Failure to morally self-regulate: deindividuation and moral disengagement from pro-environmental behavioural intentions across tourism and home contexts

ABSTRACT

This study extends current research on cross-contextual inconsistencies in pro-environmental behaviour (PEB) intentions. We compare the mechanisms influencing PEB intentions in tourism and post-visit home contexts. We find that moral obligation, moral disengagement, and perceived anonymity are consistently linked with PEB intentions in both contexts, but public self-awareness has a substantial impact on PEB intentions only during travel. While public self-awareness does not work consistently in relation to PEB intentions across tourism and home contexts, perceived anonymity has a consistently strong impact on moral disengagement and PEB intentions in both contexts. This study extends previous research on the contextual inconsistency of PEB intentions by elucidating the mechanisms by which deindividuation (anonymity and absence of public self-awareness) may limit PEB intentions through failure of moral self-regulation. This research also contributes to the literature on moral disengagement and deindividuation by demonstrating how the predictive effects of deindividuation components depend on the environment. Our findings will help practitioners and policymakers to devise effective, context-aware strategies to promote sustainable behaviours.

KEYWORDS:

• Pro-environmental behaviour
• deindividuation
• perceived anonymity
• public self-awareness
• moral obligation
• moral disengagement

Introduction

Despite the economic benefits promised by the tourism industry, its negative environmental impacts, such as greenhouse gas emissions, pollution, and solid waste, continue to be a global concern (Loureiro et al., 2022; UNWTO, 2021). Significant behavioural change by tourists is required to mitigate these negative impacts (Xu et al., 2020), making pro-environmental behaviour a well-researched topic in the tourism literature (Hoberg et al., 2021; Loureiro et al., 2022; Miller et al., 2015). Substantial research has been undertaken to understand tourists’ adoption of pro-environmental behaviour (PEB) in tourism contexts (Hoberg et al., 2021; Wu et al., 2021a) and the impact of sustainable tourism experiences on individuals’ behavioural changes on returning home (Ballantyne et al., 2018; Hughes, 2013). However, much remains unknown about the inconsistent mechanisms that influence PEB and PEB intentions across tourism and home contexts (Holmes et al., 2021; Wu et al., 2021b).

Pro-environmental decisions are highly contextual (Wu et al., 2021a, 2021b; Xu et al., 2020). Studies of the underlying mechanisms that promote or restrain socially desirable behaviours need to account for both the external conditions (Guagnano et al., 1995; Wu et al., 2021b) and the individual’s situational self-awareness in relation to those conditions (Steidle & Werth, 2014). For example, in tourism contexts we may ignore types of PEB that are well-established in home settings (Holmes et al., 2021; Miao & Wei, 2013), as individuals often travel ‘to get away from it all’, while the privacy of home provides anonymity that may encourage people to eschew social norms (Thomas & Sharp, 2013) and potentially morally disengage from PEB. Although the extant literature indicates little consistency in conservation-oriented behaviour or intentions between home and tourism contexts (Hughes, 2013; Whitmarsh et al., 2018), little is known about what causes this inconsistency and how. We leverage deindividuation theory and moral disengagement theory to extend previous insights and propose an alternative account of contextual inconsistency in PEB.

First, according to deindividuation theory, absence of public self-awareness and the condition of anonymity may make people less likely to regulate their moral behaviours with reference to social norms (Diener, 1979; Lowry et al., 2016), which may help to explain cross-contextual inconsistency in PEB and PEB intentions. However, pro-environmental literature has so far devoted limited attention to the deindividuation effect and its critical constructs, such as perceived anonymity and public self-awareness. The few existing tourism studies on deindividuation have considered the impacts of anonymity (Su et al., 2022) and self-awareness (Hwang & Lee, 2019) on tourists’ pro-social/environmental choices only in tourism settings. No previous research has examined how the deindividuation effect functions differently through moral regulation between tourism and home contexts.

Second, tourists may justify their environmentally harmful behaviours through moral disengagement mechanisms, involving selective suspension of internal moral standards that would deter them from engaging in unethical behaviours (Peng et al., 2022; Wu et al., 2021a). However, more research is needed to determine how moral disengagement is linked with deindividuation across contexts, and identify the antecedents of moral disengagement and their roles in influencing inconsistency in PEB and PEB intentions. We are unaware of any previous attempts to model the relationship between public self-awareness, moral obligation, perceived anonymity, moral disengagement, and PEB intentions.

This study identifies an important theoretical underpinning that advances understanding of cross-contextual inconsistency in PEB and PEB intentions. In particular, we: (1) compare the results between tourism and post-visit home contexts to determine whether inconsistency in PEB intentions across contexts is a function of the deindividuation effect; (2) investigate how critical factors in deindividuation (perceived anonymity and public self-awareness) may influence tourists’ moral considerations toward PEB between tourism and home contexts; and (3) examine how moral considerations and PEB intentions may be fostered through effective, context-aware strategies.

Literature review

Deindividuation theory

The concept of deindividuation, originally proposed to explain individuals’ loss of self-control in crowds due to anonymity (Le Bon, 1895/1995), was reintroduced to refer more broadly to contexts in which one is not seen or noticed as an individual (Festinger et al., 1952). Loss of individuality may lead to a decrease in inner restraints and an increase in uninhibited behaviours as a result of being unidentified and unaccountable (Festinger et al., 1952).

