Predicting Pro-Environmental Behaviours in the Public Sphere: Comparing the Influence of Social Anxiety, Self-Efficacy, Global Warming Awareness and the NEP

Abstract

Environmental sustainability depends highly on our ability to identify the determinants of various types of pro-environmental behaviours (PEBs). However, so far, public sphere pro-environmental behaviours (PBS-PEBs) have received little attention in the scientific literature. This paper aims thus to fill this gap by exploring the role of dispositional (i.e., social anxiety and self-efficacy) and social psychological factors (i.e., environmental concern and global warming awareness) in the prediction of PBS-PEBs. An online questionnaire, including measures of these factors, was administered to residents (N = 199) of various Italian cities. The results indicated that (1) dispositional social anxiety (but not self-efficacy) directly (and negatively) predicts PBS-PEBs and that its effect is comparable to that of the social psychological factors considered (environmental concern and global warming awareness); (2) global warming awareness (positively) predicts PBS-PEBs directly, while (3) environmental concern, measured through the NEP scale, does it indirectly. Implications of the results for modelling the role of dispositional anxiety, environmental concern and global warming awareness in studies on PBS-PEBs are briefly outlined.

1. Introduction

Most classic studies on the determinants of pro-environmental behaviours (PEBs) have predominantly focused on private sphere actions, such as household waste recycling, energy conservation, travel mode choice and green consumption [1,2]. In contrast, public sphere pro-environmental behaviours (PBS-PEBs) have received comparatively less attention despite their increasing significance for enhancing environmental sustainability in modern societies over the past few decades. The rise of internet-based communication and social networks has expanded opportunities for social interactions and may have brought about significant changes in how pro-environmental and sustainability actions are publicly enacted [3,4]. Therefore, the determinants of public sphere pro-environmental behaviours need further investigation and understanding. The few existing studies have primarily concentrated on classic social-psychological factors, such as values, attitudes, norms and knowledge (e.g., [5,6,7,8]). Individual dispositional traits, such as social anxiety and self-efficacy, have received comparatively less attention in relation to these types of behaviours. Furthermore, there is limited knowledge about the potential impact of global warming risk awareness and how this factor, combined with environmental concerns, may influence individual pro-environmental actions in the public sphere. Therefore, this paper aims to address these gaps by examining the role of social anxiety, self-efficacy, global warming awareness and environmental concerns as predictors of PBS-PEBs.

1.1. Distinguishing between Private and Public Sphere Pro-Environmental Behaviours

In a classic theoretical reflection on pro-environmental behaviours, Stern [9] proposed a distinction of three main categories: (a) ecological behaviours in the public sphere, (b) ecological behaviours in the private sphere and (c) other environmentally relevant behaviours.

Ecological behaviours in the private sphere are individual actions intended to impact environmental conditions locally and globally by directly reducing one’s own environmental footprint. These actions are usually based on private decision-making, involving choices that affect individuals and their close contacts. Examples include buying organic food or energy-efficient household appliances, separating household waste for recycling and opting for low-impact transportation means.

Ecological behaviours in the public sphere, on the other hand, relate to ‘environmental citizenship’. This involves participating in environmental protection activities and initiatives within the public, collective sphere. These behaviours can be categorised based on the level of personal involvement required. They range from simple actions, such as donating money to environmental organisations (non-activist behaviours in the public sphere), to more demanding forms of direct participation, such as engaging in activities, events, public petitions and communication-based actions often promoted by these organisations (environmental activism).

Finally, according to Stern [9], other “environmentally” relevant behaviours include the various actions capable of having an environmental impact, which are carried out by people in various organisations, including public and private work organisations.

Stern’s theorisations and classifications have inspired the work of dozens of researchers who have disentangled the role of a number of potential determinants of PEBs enacted in several societal contexts: from home to the workplaces and from schools and Universities to the commercial and tourism areas (e.g., [1,2,10,11,12]). However, most of these studies fall within the private sphere behaviours and the “Other” category, while public sphere behaviours have received comparatively less attention to date.

1.2. Comparing the Determinants of Public and Private Sphere PEBs

Stern [9] proposed that different types of pro-environmental behaviours could be influenced in distinct ways by various social and psychological determinants, though they may all be shaped by a general predisposition towards pro-environmental action. In other words, overarching pro-environmental attitudes, values and personal norms should be seen as broad antecedents for all types of pro-environmental behaviours, while other factors influencing these behaviours can vary significantly depending on the type of environmental action, the individual and the context.

