Individuals make ethical judgements about particular situations grounded in their ethical ideologies and moral philosophies (Barnett et al. 1998; Ko et al. 2019). Andersch et al. (2018) professed that the idealism as operationalised by Forsyth (1980) is grounded in the deontological viewpoint, which presumes that individuals are held responsible for things within their control; and the teleological viewpoint, which assumes that practices that contribute to societal goals are acceptable, even if they cause harm to individuals. In essence, deontology involves ‘modal obligation and justice’ and teleology relates to ‘perceived risk and perceived benefit’ (Agag & Colmekcioglu 2020:8). Accordingly, individuals who are highly idealistic (deontologists) believe any actions that harm other people should be completely averted, while those who are not idealistic (teleologists) do not hold such beliefs and apply more utilitarian approaches where causing harm to others is sometimes inevitable (Agag & Colmekcioglu 2020; Andersch et al. 2018). Furthermore, Agag and Colmekcioglu (2020) professed that where a pro-environmental context is concerned, that those with high levels of idealism think that saving and protecting the natural environment is the right thing to do and that harming the natural environment is completely wrong. Khan and Abbas (2023) added that it is expected that idealism will correlate positively with pro-environmental behaviours and ethical consumption. Those who are concerned about the environment will take decisions that are environmentally ethical and will be concerned about achieving environmental goals (Alyahya et al. 2023). According to Zou and Chan (2019), idealism impacts positively on individuals’ ethical judgement. In this study, idealism involves the extent to which individuals believe they should not harm the environment or others or do anything that could impact their welfare and dignity, as proposed by the Ethics Position Questionnaire of Forsyth (1980) and used subsequently by Andersch et al. (2018).

For the purpose of this study, the focus is on moral judgements regarding environmentally friendly consumption, as reflected by green attitudes. According to Wang, Shen and Chu (2021), attitude is the propensity to back or dislike particular behaviours, objects or ideas. Given the impact humans are having on the environment, green products and services are increasing rapidly in popularity and consumers’ attitudes towards these products and services depend upon whether products and services are seen as environmentally friendly (Sohaib et al. 2022). Pandey and Yadav (2023) stated that attitude influences individuals’ judgement of the benefits and costs of executing a particular behaviour. The main concern for an individual with respect to green (consumption) attitude is to ‘justify and defend the cause of actions, promote its worth, and support the environment’ (Pandey & Yadav 2023:3). Moreover, moral judgements are said to notably influence consumers’ behaviour (Andersch et al. 2018). Studies have considered moral judgements where green consumption is concerned. Sparks and Shepard (2002) considered moral judgements in relation to food production and consumption, while Huang, Ma and Yen (2022) considered moral judgements in the context of green consumption (managing and preventing food waste). To uncover moral judgements, researchers measure consumers’ attitudes (Sparks & Shepard 2002) – in this instance, green attitudes. For this study, green attitude involves the notion of whether an individual thinks environmentally green consumption is rewarding, wise and valuable to practice (Van Tonder, Fullerton & De Beer 2020).

As addressed earlier, within an environmental sustainability context, previous research has established that idealism strongly influences moral judgement of environmental issues (Zaikauskaite et al. 2020). Furthermore, Singhapakdi et al. (1995) explained that idealism impacts the extent to which customers view social responsibility and ethics as important. The authors opined that higher levels of idealism are associated with an increased importance of social responsibility and ethics among customers (Singhapakdi et al. 1995). Additionally, idealism has been identified as an important driver of ethical consumption (Singhapakdi et al. 2000). Customers who are idealists believe that ethical or green behaviours are the right thing to do (Zou & Chan 2019). Because green attitude in this study refers to the extent to which customers believe that ‘practising environmentally-friendly green consumption’ is wise, rewarding and valuable (Van Tonder et al. 2020), which is ethical in its nature, it is hypothesised that:

As highlighted in the introduction, there is evidence that customers engage in helping behaviours in a green context by solving the problems of other customers and advising other customers regarding their green purchases (Alsaggaf et al. 2019; Deng & Yang 2022; Van Tonder et al. 2023). Helping generally represents a dimension of customer citizenship behaviour and is considered as extra-role behaviour that customers undertake voluntarily at their own discretion and that may benefit a company (Gong & Wang 2022; Groth 2005; Hwang & Lyu 2020; Yi & Gong 2013). However, previous research suggests that when customers assist other customers not to purchase a given brand, it should only be perceived as a form of social support and not a dimension of citizenship behaviour, because the brand that is being avoided is being disadvantaged (Petzer et al. 2023). From a green perspective, advice to green purchasers in relation to the avoidance of harmful brands may advantage companies selling authentic green brands. Given that this study focusses on helping in relation to the avoidance of harmful brands to assist green purchasers, helping in the current study’s context is deemed a dimension of customer citizenship behaviour. For the purpose of this study, green helping behaviours focus on the degree to which customers engage in helping others to avoid products and services (brands) that could be harmful to society by adapting the measure for helping as a customer citizenship behaviour (Van Tonder et al. 2023; Yi & Gong 2013) to focus on avoidance of products and services that may harm society (Gurrieri & Drenten 2019; Petzer et al. 2023).