Social psychologists hypothesise deindividuation as a psychological process involving reduced self-awareness that leads to increased antisocial actions (Diener, 1979; Prentice-Dunn & Rogers, 1982). According to Diener (1979), absence of self-awareness is a crucial factor inducing a deindividuation state in individuals. Prentice-Dunn and Rogers (1982) specify the importance of both public self-awareness and anonymity: increased anonymity and reduced public self-awareness are likely to lower people’s concern for others’ thoughts about them, resulting in less pro-social behaviour. In contrast, when people are highly aware of the presence of others and are concerned about being observed or evaluated by others, they tend to engage in behaviours that follow the norm of the social context (Coker, 2020; Uhrich & Tombs, 2014). In a state of public self-awareness, people may worry about their image and reputation. Such uncomfortable feelings elicit their conformity with norms (either antisocial or prosocial) to meet the expectations of their social environment (Hwang & Lee, 2019; Uhrich & Tombs, 2014).

Public self-awareness

Public self-awareness, referring to individuals’ awareness of publicly displayed aspects of themselves, is associated with an increased focus on external appearance and behaviours (Prentice-Dunn & Rogers, 1982; Uhrich & Tombs, 2014). The presence of others and the perception of being observed by them are enough to cause individuals to be concerned about their public selves (Hwang & Lee, 2019). The presence of others may have significant social impacts on people’s attitudes and behaviours (Bodur et al., 2015). Specifically, the sense of being watched prompts socially desirable behaviours because those who are highly conscious of their public selves worry about what other people will think of their conduct. Public self-awareness may be activated by the presence of others, inducing actions in alignment with social norms, such as PEB (Bodur et al., 2015). Conversely, perceived lack of accountability may suppress public self-awareness, enabling individuals to escape from the pressure of social evaluation and engage in more uninhibited behaviours (Lowry et al., 2016).

According to deindividuation theory, deindividuation can be induced by lack of public self-awareness (Uhrich & Tombs, 2014), as evidenced in behavioural studies in various contexts (Bodur et al., 2015; Lowry et al., 2016). For example, Bodur et al. (2015) find that those who are exposed to high public self-awareness cues tend to have a greater preference for environmentally sustainable products. Lowry et al. (2016) indicate that low public self-awareness is associated with low moral consideration, which leads to more online antisocial behaviour such as cyberstalking. Similarly, in the context of hospitality, customers tend to tip more in restaurants when they are surrounded by strangers rather than friends (Cavusgil et al., 2022).

Recent tourism research also confirms that travellers tend to engage in more PEB (e.g. choosing green hotels) when their behaviours are more socially visible (Rahman et al., 2023). Tourists are more inclined to make pro-social choices when they believe they are being observed (Su et al., 2022). People are likely to be more conscious of others’ opinions and social expectations in tourism settings than in private settings. In the former, there may be more public self-awareness, which increases the possibility that people will adopt more sustainable behaviours. However, in domestic contexts, the effect of public self-awareness is limited, as people are likely to be less compelled to adopt PEB with little immediate public scrutiny. Therefore, the following hypothesis is proposed:

H1: Public self-awareness positively affects intentions to engage in pro-environmental behaviours in tourism contexts but not in home contexts.

Few studies have examined how public self-awareness may regulate people’s moral norms to promote PEB. In a state of high public self-awareness, individuals tend to follow moral norms and are more likely to perform socially desirable actions (Coker, 2020; Uhrich & Tombs, 2014). Public self-awareness incorporates moral aspects; therefore being aware of the presence of others may affect one’s moral judgment (Hwang & Lee, 2019). Tourists may readily experience a sense of public self-awareness, such as being noticed by travel companions or other tourists, which may be a mechanism that triggers their moral obligation for PEB. For example, when tourists become more publicly aware, they may develop greater moral concern for others, which in turn will lessen the likelihood of unethical behaviours (Su et al., 2022). In contrast, at home, the influence of public self-awareness on environmental behaviours is likely to be limited. Hence, we propose that:

H2: Public self-awareness positively affects one’s moral obligation towards PEB intentions in tourism contexts but not in home contexts.

We still know little about public self-awareness as an antecedent of moral obligation in tourism contexts, particularly in relation to pro-environmental actions. However, the positive influence of moral obligation on PEB intentions has been widely researched (Han et al., 2018; Varotto & Spagnolli, 2017; Wu et al., 2021a). These studies find that moral obligation is a significant factor in generating strong intentions to engage in PEB across multiple contexts, as long as individuals are able to recognise their moral responsibilities. In tourism contexts, tourists’ sense of moral obligation leads to heightened commitment to PEB (Han et al., 2018; Liu, Wu, & Che, 2019; Wu et al., 2021a), whereas in home settings, recognising one’s moral responsibilities is crucial in motivating the adoption of more pro-environmental actions (Massarutto et al., 2019). Thus, the following hypothesis is proposed:

H3: Moral obligation positively affects intentions to engage in pro-environmental behaviours in both tourism and home contexts.

Moral disengagement theory

Moral disengagement theory addresses the question of how individuals can live without obvious guilt or distress after engaging in reprehensible behaviours (Bandura, 2002). People normally behave in alignment with moral standards, and self-regulatory mechanisms such as self-sanctioning inhibit them from engaging in unethical behaviours. However, Bandura et al. (1996) identify a set of moral disengagement mechanisms through which moral self-sanctioning may be selectively disconnected from reprehensible actions, thereby preventing people from suffering feelings of guilt.

Moral disengagement may focus on redefining reprehensible behaviours as socially acceptable through various mechanisms (Bandura, 2002; Wu et al., 2021a): (1) moral justification, e.g. it is OK to engage in less eco-friendly behaviour if it is for the convenience of friends; (2) euphemistic labelling using neutralising language, such as referring to clearing forests to build golf courses as ‘timber harvesting’; and (3) advantageous comparison, in which reprehensible conduct is contrasted with more egregious behaviours. The mechanisms of moral disengagement may also centre around agency, whereby individuals minimise their agentive roles in reprehensible behaviours, including displacement and diffusion of responsibility (Bandura et al., 1996; Wu et al., 2022).