Comparative studies seem to support this idea, particularly in showing how general and specific pro-environmental attitudes tend to affect different types of private and public behaviours (e.g., [5,7,8,13]). For example, Gkargkavouzi and colleagues [14] compared the effects of environmental concern, connectedness to nature and ecological worldviews on private and public PEBs (the latter including being a member of an environmental group, taking part in protests, signing petitions, participating in community events, donating money to environmentalist associations and taking part in community actions). The factors considered revealed that they influence both types of behaviour, although they have varying levels of magnitude. Huang [15] studied the association of environmental beliefs, environmental self-efficacy and media use on private and public PEBs (the latter including persuading others, participating in environmental groups and supporting environmental policies), thus showing how both types of behaviours were similarly correlated with the factors considered. Boiral and colleagues [16] analysed the influence of environmental beliefs and stages of consciousness on organisational managers’ citizenship behaviour towards the environment. The behaviour investigated included actions such as making suggestions about ways to protect the environment, encouraging colleagues to adopt more environmentally conscious behaviours and volunteering for environmental projects in the organisation. Results confirmed that the factors considered in the study correlated with this type of public sphere behaviour as much as they were demonstrated to correlate with various types of private-sphere PEBs in the past.

However, some investigations have also identified potential differentiating factors. For instance, a study by Hansmann and Binder [6] discovered that green self-identity had a stronger impact on private sphere PEBs than on PBS-PEBs, whereas PBS-PEBs appeared to be more influenced by prescriptive social norms. Additionally, Liobikienė and Poškus [7] found that private sphere behaviours were more influenced by habitual (i.e., automatic) actions, whereas public sphere behaviours were more shaped by deliberate, conscious decisions.

Nevertheless, further research is required to deepen our understanding of PBS-PEBs and to identify other possible differential determinants. For example, a recent survey indicated that climate change is the most frequently discussed environmental issue on the internet, particularly among younger generations [17]. Factors such as global warming awareness may have become increasingly relevant for predicting PBS-PEBs compared to traditional forms of environmental concern and awareness, such as those measured by the NEP (New Environmental Paradigm [18]). The latter draws on the classic distinction between ecocentric and anthropocentric views of people–nature relationships, which has dominated the debates on environmental protection since the 1970s (see [19] for a recent discussion) and has revolved around whether people are part of nature or should dominate it and which needs should come first (people’s vs. nature’s). Some studies have shown that such classic views (and the measures devised to seize them in the population) are still relevant for investigating all types of PEBs [13,20], while others suggested a slightly weaker role for the NEP in predicting public sphere behaviours (e.g., [14]). Hence, the need exists to better understand whether and how the NEP may interact with the issues that animate current environmental debates (such as climate change-related ones) to eventually influence particular types of PEBs, such as PBS-PEBs.

Furthermore, it is well-established that individual dispositional traits can influence the availability to be involved in social interactions and actions. Social anxiety is one such trait, and exploring its significance for public sphere PEBs would be worth attention, as we will better discuss in the next paragraph.

1.3. Social Anxiety and Pro-Environmental Action

Social anxiety was defined as the “anxiety resulting from the prospect or presence of interpersonal evaluation in real or imagined social settings” [21] (p. 47) or as the desire to escape or actually avoid being with, talking to or interacting with others for any reason [22,23].