There are opposing views with respect to whether favourable attitudes lead to aligned behaviours, coined as the attitude-behaviour gap (Mühlthaler & Rademacher 2017; Van Tonder et al. 2023). However, Hwang and Lyu (2022) found significant and positive relationships between customer attitude and selected customer citizenship behaviours, such as feedback to service providers, tolerating others’ mistakes, advocacy and helping other customers. In addition, Wang et al. (2022) confirmed that where green attitude is concerned, resultant behavioural responses have been evident. Furthermore, Van Tonder et al. (2020) found that where customer citizenship behaviours are concerned, there are significant and positive relationships between these behaviours, including feedback to service providers and advocacy and green attitude. Accordingly, it is believed that in the context of the current investigation, those who exhibit a green attitude will engage in citizenship behaviours to help others avoid brands that may harm society. Therefore, it is hypothesised that:

As mentioned, idealism in this study involves the degree to which customers consider the dignity and welfare of others to be a core concern of a society as well as the concern that members of society should not be harmed in any way (Andersch et al. 2018). Idealists believe the most suitable deeds are those that address the above-mentioned concerns, where the dignity, welfare and possible harm of others are concerned (Ko et al. 2019). Because green customer helping behaviours involve customers’ citizenship behaviours, specifically helping others to avoid brands that may harm society (Solesvik 2017), it is hypothesised that:

There has been a growing focus on cynicism with the move from modernity to post modernity, and cynicism has been viewed as having the ability to provide deeper ‘insights into the contemporary human condition’ (Mikkonen, Moisander & Fırat 2011:102). Modern cynics can be portrayed as informed social entities who are part of contemporary society and who ensure they are not taken advantage of, benefitting from an unobstructed view of how things operate (Bertilsson 2015). Core beliefs associated with cynicism allow consumers to reclaim control of their lives, thus representing a coping strategy for consumers (Odou & De Pechpeyrou 2011). Furthermore, cynicism is more lasting than instrumental and plays an important role in defining consumers’ morality (Odou & De Pechpeyrou 2011). Moreover, it can be described as an emotional defiance, rather than being of a cognitive nature like scepticism (Ryu & Jun 2019). Consumer cynicism or customer helper cynicism further denotes a stable and learnt consumer attitude towards the consumer marketplace that is perceived as damaging because of the alleged widespread unscrupulousness of businesses operating in the consumer marketplace (Helm, Moulard & Richins 2015). In addition to the marketplace, cynicism can be directed at a particular business (Chylinski & Chu 2010). Bertilsson (2015:461) asserted that there are three focal ‘cynical discursive practices’, which involve cynicism towards the self, towards other consumers and towards the market, with the latter being relevant to this study because it involves a situation ‘where consumers display an enlightened disbelief in the morality of brands and branding’. Mikkonen et al. (2011) viewed consumer cynicism as a counteracting discursive practice and strategy that create and maintain frustration, uncertainty as well as disenchantment towards the marketplace and business.

Accordingly, in the context of this study, customer helper cynicism involves the extent to which an individual believes that businesses will do almost anything to make a profit, such as manipulating customers, making any sacrifice necessary, taking shortcuts and acting without consequences (Helm et al. 2015; McCullough, Rynarzewska & Fullerton 2022). When customers exhibit strong cynicism schemas in general, the chances are good that these customers will also be more cynical in the marketplace (Indibara & Varshney 2021). Aligned with this notion, it is further postulated that cynicism possibly explains several customer behaviours directed at the marketplace (Helm et al. 2015) and that it leads to anti-consumption behaviours (Odou & De Pechpeyrou 2011). Cynicism involves a range of affective, cognitive and behavioural responses from customers and can be articulated through customers’ ‘initial suspicion, defensive attempts, and eventual alienation’ (Chylinski & Chu 2010:799). Cynical consumers may decide to exit from the marketplace or act in a non-conforming manner (Indibara & Varshney 2021).

Given the above perspectives, it seems plausible that the relationship between idealism and green attitudes may be weakened when the green customer helper has a cynical attitude towards businesses in the consumer marketplace. Green customer helpers who support idealistic beliefs may be less likely to think that environmentally green consumption is rewarding, valuable and wise to practice if they believe that businesses will do almost anything to make a profit, such as manipulating customers, making any sacrifice necessary, taking shortcuts and doing anything they can without consequences.