Distortion of harms depicts how people ignore and distort harms caused by their reprehensible actions. When the harmful consequences of one’s actions are minimised and distorted, self-sanctioning is less likely to be activated (Bandura, 2002). For example, people may argue that engaging in deviant behaviours, such as excessive water usage, for a short time while travelling may not cause great harm to their destination (Su et al., 2022). Moral disengagement may also involve dehumanisation, referring to dehumanising victims of the consequences of misbehaviours by divesting them of their human qualities/values, and attribution of blame, placing blame on others or on circumstances. Through these moral disengagement mechanisms, people may achieve self-exoneration and even feel self-righteous (Bandura, 2002; Bandura et al., 1996).

The various mechanisms outlined above illustrate how individuals may justify their misdemeanours to themselves. Moral disengagement theory has been applied in several disciplines to explain (un)ethical behaviour (Moore et al., 2012; Peng et al., 2022; Wu et al., 2022). However, few tourism studies have used it to explain deviant behaviours and/or inhibition of PEB (Sharma, 2020; Su et al., 2022; Wu et al., 2021a).

Perceived anonymity

Perceived anonymity refers to the extent to which people feel they are unidentifiable (Hite et al., 2014), which is closely linked with deindividuation. According to deindividuation theory, the feeling of having a hidden identity (i.e. perceived anonymity) may evoke deindividuation, which will result in behaviours that violate norms (Vilanova et al., 2017). The relationship between individuals’ perceived anonymity and their behaviour is well documented (Hite et al., 2014). People are more likely to engage in pro-social activities when they are identified as individuals, and are more inclined to perform anti-normative behaviours when they are anonymous.

Deindividuation explains how perceived anonymity expedites loss of the self-regulation that normally controls one’s anti-normal actions (Vilanova et al., 2017). Individuals who feel anonymous in a situation are more likely to engage in behaviours that they would not perform in highly identifiable contexts. Perceived anonymity may occur both at home and on holiday. The safety of a home environment provides anonymity, while the fact of not being known or identifiable by referent others during travel may also generate that sense of anonymity, especially when immersed in crowds of tourists (Sharma, 2020; Su et al., 2022). This leads us to suggest that:

H4: Perceived anonymity negatively affects one’s intention to engage in pro-environmental behaviours in both tourism and home contexts.

According to deindividuation theory (Prentice-Dunn & Rogers, 1982), when individuals feel anonymous, they are more likely to morally disengage from socially expected behaviours. If people experience a sense of anonymity in unusual places or among strangers when they are on holiday, they will be more likely to morally disengage from PEB without worrying about being judged by their usual social circles (Su et al., 2022). Deindividuation theory suggests that perceived anonymity contributes to deindividuation and increases moral disengagement, whereby people feel less personally responsible for their behaviours (Vilanova et al., 2017). In home contexts, this may be manifested as a diminished sense of personal responsibility for upholding moral principles, thus fostering moral disengagement. Accordingly, we propose that:

H5: Perceived anonymity positively affects one’s moral disengagement in both tourism and home contexts.

The anonymity of tourism contexts, which are typically far away from home, provides a liminal space with potential to engage in actions outside the norm (Sharma, 2020). Thus, we can assume that moral disengagement is easily possible as people’s moral behaviours are put to the test while on holiday. Recent literature provides strong empirical evidence that moral disengagement is a significant driver of tourists’ deviant behavioural intentions (Su et al., 2022) and a critical inhibitor of their PEB intentions (Wu et al., 2021a). In an example from daily life, there is evidence that moral disengagement had negative impacts on people’s intentions to comply with social distancing measures during the COVID-19 pandemic (Wu et al., 2022). Hence, we propose that:

H6: Moral disengagement negatively affects one’s intention to engage in pro-environmental behaviours in both tourism and home contexts.

Subjective feelings of anonymity may induce deindividuation, which is strongly associated with lower public self-awareness, especially when people are immersed in a group or crowd (Prentice-Dunn & Rogers, 1982; Steidle & Werth, 2014). Perceived anonymity is seen as an antecedent to loss of public self-awareness (Lea et al., 2001). For example, Steidle and Werth (2014) find that people who stay in dark environments in relatively anonymous conditions tend to feel less worried about others’ impressions of them. Similar results are found in relation to perceived anonymity in virtual spaces, particularly on social media (Lowry et al., 2016). This reduction in public self-awareness may result in unethical behaviours that would not be exhibited in face-to-face interactions (Mkono, 2018; Peng et al., 2022). People may feel that, compared with behaviours in more public areas, their environmental actions at home are more private and less noticeable to others, and this perception of anonymity may diminish their public self-awareness. Hence, we propose that:

H7: Perceived anonymity negatively affects one’s public self-awareness in both tourism and home contexts.

Figure 1 summarises our hypotheses in a conceptual framework.

Methodology

Sampling and procedures

To compare the results between tourism and post-visit home contexts, we conducted two rounds of surveys. First, we ran an onsite survey (T1) at Dinghu Mountain National Nature Reserve in China’s Guangdong Province. This was followed by an online survey one month after the tourists’ visit (T2). As a UNESCO biosphere reserve, Dinghu Mountain is a unique ecosystem of tropical and sub-tropical forest offering high-quality environmental education and tourism activities. The reserve attracts millions of tourists annually, providing them with opportunities to discover, appreciate, protect, and enjoy the natural and cultural heritage. Its environmental education and interpretation initiatives include an ecological science exhibition in the visitors’ centre, narrative brochures, public notices, environmental education seminars, and volunteering on conservation projects. These factors made Dinghu Mountain Reserve appropriate for administration of our onsite survey (T1).

Figure 1. Research model.