Social anxiety tends to directly or indirectly influence everyday individual behaviour in many different life domains (e.g., [24,25,26,27,28]). We believe that social anxiety may negatively affect pro-environmental behaviours (PEBs) when these require interactions with others in order to discuss, for example, actions to take. Social anxiety has also been found to be negatively associated with pro-social behaviour (e.g., [29]). Due to their intrinsic cooperative nature, pro-environmental behaviours (PEBs) have been both conceptually and empirically associated with pro-social behaviour as well (e.g., [30,31], see also [32]). Therefore, there are reasons to believe that social anxiety may be a negative predictor of PEBs, especially in the public sphere. However, no empirical investigation has yet verified this hypothesis. Previous studies on anxiety and PEBs have focused on other types of anxiety (e.g., general trait and state anxiety and climate change anxiety (e.g., [33,34,35,36])) and have shown the complexity of this relationship by producing contradictory results. For example, studies have shown that when anxiety is measured as a component of the broader neuroticism dimension (following both the Big Five and HEXACO measures (e.g., [37,38,39]), trait anxiety per se tends not to correlate with PEBs and, at best, shows only low correlations with pro-environmental attitudes (e.g., [34], see also [40,41,42,43]). When trait anxiety is measured via its emotional correlates, such as worry and fear, a greater association between trait anxiety and PEBs can be found [44,45,46]. However, even in this case, the results are not unequivocal. For example, Verplanken and colleagues [45] found that concern about global warming may increase pro-environmental intents and actions in certain individuals with psychopathological dysfunctions but decrease pro-environmental participation in others. However, no study has so far explored whether such findings may apply to the possible impact of social anxiety on PBS-PEBs. Moreover, although previous investigations have confronted eco-anxiety with public and private PEBs [36], to our knowledge, no study has so far included social anxiety in the analyses. Therefore, it would be important to determine if social anxiety might directly or indirectly predict PBS-PEBs and whether it could predict PBS-PEBs over and above other possible direct predictors of PEBs, such as environmental concern and global warming awareness (e.g., [46,47]). Furthermore, it seems reasonable to expect that the effects of social anxiety on PBS-PEBs may interact with those of one of its known predictors: the perception of self-efficacy (e.g., [48,49]).

1.4. Self-Efficacy and Pro-Environmental Behaviour

The concept of “self-perception of efficacy” is used to evaluate “how people judge their capabilities and how these affect their motivation and behaviours” ([50], p. 122, see also [51,52,53,54]). According to Bandura, people’s efficacy in dealing with their environment is not just a matter of knowing what to do, nor is it a stable skill that can be acquired once and for all. It depends on how well individuals are able to organise their cognitive, social and behavioural skills to accomplish such a task in a dynamic environment [50]. Since failure in adequately assessing one’s own abilities may lead to more or less adverse consequences for actions, people tend to reflect on such abilities and develop their own ideas regarding their capabilities of carrying out actions effectively. Thus, perceived self-efficacy is the personal assessment of one’s ability to carry out the necessary actions to handle a certain circumstance at a specific time in life. According to Bandura, perceived self-efficacy may produce a dispositional tendency and have a significant impact on an individual’s cognitive processes and emotional responses to perspective actions. People may thus develop a more or less stable tendency to perceive themselves as self-effective or ineffective when it comes to specific behaviours or even when it comes to actions in general. Dispositional self-efficacy may lead individuals to systematically experience positive (or negative) emotional states when faced with the choice of whether to carry out (or not) a specific action that they believe they are capable (or not) of performing successfully. Anticipating negative emotional states, for example, serves as a useful mechanism to keep people from placing themselves in risky circumstances and may be experienced as anxiety. Hence, general perceived self-efficacy is known to be negatively related to anxiety in general [50], including social anxiety [55], and has shown some direct effects on PEBs as well (e.g., [56]). However, most investigations to date have focused their attention on behaviour-specific self-efficacy (also called green or environmental self-efficacy; e.g., [15,57,58,59]), while the role of trait self-efficacy on PEBs remains under-investigated [56]. Moreover, a study by Leary and Atherton [49] showed that self-efficacy and social anxiety may interact to eventually inhibit social behaviour. It would thus be worth verifying whether general (trait) self-efficacy could have twofold direct and indirect effects on PBS-PEBs via social anxiety, as, to our knowledge, no investigation has so far tested this hypothesis.

1.5. Goals and Hypotheses

The general goal of the analyses discussed in this paper was to explore the role of some dispositional (i.e., social anxiety and self-efficacy) and social-psychological (NEP and global warming awareness) factors in the prediction of PBS-PEBs. In particular, we expected (H1) social anxiety and self-efficacy to be able to negatively predict self-reported PBS-PEBs over and above environmental concerns and global warming awareness.

Moreover, we expected that (H2) self-efficacy could have a twofold direct and indirect effect on PBS-PEBs via social anxiety and that (H3) global warming awareness could play a more relevant role in predicting PBS-PEBs (i.e., it would show higher correlations and predictive power) compared with classic measures of environmental concern, such as NEP.

2. Materials and Methods

2.1. Participants

The study involved 218 participants of Italian nationality, of which 60.3% were females, and the rest were males (M age = 31.50, SD = 15.65, ranging from 18 to 75 years). The participants were all of Italian nationality, and most of them (61.3%) were recruited among students attending psychology courses at Sapienza University and their acquaintances. Mean age of the participants was 28.15 (SD = 12.25), and 71.4% of all respondents were 28 years old or younger. Most of the participants completed secondary school (20.6%), followed by those who completed undergraduate or graduate courses (8.5%) or who had a post-graduate degree (3.0%).