Moreover, given the plausible relationship between green attitude and green customer helping behaviour, a moderated mediation effect may be present. Consumers with idealistic perspectives may have green attitudes and subsequently engage in green helping behaviours. However, this proposed indirect effect may be moderated by customer helper cynicism, assuming that customer helper cynicism moderates the relationship between idealism and green attitude. Considering the fundamentals of customer helper cynicism, green attitude is less likely to mediate the indirect effect when customer helper cynicism beliefs are high. Against the background provided above, it is hypothesised that:

Figure 1 provides an outline of the research model of this study.

A self-administered questionnaire was developed to survey consumers in the US and South Africa, who were at least 18 years old (being adults) and consented to participate in the study addressing matters of ethical concern. Adult respondents were targeted, because they would be able to provide rationalised responses when completing the survey. The study was duplicated in the US with the intention to seek further validation for the established model. The questionnaire included a cover letter that provided more information about the objective of the study, the type of respondent being surveyed, the treatment of the research data and the project leaders. Contact details were also provided, should the respondents want more information about the study results. The respondents had to give their consent, before they were allowed to continue with the survey. The first part of the survey sourced the demographic information from the respondents. The section included a question about the age of the respondents that was helpful to ensure that no respondent under the age of 18 would participate in the study. The latter part of the survey collected information about the constructs that were examined in this study.

To measure the study’s four constructs – namely idealism, customer helper cynicism, green attitude and green customer helping behaviour – a six-point labelled Likert-type scale was used, where 1 implied ‘strongly disagree’ and 6 implied ‘strongly agree’. Idealism was measured using five items from a scale adopted from Andersch et al. (2018), while customer helper cynicism was measured using a shortened (four-item) version of the scale from Helm et al. (2015), as adapted and pre-tested by McCullough et al. (2022). Green attitude was measured using three items from a scale adapted from Van Tonder et al. (2020) and green customer helping behaviour was measured using four items from a scale adapted from Yi and Gong (2013). Annexure 1 summarises the scales’ items that were employed in the survey.

In the US, data were collected with the aid of an online consumer panel (Dynata). In South Africa, the survey was distributed to a national consumer panel of a field services company (Consulta). In each country, the consumer panel received an email invitation from the respective service provider, explaining the study’s intent and asking them to complete the online self-administered survey. Upon voluntarily consent, the respondents were allowed to proceed to the survey. The researchers adhered to the prescribed ethical guidelines of their respective universities and no personal data were sourced from the respondents. Table 1 outlines the demographic profile of respondents for the South Africa and the US studies.

It is evident from Table 1 that the typical respondent taking part in the South Africa study was male or female, older, married, well-educated, and full-time employed at the time the survey was completed. The typical respondent who participated in the US study was also a well-educated male or female of a more mature age, married or single, and full-time employed or retired when the survey was completed.

Confirmatory factor analysis (CFA) was performed in Mplus 8.5 to assess measurement model validity (Hair et al. 2019). The Mplus programme was favoured, because it allowed for the use of the maximum likelihood procedure in the assessment and to obtain more accurate results (Byrne 2001). Composite reliability (CR) measures were employed to assess construct reliability (Hair et al. 2019). The factor scores from the analysis were retained, and the research hypotheses were subsequently assessed using Model 7 of the Hayes (2013) PROCESS macro in RStudio (RStudio Team 2020). The mediation hypothesis was tested, with the aid of bootstrap resampling (5000 draws and 95% confidence interval [CI]).

The authors adhered to all ethical standards as prescribed by their respective institutions, with ethical clearance number: NWU-00016-21-A4.

Individual measurement models were first compiled for South Africa and the US, comprising the four constructs of interest to this study. Within both models, it became evident that model fit could be improved by deleting the first item of the idealism scale (that obtained a standardised factor loading lower than 0.5) and by acknowledging covariances between the third and fourth items of the green customer helping behaviour scale. Subsequently, both measurement models delivered adequate model fit statistics (Hair et al. 2019) – South Africa: χ² (df = 83) = 180.046; (χ²/df = 2.17), comparative fit index (CFI) = 0.96, Tucker-Lewis index (TLI) = 0.95, and root mean square error of approximation (RMSEA) = 0.068; and US: χ² (df = 83) = 210.020, (χ²/df = 2.53), CFI = 0.97, TLI = 0.96, and RMSEA = 0.066.

In both countries, all scale items of the revised measurement models exceeded 0.5 and loaded significantly onto their individual constructs (p < 0.001). No convergent or discriminant validity issues were detected either. For all constructs measured, the average variance extracted (AVE) values exceeded 0.5 and the shared variance between any pair of constructs was never larger than their separate AVE values. In both data sets, the CR measures of all four constructs also far exceeded the 0.7 cut-off, providing further evidence that reliability was obtained (Fornell & Larcker 1981; Hair et al. 2019).