Convenience sampling was used to collect survey responses at three popular tourist spots, administered by trained investigators during both weekdays and weekends. To lower the potential social desirability bias of face-to-face survey distribution, the investigators administered three questionnaires to three respondents at the same time, and explained that after completing the survey, their questionnaires would be put directly into a folder and would not be judged by the investigator. Over 330 questionnaires were handed out and 285 valid surveys were retained for data analysis. Of the latter, 169 were from female and 117 from male visitors. The 16–35 and 36–45 age groups were best represented. Over 75% of the respondents had a university-level education.

The post-visit survey questionnaires were distributed to the same group of respondents through an online survey link via WeChat (the most prevalent online communication tool in China) approximately one month after their visit. These follow-up survey links were sent from an official WeChat account with the name ‘University Research Project’. During the onsite data collection, those who agreed to participate in the follow-up survey were clearly informed that their WeChat contact would be coded with a unique number for data tracking purposes to avoid personally identifiable information. Before the follow-up surveys were distributed, information was sent to the participants to reassure them that the goal was to gather different views, that there were no right or wrong answers, and that their responses would be retained confidentially and anonymously.

Of the original sample from the onsite survey (T1 = 285), 170 respondents made valid responses to the post-visit survey (T2). This relatively small sample size of 170 is considered acceptable owing to the challenges of obtaining data from participants in longitudinal follow-up studies (Barnes et al., 2016). Following Vandenberg and Self (1993), we calculated the potential attrition effect of our longitudinal study by conducting non-response analysis, and confirmed that there was no significant difference in gender, age, education, or income between the respondents at T2 and the entire set of respondents at T1.

According to Hair et al. (2010), a ratio of at least five responses for each item is considered acceptable. The onsite survey contained 24 items and the post-visit survey 32, thus requiring at least 120 responses for T1 and 160 for T2. Our sample sizes for both T1 (n = 285) and T2 (n = 170) exceeded these requirements. Of the 170 participants in T2, 40% were male and 60% female, most were aged between 16 and 45, and over 80% had or were studying for an undergraduate degree. Although the primary data were collected before 31 January 2020 (pre-COVID-19), our findings remain applicable because we focus on microsystem interactions between individuals (Li et al., 2023).

Measurement

Our survey instrument was based on a detailed review of the relevant literature and was initially developed in English. Bilingual (Chinese-English) researchers were involved in the back-translation process, and a few minor discrepancies in the questions were discussed and resolved. We then conducted pre-tests by seeking opinions from tourism scholars, industry practitioners, and tourists who had been to the Dinghu Mountain Nature Reserve. The wording of some items was improved, and the questionnaire was further modified for congruence with the research context. This procedure ensured content validity. Seven-point Likert scales were used for the questionnaire items, ranging from (1) strongly disagree to (7) strongly agree.

The scale for PEB intentions, adapted from Hughes et al. (2011) and Ramkissoon et al. (2013), has been widely validated in previous studies of PEB intention changes (e.g. He et al., 2022; Wu et al., 2021b). The survey contained eight items relating to PEB intentions, including ‘Remind others to avoid environmentally harmful behaviours’. Notably, the measures for actual PEBs and PEB intentions mirrored each other. For example, participants were asked to indicate their intention to ‘Use green (non-plastic) shopping bags’ for PEB intentions in both T1 and T2, and were asked to report their frequency of use of green shopping bags as an actual PEB in T2, on a seven-point scale from (1) never to (7) always. Actual PEB was only assessed in the post-visit home context (T2), as the measures were chosen to reflect daily sustainable actions rather than tourism context-specific actions. This aligns with Hughes et al.’s (2011) scale measurement approach for longitudinal studies with follow-up surveys, helping to examine changes between tourism and post-visit home contexts and determine potential long-term impacts of travel experiences on sustainable living (Wu et al., 2021b).

A three-item scale was adapted from Govern and Marsch (2001) to measure public self-awareness across T1 and T2, including ‘Right now, I am concerned about the way I present myself’. To measure perceived anonymity from T1 to T2, Hite et al.’s (2014) three-item scale was adapted, including ‘It is difficult for others to identify me as an individual’. Moral obligation was measured across T1 and T2 using Onwezen, Antonides, and Bartels’s (2013) scale, also with three items; for example, ‘I feel an obligation to behave in an environmentally friendly way’. An eight-item scale was adapted from Bandura et al. (1996) and Moore et al. (2012) to examine moral disengagement from T1 to T2; for example, ‘Considering the pollution caused by big enterprises, it’s hardly a sin for us to perform environmentally harmful behaviours sometimes’. The internal consistency of the constructs was examined using Cronbach’s alpha and composite reliability. The results showed that the Cronbach’s alpha of each construct was greater than 0.80 and each composite reliability value exceeded the 0.8 threshold, indicating the instrument’s reasonable reliability and internal consistency (Hair et al., 2010). For details of the scales, see Appendix.

Data analysis

Data analysis was carried out in three steps. First, we conducted descriptive analyses and a Pearson’s correlation analysis to assess the respondents’ profiles and the mean and standard deviations of each construct, and to understand the associations between constructs (see Table 1). Second, a paired-sample t-test was performed to compare potential changes in key constructs between the tourism and post-visit home contexts (see Figure 2). The potential attrition effect owing to discontinued participants between the onsite (n = 285) and post-visit surveys (n = 170) was also examined following Vandenberg and Self’s (1993) procedure. The results revealed no significant differences in the sample profile between those who participated in both surveys and those who engaged only in the initial onsite survey, indicating limited threat from attrition in this research. Third, to test the conceptual framework and further assess the hypotheses, structural equation modelling (SEM) was performed using AMOS 26.0. Prior to SEM analysis, reliability and validity tests were conducted to evaluate the quality of the measurement, including examinations of Cronbach’s alpha, composite reliability, convergent validity, and discriminant validity. Confirmatory factor analysis (CFA) was employed to assess the constructs’ validity and obtain the results of factor loadings. The overall fit of the measurement and structural models was examined using CFA and SEM respectively.