2.2. Measures

Data from a comprehensive online cross-sectional survey was utilised to test our hypotheses. The questionnaire comprised several Likert-type scales measuring the variables considered (general pro-environmental attitudes, climate change awareness, social anxiety, self-efficacy and PBS-PEBs). The questionnaire also included some other scales that were not the object of discussion of the present analyses.

Public sphere pro-environmental behaviours (PBS-PEBs) were measured using the environmental action scale (EAS) [60]. The scale contains 18 items and was designed to assess individual civic engagement in the public sphere PEB. Respondents reported how frequently (from 1–never to 5–often) they had performed various environmental actions over the past six months. Most of the actions listed typically took place in the public sphere and involved participating in or organising group or community-based environmental activities. These activities ranged from educational events to protests and petition writing, as well as volunteering for environmental organisations and engaging in demonstrative activities (e.g., planting trees, cleaning up the neighbourhood).

The revised new environmental paradigm scale (RNEP; 15 items) [18] was used to measure environmental concern. This widely recognised measure of environmental concern was initially developed in the late 1970s [61] and updated to new theorisations regarding ecological issues in 2000. The original scale measured agreement with three main sets of beliefs, regarding in particular: (1) people’s impact on nature balance, (2) the necessity to limit socio-economic growth and (3) people’s rights to rule over nature and use it to fulfil their needs. The revised NEP (now called New Ecological Paradigm) accounts for a broader range of ecological facets by adding items focused on the so-called human exemptionalism—i.e., the beliefs that humans are not subject to laws and constraints of nature—as well as items measuring people’s awareness of imminent ecocrises—i.e., potentially catastrophic events caused by anti-ecological behaviours. In addition, the new scale included a more balanced set of items reporting pro- and anti-NEP views. Responses on the items of this revised NEP were collected on a 5-point scale, ranging from 1 (strongly disagree) to 5 (strongly agree).

Global warming awareness was measured using six items from Ma and colleagues’ [62] measure of global warming risk perception, concerns and adaptation. Specifically, we focused on items that assessed individuals’ awareness of the potentially harmful impacts of global warming both in general and to humans in particular. Additionally, we included items that measured respondents’ levels of concern and interest in acquiring more information on the topic, which we considered indirect evidence of their concern level. Responses were recorded on a 5-point scale, ranging from 1 (completely agree) to 5 (completely disagree).

Social anxiety was measured through an ad hoc-built scale by A. Kosic [63]. The scale is composed of 7 items measuring how people feel in social interactions. Responses to this scale also were on a 5-point format from 1 (completely agree) to 5 (completely disagree). Items: I am a shy person; I usually feel tense and nervous when I have to talk to someone; I feel insecure when I have to express an opinion; I feel embarrassed when I have to say something in front of others; I have difficulty speaking in a group of people; I have difficulty approaching someone I like; and I usually feel comfortable when I have to talk to someone in a position of authority.

Self-efficacy was measured using Sibilia and colleagues’ [64] Italian adaptation of the general self-efficacy scale by Jerusalem and Schwarzer (1981). The authors describe this scale as “a 10-item psychometric measure that was designed to assess optimistic self-beliefs to cope with a variety of difficult demands in life. The scale was originally developed in German by Matthias Jerusalem and Ralf Schwarzer in 1981 and has been used in many studies with hundreds of thousands of participants. In contrast to other scales that were designed to assess optimism, this one explicitly refers to personal agency, i.e., the belief that one’s actions are responsible for successful outcomes” (http://userpage.fu-berlin.de/~health/selfscal.htm; accessed on 10 September 2024). This scale also used a 5-point response scale from 1 (completely agree) to 5 (completely disagree).

2.3. Data Analysis

A set of EFAs (exploratory factor analyses) using PFA (principal axis factoring) were performed on all the scales used for our investigation. Indicators with low factor loadings (<0.30) were excluded from further analysis, and Cronbach’s alpha was computed to assess measure reliability for each dimension. The scores for all items were averaged to calculate a single index for all variables. Subsequently, bivariate correlations (Spearman’s Rho) were computed in order to explore the relationship among all aggregated indexes and to evaluate the plausibility of our hypotheses. A series of path analyses using SEM structural equation modelling (SEM) were then performed, using AMOS to jointly test all our hypotheses regarding the predictive role of the factors considered in our study.