Subsequently, a multi-group CFA was conducted. Within the multi-group model, the first item of the idealism scale was omitted again, and the covariance between the third and fourth items of the green customer helping behaviour scale was accounted for in the model. Measurement invariance was achieved, as evidenced by the delta change between the CFI values of the configural, metric, and scalar model results that never exceeded 0.01 (Cheung & Rensvold 2002). Moreover, satisfactory model fit statistics were obtained (Hair et al. 2019): χ² (df = 188) = 491.630; (χ²/df = 2.62); CFI = 0.96; TLI = 0.95, and RMSEA = 0.073. The chi-square contribution for the South Africa sample was 246.07 and the chi-square contribution for the US sample was 245.56. As further indicated in Table 2, the multi-group’s standardised factor loadings exceeded 0.5 and loaded significantly onto their individual constructs (p < 0.001).

There were no convergent or discriminant validity issues in the multi-group model either. Table 3 shows that the AVE values for both the South Africa and US samples are above 0.5 and the shared variance between any pair of constructs are not larger than their separate AVE values.

Table 2 shows that, in both data sets, the CR measures of all four constructs far exceed the 0.7 cut-off, providing further evidence that reliability was obtained (Fornell & Larcker 1981; Hair et al. 2019). Accordingly, latent variable factor scores of the multi-group model were saved and carried over to RStudio for further analysis of the research model.

Table 4 summarises the output, as produced from the analysis of Model 7 of the Hayes (2013) PROCESS macro in RStudio.

Individual analysis was conducted by country. As per Table 4 (Model 7, Stages 1 and 2), the unstandardised coefficients studied are all significant, except for the relationship between customer helper cynicism and green attitude within the US sample. Within the South Africa sample, the lowest coefficient was -0.24 and the highest coefficient was 0.47. The coefficients obtained within the US sample were somewhat lower than the corresponding coefficients from the South Africa sample, ranging from -0.13 to 0.55. Moreover, in both samples, idealism had a strong effect on green attitude (South Africa: coefficient = 0.47, p < 0.001; US: coefficient = 0.41, p < 0.001), while the relationship between idealism and green customer helping behaviour was weaker (South Africa: coefficient = 0.23, p < 0.05; US: coefficient = 0.18, p < 0.05), but still significant. Furthermore, green attitude had a stronger effect on green customer helping behaviour in the US sample, compared to the South Africa sample (South Africa: coefficient = 0.38, p < 0.001; US: coefficient = 0.55, p < 0.001).

Table 4 (Model 7, Stage 1) further depicts that the coefficient of the product between idealism and customer helper cynicism is significant and negative in both samples (South Africa: coefficient = -0.24, p < 0.001; US: coefficient = -0.13, p < 0.05). Accordingly, in both samples, customer helper cynicism moderated the relationship between idealism and green attitude. In the South Africa sample, the moderated direct effect of idealism on green attitude varied between high (point estimate: 0.24, p < 0.01), moderate (point estimate: 0.43, p < 0.001), and low (point estimate: 0.65, p < 0.001) levels of customer helper cynicism. Similarly, in the US sample, the moderated direct effect of idealism on green attitude varied between high (point estimate: 0.33, p < 0.001), moderate (point estimate: 0.44, p < 0.001), and low (point estimate: 0.55, p < 0.001) levels of cynicism.

Additionally, in the South Africa sample, the index of moderated mediation was significant (coefficient = -0.092, standard error [SE] = 0.05, 95% bootstrap CI = -0.189 to -0.009). These findings indicated that customer helper cynicism moderated the indirect effect of idealism on green customer helping behaviour through green attitude (Hayes 2018). The indirect effect of idealism on green customer helping behaviour was larger at lower levels of customer helper cynicism (coefficient = 0.24, SE = 0.08, 95% bootstrap CI [0.100, 0.406]), significant and weaker at moderate levels of customer helper cynicism (coefficient = 0.16, SE = 0.05, 95% bootstrap CI [0.071, 0.269]), and even weaker at high levels of customer cynicism (coefficient = 0.092, SE = 0.047, 95% bootstrap CI [0.020, 0.203]). In the US sample, the index of moderated mediation was not significant (coefficient = -0.071, SE = 0.04, 95% bootstrap CI = -0.158 to 0.011). These findings indicated that customer helper cynicism did not moderate the indirect effect of idealism on green customer helping behaviour through green attitude in the US sample (Hayes 2018). In view of these perspectives, H1–H4 were supported in both samples, while H5 was only supported in the South Africa sample. Figure 2 illustrates the research results.