Figure 2. Results of paired-sample t-test (tourism versus home). Note: PEB = pro-environmental behaviour, T1 = onsite survey (tourism context), T2 = post-visit survey one month after T1 (home context).

Table 1. Descriptive statistics and correlation among factors (N = 285).

T1(tourism) n = 285	Mean	SD	1	2	3	4	5
1. Public Self-Awareness	5.381	1.149	(0.780)
2. Moral Obligation	6.612	0.742	0.171**	(0.866)
3. Perceived Anonymity	2.319	1.496	−0.052	−0.063	(0.938)
4. Moral Disengagement	2.289	1.103	−0.082	−0.122*	0.579**	(0.732)
5. PEB intention	5.927	0.764	0.277**	0.202**	−0.286**	−0.282**	(0.660)
T2(home) n = 170	Mean	SD	1	2	3	4	5	6
1. Public Self-Awareness	4.986	1.559	(0.949)
2. Moral Obligation	6.146	0.792	0.179*	(0.924)
3. Perceived Anonymity	2.680	1.532	0.089	−0.226**	(0.962)
4. Moral Disengagement	2.638	1.073	0.030	−0.406**	0.560**	(0.756)
5. PEB intention	5.791	0.769	0.041	0.568**	−0.364**	−0.404**	(0.735)
6. PEB	4.866	0.978	−0.060	0.400**	−0.327**	−0.285**	0.672**	(0.713)

Note: a. *p < 0.05 and **p < 0.01, b. The numbers enclosed in parentheses and presented diagonally indicate the square roots of the AVEs. c. SD = Standard Deviation, PEB = pro-environmental behaviour

Findings

Results of paired-sample t-test

Figure 2 illustrates the results of the paired-sample t-test between the onsite and post-visit surveys. Significant differences were found between the tourism and post-visit home contexts regarding moral obligation (p < 0.01), moral disengagement (p < 0.05), perceived anonymity (p < 0.05), and public self-awareness (p < 0.001). Respondents’ moral obligation was found to be significantly lower in T2 than in T1, and this was consistent with changes in public self-awareness (T1 = 5.381 to T2 = 4.986). As shown in Figure 2, the changes in moral disengagement and perceived anonymity were quite similar (both increased significantly between T1 and T2), suggesting a strong correlation between these two variables.

Notably, no significant change in PEB intentions was observed between the tourism and post-visit contexts (p > 0.05), and self-stated PEB was much lower than both T1 and T2 PEB intentions. This is in line with the attitude – behaviour inconsistency of PEB discussed in previous studies (Holmes et al., 2021; Whitmarsh et al., 2018; Xu et al., 2020). Our findings of significant declines in public self-awareness and moral obligation and significant increases in perceived anonymity and moral disengagement suggest how the deindividuation effect may work to explain the attitude – behaviour disconnect through failure of self-regulation with regard to moral agency. To confirm the critical components of the deindividuation effect and how they influence moral considerations in tourism and post-visit contexts, SEM analysis of both contexts was conducted.

Measurement model

The results of CFA (in both T1 and T2) confirmed that all the standardised factor loadings exceeded the 0.50 threshold (see Appendix 1). The average variance extracted (AVE) for each construct was also calculated, with all values being greater than 0.50 except for PEB intention, indicating adequate convergent validity. However, as the AVE value of PEB intention is 0.435, only slightly below 0.50 (only in T1) with high values for both composite reliability and Cronbach’s alpha, this can be seen as acceptable (Kuo et al., 2022). In addition, the square roots of the AVE of each construct are higher than the corresponding correlation coefficients (see Table 1), suggesting satisfactory discriminant validity (Hair et al., 2010). Overall, the results reflect good measurement quality.

Several model fit indices were considered to assess the measurement model through CFA, including the chi-squared test (χ2) with degrees of freedom, the comparative fit index (CFI), the Tucker-Lewis index (TLI), and the root mean square error of approximation (RMSEA). Previous literature suggests an χ2/df lower than five, a CFI greater than 0.90, and a TLI greater than 0.80 as thresholds for acceptable model fit (Hair et al., 2010). Although RMSEA, as a measure of the average standardised residual per degree of freedom, should be less than 0.10, a value below 0.08 suggests a good model fit and 0.08 to 0.10 a mediocre fit (Hair et al., 2010). Based on these rules, the measurement model in this study fits the data: χ2/df = 2.12, CFI = 0.931, TLI = 0.922, RMSEA = 0.063 (T1 CFA); and χ2/df = 1.996, CFI = 0.907, TLI = 0.896, RMSEA = 0.077 (T2 CFA).

Structural model and discussion of results

Like the measurement model, the structural model yields a reasonable fit with the data: χ2/df = 2.122, CFI = 0.930, TLI = 0.921, RMSEA = 0.063 (T1 SEM); χ2/df = 2.008, CFI = 0.905, TLI = 0.895, RMSEA = 0.077 (T2 SEM). In the tourism context (T1), the results of the SEM (see Figure 3) suggest that public self-awareness positively affects moral obligation (β = 0.181, p < 0.01) and PEB intention (β = 0.246, p < −.001), and that moral obligation has a positive impact on PEB intention (β = 0.162, p < 0.05). However, the results for the post-visit home context (see Figure 4) show that the influence of public self-awareness on moral obligation (β = 0.132, p > 0.05) and PEB intention (β = −0.008, p > 0.05) appears to be limited, yet moral obligation still shows a strong impact on PEB intention (β = 0.483, p < 0.001). These results support H1 and H2, as public self-awareness only significantly affects moral obligation and PEB intention in the tourism but not the home context. H3 is also supported, suggesting consistent positive impacts of moral obligation on PEB intention across tourism and home contexts.