3. Results

3.1. Dimensionality and Validity of Measures

A single-factor solution was chosen for all the scales used to measure our constructs. The explained variance (EV) for each scale was as follows: EAS (EV = 45.72%), self-efficacy (EV = 40.96%), social anxiety (EV = 47.51%), NEP (EV = 22.19%) and climate change awareness (EV = 52.30%). Descriptive statistics (means and standard deviations), factor loadings and internal consistency for the items of all the measures used in our study are reported in

Table 1. Cronbach’s alpha (α), means (M), standard deviations (SD) and factor loadings (Loads) for the items of the environmental action scale (EAS).

Scale/Subscale/Items	α	M	SD	Loads
EAS	0.93
Spent time working with a group/organization that deals with the protection of the environment.		1.79	1.19	0.86
Consciously made time to be able to work on environmental issues (e.g., working part time to allow time for environmental pursuits, working in an environmental job or choosing environmental activities over other leisure activities).		1.76	1.25	0.83
Participated in a community event which focused on environmental awareness.		1.78	1.25	0.81
Became involved with an environmental group or political party (e.g., volunteer, summer job, etc.).		1.83	1.29	0.76
Organised an environmental protest/rally.		1.29	0.86	0.75
Took part in a protest/rally about an environmental issue.		2.14	1.41	0.72
Participated in activities for protecting the environment (e.g., planting trees, cleaning streets and green areas).		2.12	1.34	0.71
Organised an educational event (e.g., workshop) related to environmental issues.		1.43	0.96	0.71
Personally wrote to or called a politician/government official about an environmental issue.		1.41	0.95	0.70
Financially supported an environmental cause.		1.90	1.22	0.69
Participated in an educational event (e.g., workshop) related to the environment.		2.19	1.31	0.67
Organised a boycott against a company engaging in environmentally harmful practices.		1.31	0.84	0.65
Organised a petition (including online petitions) for an environmental cause.		1.43	1.02	0.62
Used traditional methods (e.g., letters to the editor, articles) to raise awareness about environmental issues		2.01	1.34	0.60
Organised a community event which focused on environmental awareness.		1.31	0.92	0.57
Talked with others about environmental issues (e.g., spouse, partner, parent(s), children or friends).		3.75	1.09	0.47
Educated myself about environmental issues (e.g., through media, television, internet, blogs, etc.).		3.57	0.95	0.42
Used online tools (e.g., YouTube, Facebook, Wikipedia, MySpace Blogs) to raise awareness about environmental issues.		3.21	1.25	0.39

Note: N = 199; R = reverse scored item.

,

Table 2. Cronbach’s alpha (α), means (M), standard deviations (SD) and factor loadings (Loads) for the items of self-efficacy (SE) and social anxiety.

Scale/Subscale/Items	α	M	SD	Loads
Self-Efficacy (SE)	0.86
I am confident to be able to effectively deal with unexpected events		3.51	0.92	0.78
Thanks to my resources, I know how to handle unexpected situations		3.59	0.92	0.78
When I am faced with a problem, I usually find several solutions		3.39	0.86	0.77
No matter what happens to me, I can usually handle it		3.41	0.85	0.75
I can solve most problems if I put in the necessary effort		3.99	0.76	0.65
I remain calm when facing difficulties because I can trust my ability to face them		3.37	0.98	0.60
If I am in trouble, I can always think of something to do		3.45	0.91	0.54
I can always solve difficult problems if I try hard enough		3.89	0.84	0.51
If anyone contrasts me, I can find a way or system to obtain what I want		3.42	0.86	0.46
It is easy for me to keep up my intentions and meet my goals		3.37	0.96	0.43
Social-Anxiety (SA)	0.89
I feel embarrassed when I have to say something in front of others		2.72	1.21	0.85
It is difficult for me to speak in a group of people		2.68	1.34	0.82
I am usually tense and nervous when I have to talk to someone		2.84	1.13	0.81
I am unsure when I have to express an opinion.		2.82	1.25	0.74
I am a shy person		3.02	1.23	0.73
I feel comfortable in most social situations (R)I rarely feel embarrassed in front of others (R)		3.35	1.08	0.58
I usually feel at ease when I have to talk to someone who represents an authority (R)		2.97	1.13	0.56
I have trouble approaching someone I like		3.12	1.17	0.54
I consider myself a person who can communicate well with others (R)		3.10	1.31	0.50

Note: N = 199; R = reverse scored item.

and

Table 3. Cronbach’s alpha (α), means (M), standard deviations (SD) and factor loadings (Loads) for the items of the New Environmental Paradigm scale (NEP) and Global Warming Awareness scale (GWA).

