioral intentions are mediators between attitudes and behavior. Nevertheless, the reliability of the findings can be increased by extending measures by testing actual behavior in the future.

In the Hungarian context, dissent with the government about the COVID-situation was positively related to a global (human), and negatively related to a local (national) identification. However, identifications are politicized and context-dependent, so it is important to investigate this question in other contexts, as dissent would be different in countries with liberal vs. conservative governments. In the present study, political action had a limited scope, as we measured an opposition to a right-wing government. In fact, political action can have many different forms and messages, for example, it is timely to investigate the motivations of protesting against prevention measures: how local and global identities make people accept or oppose the measures their governments introduce, and when do they stand up against them.

Conclusion

Identification with all humanity was rarely investigated in connection with political/collective action, but prosocial action. Our results supported that identification with all humanity is a relevant factor in predicting collective action. In the context of the pandemic, the most inclusive form of identification (with the human community) was able to increase people's willingness for acting on behalf of groups in need, and also for questioning political decisions not serving the community.

Identification with all humanity may be a novel way of connection for many, but the pandemic teaches the lesson, how interdependent people in the world are, and how crucial it is to cooperate and act on this identification. Intervention programs are advised to be designed to help groups of children and adults alike to develop and strengthen this new form of identity, that was not so much part of everyday life before. Beyond the pandemic, climate change as a global crisis also makes people realize that humanity should be the new scope of identification and action in the future.

Our results also highlight that when it comes to prosocial behavior, national identity can be also the basis for mobilizing people. However, when it came to political action that involves confrontation and conflict with authorities, it was only human identity that had the potential for mobilization. Therefore, we concluded that identification with all humanity was a more general predictor of both behavioral intentions, while national identification, being a politicized factor, had a mixed impact. In a crisis like the pandemic, both solidarity and cooperation with others and stepping up against unacceptable measures of authorities are tools citizens can turn to, to shape events for a social change they need globally.

Data Availability Statement

The raw data supporting the conclusions of this article will be made available by the authors, without undue reservation.

Ethics Statement

The studies involving human participants were reviewed and approved by Research Ethics Committee, Faculty of Education and Psychology of ELTE. The patients/participants provided their written informed consent to participate in this study.

Author Contributions

ME, BN, SC, and AK developed the research questions, constructed the survey, coordinated online data collection, and undertook data preparation. NL formulated the hypotheses and prepared and revised the manuscript. NL and ME conducted the data analysis in consultation with MH. AK, BN, and MH commented on the manuscript. All authors contributed to the article and approved the submitted version.

Conflict of Interest

The authors declare that the research was conducted in the absence of any commercial or financial relationships that could be construed as a potential conflict of interest.

Publisher's Note

All claims expressed in this article are solely those of the authors and do not necessarily represent those of their affiliated organizations, or those of the publisher, the editors and the reviewers. Any product that may be evaluated in this article, or claim that may be made by its manufacturer, is not guaranteed or endorsed by the publisher.

Supplementary Material

The Supplementary Material for this article can be found online at: https://www.frontiersin.org/articles/10.3389/fpos.2022.855148/full#supplementary-material

References