Figure 3. SEM results for tourism context

Figure 4. SEM results for one-month post-visit home context.

These findings demonstrate that public self-awareness may be a critical antecedent of moral consideration and PEB intentions in a tourism context, in the presence of other people, but is less important in a private, home context. Although we did not hypothesise the intention – behaviour link, PEB intention is found to have a significantly strong influence on PEB in the post-visit home context (see Figure 4), which is consistent with previous research showing that intention is a good predictor of behaviour within the same context (Wu et al., 2021b).

Notably, in both contexts, perceived anonymity has significant negative impacts on PEB intention (T1 β = −0.231, p < 0.01; T2 β = −0.255, p < 0.01), supporting H4, and strong positive effects on moral disengagement (T1 β = 0.618, p < 0.001; T2 β = 0.471, p < 0.001), supporting H5. The effect of moral disengagement on PEB intention appears not to be significant in the SEM models (T1 β = −0.101, p > 0.05; T2 β = −0.058, p > 0.05). However, the results of the correlation analysis, showing that moral disengagement is negatively and significantly associated with PEB intention, suggest possible suppression effects. In other words, a change in the significance (or direction) of the relationship between an influential factor and the dependent factor may result when a third influential factor is added to the model (Lankau et al., 2006).

Thus, following Lankau et al.’s (2006) approach to examining suppression effects, the hypothesised factors influencing PEB intention were inserted into the model step by step. Table 2 shows the model fit information for each model. Contrary to the full model results, a significantly negative impact of moral disengagement (β = −0.282, p < 0.001) on PEB intention is found in Model 1 (β = −0.282, p < 0.001), Model 2 (β = −0.147, p < 0.05), and Model 4 (β = −0.135, p < 0.05). The effects of other influential factors are consistent with the results of the full model.

Table 2. Suppressor effect analysis.

T1(n = 285) Predicted constructs	PEB Intention (dependent variable)
	Full model	Model 1	Model 2	Model 3	Model 4
Moral disengagement	−0.101	−0.282***	−0.147*	−0.111	−0.135*
Perceived anonymity	−0.231**		−0.237***	−0.246***	−0.234***
Moral obligation	0.162*			0.209**
Public self-awareness	0.246***				0.279***
Model fit	χ2/df = 2.122; CFI = 0.930, RMSEA = 0.063	χ2/df = 2.270; CFI = 0.892, RMSEA = 0.077	χ2/df = 2.669; CFI = 0.894, RMSEA = 0.077	χ2/df = 2.639; CFI = 0.896, RMSEA = 0.076	χ2/df = 2.610; CFI = 0.898, RMSEA = 0.075
T2 (n = 170) Predicted constructs	PEB Intention (dependent variable)
	Full model	Model 1	Model 2	Model 3	Model 4
Moral disengagement	−0.058	−0.463***	−0.361***	−0.054	−0.356***
Perceived anonymity	−0.255**		−0.180*	−0.260**	−0.187*
Moral obligation	0.483***			0.478***
Public self-awareness	−0.008				0.053
Model fit	χ2/df = 2.008; CFI = 0.905, RMSEA = 0.077	χ2/df = 2.053; CFI = 0.899, RMSEA = 0.079	χ2/df = 2.049; CFI = 0.900, RMSEA = 0.079	χ2/df = 1.990; CFI = 0.906, RMSEA = 0.077	χ2/df = 2.052; CFI = 0.900, RMSEA = 0.079

Note: a. *p < 0.05, **p < 0.01, ***p < 0.001, b. PEB intention = pro-environmental behavioural intention.

Combining the findings of the correlation analysis with the additional analysis for suppressor effects, we can still conclude that moral disengagement has a negative impact on PEB intention, supporting H6, as seen in the literature on tourism and other contexts (Wu et al., 2021a). Unlike previous studies (Mkono, 2018; Peng et al., 2022; Steidle & Werth, 2014), the proposed negative impact of perceived anonymity is found not to be significant for public self-awareness in either the tourism or the post-visit home context (T1 β = −0.049, p > 0.05; T2 β = 0.069, p > 0.05). Hence, H7 is not supported.

Discussion and conclusion

This study responds to calls to explore contextual inconsistencies in mechanisms influencing PEB and PEB intentions (Holmes et al., 2021; Whitmarsh et al., 2018; Xu et al., 2020). Our study goes beyond understanding what predicts PEB intentions in tourism and home contexts and the inconsistencies of these predicting effects by examining critical components of the deindividuation effect (public self-awareness and perceived anonymity) as predictors of moral considerations (moral obligation and moral disengagement) toward PEB between contexts.

Our findings reveal a nuanced view of contextual inconsistency that manifests in two distinct dimensions. First, variability in factor scores across contexts is observed in the results of the paired-sample t-test (see Figure 2). The significant differences in moral obligation, moral disengagement, perceived anonymity, and public self-awareness between the tourism and home contexts suggest that any conclusions drawn from research are constrained by the environment in which it is carried out, and that context-specific sustainability interventions must be designed (Whitmarsh et al., 2018; Wu et al., 2021b). Second, we find dynamic relationships between the studied factors from the tourism to home contexts. Specifically, our findings demonstrate that moral obligation, moral disengagement, and perceived anonymity are consistently associated with PEB intentions in both tourism and post-visit home contexts, while public self-awareness only significantly affects PEB intentions in the tourism context.