Scale/Subscale/Items	α	M	SD	Loads
NEP	0.79
Humans are severely abusing the environment		4.45	0.77	0.69
The so-called “ecological crisis” facing humankind has been greatly exaggerated (R)		1.82	0.92	0.66
When humans interfere with nature, it often produces disastrous consequences		4.25	0.78	0.53
Plants and animals have as much right as humans to exist		4.32	0.91	0.52
The earth has plenty of natural resources if we just learn how to develop them (R)		4.37	0.81	0.51
Humans were meant to rule over the rest of nature (R)		1.70	0.96	0.49
If things continue on their present course, we will soon experience a major ecological catastrophe		4.19	0.97	0.49
The balance of nature is strong enough to cope with the impacts of modern industrial nations (R)		2.27	1.03	0.46
Humans have the right to modify the natural environment to suit their needs (R)		2.17	1.030	0.39
Humans will eventually learn enough about how nature works to be able to control it (R)		2.52	1.05	0.38
The earth is like a spaceship with very limited room and resources		3.56	1.10	0.37
Despite our special abilities, humans are still subject to the laws of nature		4.16	0.90	0.37
The balance of nature is very delicate and easily upset		4.10	0.90	0.36
We are approaching the limit of the number of people the earth can support		3.60	1.14	0.34
Human ingenuity will ensure that we do NOT make the earth unviable (R)		2.80	1.12	0.30
Global Warming Awareness (GWA)	0.86
I am interested in learning about the impacts of global warming on our living environment		4.36	0.82	0.79
I am currently very concerned about global warming		4.16	0.90	0.77
Global warming is harmful		4.73	0.59	0.73
Global warming affects humans		4.62	0.73	0.72
I am interested in learning about the measures to deal with global warming		4.23	0.90	0.69
I understand the impacts of global warming		4.25	0.80	0.63

Note: N = 199; R = reverse scored item.

. All measures showed acceptable (>0.70) to good (>0.80) reliability scores.

3.2. Descriptive Statistics and Bivariate Correlations

Bivariate correlations among all factors considered in the study are reported in

Table 4. Means (M), standard deviations (SD) and bivariate correlations among the factors considered in the study.

	M	SD	EAS	NEP	SA	SE	GWA	AGE	EDU
EAS	2.01	0.79	1	0.20 *	−0.29 **	0.06	0.28 **	0.22 **	0.22 **
NEP	3.98	0.49		1	−0.09	0.05	0.42 **	0.09	−0.11
SA	2.80	0.85			1	−0.39 **	−0.00	−0.16 *	−0.14 *
SE	3.54	0.59				1	0.10	−0.04	0.16 *
GWA	4.39	0.61					1	−0.00	−0.06
ETA	31.50	15.65						1	0.30 **
EDU	2.44	0.81							1

Note: ** p < 0.01; * p < 0.05; N = 199; EAS = Environmental Action Scale; NEP = New Ecological Paradigm; SA = social anxiety; SE = self-efficacy; GWA = Global Warming Awareness; EDU = education.

, together with means and standard deviations for all aggregated indexes. The pattern of correlations showed that one dispositional factor (social anxiety) and one social psychological factor (climate change awareness) had the highest correlations (respectively, negative and positive) with pro-environmental behaviours compared with the NEP (social psychological factor, which was also positively correlated), while self-efficacy was, surprisingly, not significantly correlated with the same behavioural measure. However, self-efficacy was negatively correlated with social anxiety, while the NEP was positively correlated with climate change awareness.

Regarding socio-demographics, no statistically significant gender differences in responses were found for most of the factors investigated, except for social anxiety, with women showing slightly higher (but statistically significant) scores than men. Gender was, thus, excluded from subsequent analyses. However, both age and education correlated negatively with social anxiety and positively with EAS; hence, we decided to include them in the subsequent analyses to control for their effects.

3.3. Path Analysis

The path analyses performed on our data followed the main goal of our study, which was of an explorative nature.

Since our goal was to compare the effects of two dispositional and two social–psychological factors in the prediction of EAS, a first path analysis was performed, introducing all factors (SA, SE, NEP and CCA) and the socio-demos (age and education) all together as direct predictors of EAS. However, unfortunately, the initial model showed poor fit indexes (χ²₍₁₅₎ = 124.186; p = 0.000; RMSEA = 0.19; SRMR = 0.15; CFI = 0.24; NNFI = −0.06). The inclusion of the correlations among the constructs did not improve the model fit substantially. This is because, although correlated with EAS, the NEP did not provide a significant independent contribution to the prediction of EAS, while SE confirmed its irrelevance in directly predicting EAS in this case. Moreover, age did not directly add to the prediction of EAS, either. However, the inspection of the modification indices (that we had explicitly requested as an output) suggested that the model could be substantially improved with the introduction of a direct path from NEP to CCA and from SE to SA. Furthermore, age emerged as a direct negative predictor of social anxiety. Since these modifications seemed theoretically plausible, we modified the model accordingly. The final model (Figure 1) showed good fit indexes (χ²₍₁₄₎ = 16.977; p = 0.257; RMSEA = 0.03; SRMR = 0.055; CFI = 0.98; NNFI = 0.97) and explained 17% of variance in EAS. Results of the Bollen–Stine bootstrap test [65]; see also [66]’s analyses performed on our data (bootstrap samples = 5000; p = 0.261), which supported the fit indexes statistical reliability. For these reasons, we concluded that our data fully supported H3, while H1 was supported only partially by the confirmation of the direct effect of social anxiety on EAS. The hypothesised direct effect of self-efficacy (H2) and the NEP on EAS were instead refuted.

4. Discussion

Our findings expanded the existing literature on public sphere pro-environmental behaviours (PBS-PEBs) by identifying two additional, relevant antecedents of these behaviours: social anxiety (SA) and global warming awareness (GWA).

In our study, global warming awareness was found to be a direct positive predictor of PBS-PEBs, and its effects appeared to be stronger than those from traditional measures of environmental concern, such as the NEP. More specifically, global warming awareness ended up mediating the influence of the NEP on PBS-PEBs. This result seems consistent with theories, for example, Stern and colleagues [67] Value–Beliefs–Norms theory of non-activist behaviour, which considers the influence of general environmental attitudes (also known as environmental worldviews) on PEBs to be mediated by other factors, including moral norms and awareness of consequences. Another factor that has been shown to mediate the influence of the NEP on PEBs is, for example, self-identity [68,69]. However, so far, such effects have been mostly explained on the basis of the “compatibility” principle (e.g., [70,71]), occurring, particularly when the mediator is conceptually more related (in terms of action, context and/or time) to the target behaviour compared to the indirect predictor (in this case, the NEP).

However, in our case, it seems that the mediating role of GWA on PBS-PEBs could be better explained on the basis of what we could call here a “trendiness” effect (a process that makes some issues appear more cogent than others in a given time period). Considering the relevance that climate change-related issues have been gaining in public discourse on social media, the greater association between GWA and PBS-PEBs becomes, thus, not surprising. Various authors have emphasised the powerful agenda-setting effect [72,73] that social media networks may have on general and specific sectors of public opinion [74,75]. In this case, such an effect could have led global warming issues to appear more cogent and timely and thus conceptually more tightly associated with environmental actions.

Future studies should thus examine the possibility that individual concerns and awareness related to trending environmental topics (such as climate change issues at present) could become comparatively more effective in directly predicting environmental actions (either in the public or private sphere) compared to topics concerning traditional environmental conceptual frameworks (such as those expressed by the NEP).

However, our data also indicated that the decision to enact PBS-PEBs may depend upon one’s own level of social anxiety. As a matter of fact, in our study, people showing higher levels of social anxiety demonstrated a lower propensity towards enacting these types of behaviours. This is not surprising, considering that PBS-PEBs imply, by definition, a certain degree of social interaction, while some of them (for example, organising collective demonstrations, petitions and protests) require a strong perceived social efficacy. Only people who have developed a certain familiarity with social interactions and who enjoy the challenges posed by collective confrontation might want to engage in such actions. Future studies could, thus, better investigate whether this factor could become another useful parameter for distinguishing those who prefer pro-environmental behaviours in the public domain from those who would rather act in the private one.