The limited impact of public self-awareness in the post-visit home context may be attributable to ‘the watching eye phenomenon’ (Bodur et al., 2015; Coker, 2020). As suggested by deindividuation theory (Prentice-Dunn & Rogers, 1982), public self-awareness is activated by the condition of perceiving being observed by others. The transition from a public space (tourism context) to a more personal environment (home context) will deactivate public self-awareness. This is in line with our finding that in the tourism context, perceived public self-awareness has a significant positive impact on both moral obligation and PEB intentions, while in the home context, with a significant decrease in perceptions of being watched by others, the effect of public self-awareness appears to be no longer salient. This supports Whitmarsh et al.’s (2018) finding that different predictors may explain PEB and PEB intentions across home and tourism contexts.

This inconsistent effect of public self-awareness cautions against assuming that the mechanism of deindividuation functions consistently across contexts. Perceived anonymity, as a key component of deindividuation, consistently results in higher moral disengagement and lower PEB intentions in both tourism and home contexts, whereas public self-awareness only works on moral self-regulation with regard to PEB intentions in tourism settings, but not at home. Other important predictors of PEB intentions or PEB in home contexts, such as the availability of recycling facilities (Whitmarsh et al., 2018; Wu et al., 2021b), may suppress the impact of public self-awareness. The consistently strong impacts of perceived anonymity on both moral disengagement and PEB intentions echo the results in the extant tourism literature and studies in other contexts (Su et al., 2022), showing that perceived anonymity is a key psychological process underlying moral disengagement and unethical actions. In other words, people are more likely to morally disengage from PEB when they feel safely anonymous, whether as part of a crowd of tourists or in a home context (Sharma, 2020; Su et al., 2022).

Our finding of a link between perceived anonymity and public self-awareness differs from our expectation based on previous literature, that people’s perceptions of anonymity will reduce their public self-awareness (Hwang & Lee, 2019; Peng et al., 2022). These discrepant results can be explained by deindividuation research suggesting that control of human behaviour through anonymity is both inconclusive and highly dependent on the situation (Kim et al., 2019; Lea et al., 2001). Anonymity may only effectively impact on public self-awareness when it occurs in a ‘group’ context (Diener, 1979; Su et al., 2022), where people are identified as group members rather than as individuals (Kim et al., 2019).

Theoretical implications

Although contextual inconsistency in PEB and PEB intentions has been identified in previous scholarly work (Ballantyne et al., 2018; Holmes et al., 2021; Hughes, 2013), much remains to be elucidated regarding what predicts and what constrains their formation across contexts. Our research explains the mechanisms through which the processes of deindividuation and moral disengagement together may inhibit individuals’ self-stated PEB intentions and PEB in both tourism and home contexts. It thus complements and extends previous insights into contextual inconsistency in PEB and PEB intentions. Furthermore, we identify an important theoretical underpinning of contextual inconsistency of PEB from a novel theoretical perspective, with valuable comparative data between contexts.

Only recently have tourism studies mentioned the deindividuation effect as a theoretical basis for investigating unethical behaviours (Hwang & Lee, 2019; Sharma, 2020), paving the way for a more nuanced understanding of the mechanism of deindividuation in restraining tourists’ PEB. Our findings reveal that the deindividuation state restrains PEB intentions owing to increased perceived anonymity and reduced public self-awareness in tourism settings, while in the home context, only the effect of perceived anonymity is salient. Therefore, our study advances knowledge of deindividuation and PEB by illustrating how deindividuation affects moral regulation of PEB. Our study also adds value to the deindividuation literature by revealing differences in the predictive impacts of deindividuation components across contexts.

Furthermore, our conceptual framework disentangles interrelationships between critical constructs of deindividuation theory, deepening understanding of the relationship between perceived anonymity and public self-awareness by showing that perceived anonymity does not necessarily reduce public self-awareness. In contrast to previous deindividuation literature, which asserts that anonymity is closely linked with public self-awareness (Prentice-Dunn & Rogers, 1982; Steidle & Werth, 2014), our findings echo previous criticism of classic deindividuation theory for overlooking group conditioning (Kim et al., 2019). Our study also provides significant theoretical insights by identifying a consistent antecedent of moral disengagement and PEB intentions across tourism and home contexts, and showing how moral disengagement is linked to deindividuation in both contexts.

Practical implications

Our findings of contextual inconsistency in the mechanisms influencing PEB intentions suggest that behavioural change strategies (e.g. environmental campaigning) should be context-specific. Policymakers should not rely on environmental education programmes alone to induce tourists’ PEB change, nor make assumptions about the impacts of daily-life PEB on tourists’ behaviours while on holiday. Although direct interactions with nature may open a window to increase tourists’ PEB intentions, their return to daily life settings is likely to trigger habitual behaviours, whether environmentally responsible or not.

The tourism industry has already learned that public self-awareness and perceived anonymity can be used to market a sense of freedom to misbehave, as exemplified in the well-known slogan ‘What happens in Vegas stays in Vegas’. The challenge is to learn how to tap into variables that will make people behave morally and consistently across different contexts. To achieve this, understanding how PEB and PEB intentions may differ across contexts is essential in designing effective behavioural change strategies (Guagnano et al., 1995; Whitmarsh et al., 2018; Xu et al., 2020). Context-specific perceptions may also help to explain the challenges of promoting positive sustainability spillover across time and context (Nilsson et al., 2017) and why a simple cookie-cutter approach cannot be used to promote PEB.

Our findings can be used by tourism industry practitioners and policymakers alike to nudge individuals to engage in PEB. In both contexts, strategies that help to reduce perceived anonymity, mitigate moral disengagement, and enhance moral obligation need to be developed, as well as communication messages that help to increase tourists’ sense of public self-awareness. For example, for tourism destinations, sharing positive stories of tourists (e.g. visitors joining volunteering projects onsite) through both social media and traditional media channels onsite (e.g. tourist information boards or TV screens located in the tourist centre) would help to increase tourists’ public self-awareness, thus fulfilling their need for reputation and encouraging more PEB.