Our analyses also confirmed the well-known link between self-efficacy and social anxiety but not the link between self-efficacy and PBS-PEBs. Further study is necessary to fully understand this outcome. Previous investigations that had shown the influence of perceived self-efficacy on both public and private PEBs had used context-specific measures of self-efficacy, typically called green or environmental self-efficacy. That is, they measured self-efficacy as the perceived individual ability to carry out particular types of PEBs (e.g., [76], see also, [15,59]). Our investigation, instead, focused on trait self-efficacy (e.g., [56]), that is, on a general and stable dispositional tendency in life. Hence, if supported by future investigations, our results would suggest that the individual dispositional tendency to perceive oneself as more or less effective in general is irrelevant for this type of behaviour. It would be, thus, worthwhile to evaluate whether other types of perceived efficacy, such as collective efficacy, could be more pertinent to this type of environmental behaviour [77,78]. Previous studies have shown that collective and individual (environmental) self-efficacy might be more or less effective in predicting PEBs depending on the particular characteristics of the behaviours considered [76], as well as the communities in which people live [58]. Hence, our investigation supports the contention that more refined studies are needed to better disentangle the role of perceived efficacy (in its various forms and contexts of application) in the predictions of various types of PEBs.

Finally, our data also bring some new evidence regarding the role of socio-demographics such as education, age and gender in the prediction of PBS-PEBs. In particular, we found education level to directly (and positively) predict participation in PBS-PEBs. This parallels the results of classic studies on pro-environmental behaviours of the private sphere, which has also shown PEBs to be most likely endorsed by people with higher education levels on several occasions [1]. However, in our investigation, age predicted PBS-PEBs only indirectly via its (negative) association with social anxiety, while gender did not contribute to predicting PBS-PEBs in this particular case—not directly or indirectly. This contrasts with past investigations that have shown women to be more involved in private PEBs and men in public ones (e.g., [79]). Nevertheless, it confirms recent results that had shown no, or negligible, gender differences in engaging in PBS-PEBs [80]. Future studies could thus evaluate whether our result was due to the particular characteristics of our sample (see “Limits of this study” below) or, more simply, to be a sign of an ongoing gender levelling in PBS-PEBs conceptualisation in the newest generations.

5. Limitations

This study presents some limitations that need to be acknowledged. First of all, the generalisability of results to other social and/or cultural contexts might be limited, in this case, due to the high heterogeneity of our sample (which included a high proportion of young Italian university students). Previous investigations have shown that the determinants of PEBs (and PEBs themselves) may vary across social-cultural contexts [81]. For example, our sample’s characteristics could offer an alternative explanation for the failure of the NEP scale to directly influence EAS. Such failure could, indeed, be due to a ceiling effect, which has often been recorded when this scale was administered to particular samples of the population (e.g., environmentalists, ‘white collars’, students and highly educated people in general [82]). Sample characteristics may also have produced some age-cohort effects. In other words, the NEP content may refer to a conceptualisation of environmental issues, which might appear surpassed to the eyes of the youngest generations, who could rather be more influenced by the current views and discourses on global warming (the trendiness effect, previously mentioned). Future studies should thus evaluate whether our results can be replicated on more heterogeneous samples drawn from different socio-cultural, demographical and geographical contexts.

However, results regarding the NEP may also be due to our choice of using a single-factor solution for the items of this scale. Although our reliability indices were in line with previous investigations [61,68,69], the results for this scale have often been shown to vary substantially depending on the number of factorial solutions chosen [82].

Finally, the proportion of variance explained by our model could be substantially increased by identifying other relevant predictors that are able to capture the full range of influences on PBS-PEBs.

6. Conclusions

In this study, we discussed the importance of deepening the understanding of the potential determinants of private sphere pro-environmental behaviours (PBS-PEBs). We argued that the introduction of internet-mediated social interactions through social media networks may have contributed to their increasing importance in addressing environmental challenges in recent decades. Social media has increased the opportunities for people to interact within and across national borders on several issues of social relevance, which has greatly amplified the possibility of exchanging information, experiences, opinions, and practices. Although studies in communication research have previously suggested that these kinds of drastic shifts might have a significant influence on environmental behaviours [4], see also [15,59], scientific investigations on the topic are still in the early stages. Therefore, we thought that there was a need to broaden our research horizons and methodologies in order to better understand the determinants of PBS-PEBs in environmental studies. By identifying two potential additional predictors (social anxiety and global warming awareness), our data confirmed our hypotheses and reached its intended goals, which were precisely to demonstrate that we still know very little about PBS-PEBs and that we are still far from having identified all the various components of a comprehensive model of PBS-PEBs.