As suggested by our findings, individuals are more likely to justify their environmentally harmful behaviours or morally disengage from PEB when they feel that they are anonymous. To reduce the negative impact of perceived anonymity, nature-based destinations might consider providing guided tours for small groups of tourists and implementing anonymous reporting channels for people to report irresponsible behaviours within the destination. These interventions, together with onsite welcome sessions highlighting individual responsibility and the importance of sustainable actions, would also help to mitigate the possibility of moral disengagement from PEB. With regard to home contexts, face-to-face gatherings might be organised by local communities to encourage social interactions among residents, build a strong sense of connectedness, and reduce feelings of anonymity. Communicating sustainability initiatives through online platforms (e.g. neighbourhood social media groups) and newsletters would also help to reduce the sense of anonymity and diffusion of responsibility that lead to moral disengagement from daily PEB.

Limitations and future research

This study has several limitations that suggest interesting directions for future research. First, in our study we only compared how deindividuation affects PEB intentions and PEB through moral regulation between tourism and home contexts. Furthermore, our onsite survey was conducted in a nature-based destination, a context that may have had a natural influence on sustainable intentions and behaviours. Future studies might expand the research context to workplaces and other tourism-related contexts (e.g. accommodation) to extend knowledge of PEB inconsistency across contexts. Second, our study was based on data from China, which is likely to differ from research based on a Western context (e.g. Holmes et al., 2021; Whitmarsh et al., 2018). Cross-contextual and cross-cultural studies might validate our research framework and findings. Longer time frames or experimental designs might also increase external validity.

Third, our study focuses mainly on perceived anonymity and public self-awareness as critical components of the deindividuation effect. Further research might include the effect of private self-awareness, and compare private and public self-awareness. As suggested by Kim et al. (2019), in anonymous contexts, private and public self-awareness may affect individuals’ behaviour independently or concurrently. Fourth, our inconclusive findings on the impact of anonymity on public self-awareness suggest the need for studies to investigate whether personal versus group identity may operate concurrently in tourism contexts, and how and when one’s group and personal identities conflict. Travelling alone and travelling within a group might be compared to generate interesting insights. Finally, although we made efforts to lower potential social desirability bias during our onsite survey, it could not be completely eliminated, as is consistent with previous research. This calls for an effective procedure to deal with this issue in the future.

Disclosure statement

No potential conflict of interest was reported by the author(s).

Additional information

Funding

This work was supported by National Natural Science Foundation of China [grant number: 72074233].

Appendix 1.

Confirmatory factor analysis

Table

Factor	Item	T1 (tourism context), n = 285				T2 (home context), n = 170
		Factor loading	Composite reliability	Cronbach’s alpha	AVE	Factor loading	Composite reliability	Cronbach’s alpha	AVE
Public Self-Awareness	Right now, I am self-conscious about the way I look	0.585	0.820	0.807	0.609	0.920	0.965	0.963	0.901
	Right now, I am concerned about what other people think of me	0.850				0.981
	Right now, I am concerned about the way I present myself	0.873				0.946
Moral obligation	I should protect the environment	0.915	0.900	0.897	0.750	0.873	0.947	0.945	0.855
	It’s important that people in general protect the environment	0.870				0.967
	Because of my own values, I feel an obligation to behave in an environmentally friendly way	0.810				0.932
Perceived Anonymity	It is difficult for others to identify me as an individual	0.909	0.956	0.956	0.879	0.939	0.974	0.973	0.925
	I am confident that others do not know who I am	0.950				0.992
	I believe that my personal identity remains unknown to others	0.953				0.954
Moral disengagement	For the convenience of my friends/relatives, it’s alright to perform less eco-friendly behaviours sometimes	0.509	0.887	0.872	0.536	0.770	0.902	0.897	0.572
	Considering the pollution caused by big enterprises, it’s hardly a sin for us to perform environmentally harmful behaviours sometimes	0.692				0.806
	People cannot be blamed for doing environmentally harmful behaviours if their leaders pressure them to do so	0.780				0.732
	In contexts where everyone is engaged in environmentally harmful behaviours, we can’t be blamed for doing it too	0.825				0.788
	Conducting environmentally harmful behaviours for a little while doesn’t cause big harm to the environment	0.881				0.882
	Wearing animal fur is okay; anyhow, animals are not human beings	0.788				0.709
	If I’m engaged in less eco-friendly behaviours, it’s probably because the government is not doing enough to support the environment	0.572				0.571
Factor	Item	T1 (tourism context), n = 285				T2 (home context), n = 170
		Factor loading	Composite reliability	Cronbach’s alpha	AVE	Factor loading	Composite reliability	Cronbach’s alpha	AVE
Pro-environmental intention	Conserve resource and energy	0.530	0.859	0.857	0.435	0.661	0.903	0.900	0.540
	Recycle	0.652				0.743
	Sort garbage	0.555				0.733
	Use ‘green’ (non-plastic) shopping bags	0.698				0.709
	Remind others to avoid doing environmentally harmful behaviours	0.701				0.777
	Look for environmental information on TV, in print, or on the Internet	0.771				0.856
	Donate money to support environmental conservation	0.715				0.737
	Volunteer my time to projects that help the environment	0.619				0.643
Pro-environmental behaviour	Conserve resource and energy					0.716	0.891	0.886	0.509
	Recycle					0.656
	Sort garbage					0.672
	Use ‘green’ (non-plastic) shopping bags					0.619
	Remind others to avoid doing environmentally harmful behaviours					0.834
	Look for environmental information on TV, in print, or on the Internet					0.883
	Donate money to support environmental conservation					0.677
	Volunteer my time to projects that help the environment					0.601