Introduction

Unethical consumer behavior is becoming an increasingly global issue, with recent data indicating a surge in such behavior. For example, in the United States, consumer return fraud is suspected to have increased by 76% since 2018 (National Retail Federation, 2019). Similarly, increased consumption of counterfeits has resulted in such products accounting for 3.3% of total world trade (Organisation for Economic Co-operation & Development, 2019). More specifically, changes in the macro environment, characterized by increased Internet penetration rates, offer a unique array of misbehaviors and distinctive ethical considerations, which transcend national borders (Freestone & Mitchell, 2004; Nawrotzki, 2012). For example, digital piracy costs the US economy an estimated 29.2 billion to 71 billion USD each year (Jugović Spajić, 2022). Meanwhile, improper sharing of passwords for streaming services leads to billions of dollars of lost revenue for streaming services (Anderson, 2021).

Unethical consumer behavior has a diverse range of negative effects for businesses and other stakeholders including direct and indirect financial losses, psychological impacts, and adverse consumption experiences for other consumers (Harris & Reynolds, 2003). Additionally, such behavior presents costs and challenges for governments to effectively protect citizens, the environment, and impose adequate legislation. Thus, it is imperative for managers and policymakers to understand how consumers form ethical judgments, so that appropriate preventative measures can be developed to minimize the consequences of unethical consumer behaviors.

Over the last three decades, researchers in consumer behavior and marketing have shown significant interest in understanding consumers’ ethical judgments, intentions, and behaviors. Frequently explored variables in consumer ethics are deontological and teleological evaluations, which are posited to directly affect ethical judgments (Hunt & Vitell, 1986, 2006) and empirically shown to affect ethical intentions (e.g., Chan et al., 2008; DeConinck & Lewis, 1997). Deontological evaluations relate to the perceived inherent rightness/wrongness of each course of action, whereas teleological evaluations refer to the perceived goodness/badness relative to the perceived consequences of each alternative (Hunt & Vitell, 2006). For example, if a consumer receives too much change in a service encounter, their ethical judgment and subsequent intention may result from both a deontological evaluation and a teleological evaluation. In this example, a deontological evaluation will relate to how right/wrong they deem alternatives to be, such as keeping versus returning the surplus, whereas a teleological evaluation will assess the goodness/badness of the consequences, such as their gain, their loss, or the service provider’s loss.

Previous studies provide a wealth of empirical validation for the effects of deontological and teleological evaluations on ethical judgments and intentions. However, each type of moral evaluation’s respective weighting and magnitude of effect remains a disputed topic. Several studies provide empirical support for the relative strength of deontological evaluations (Chang, 2021; Hunt & Vásquez-Párraga, 1993; Vitell et al., 2001), while other studies suggest the greater relative impact of teleological evaluations (Chan et al., 2008; Mayo & Marks, 1990). Accordingly, critical questions remain: To what extent do deontological and teleological evaluations influence ethical judgments and intentions? And under what circumstances?

We conducted a meta-analysis of 316 effect size estimates from 53 research articles in an attempt to uncover the relative effect of deontological and teleological evaluations. Various moderators were identified to examine the circumstances under which these effects might differ. These moderators relate to: (1) the contextual elements of the ethical issue, (2) the stakeholders, and (3) the methodological characteristics of primary studies.

Our study makes a number of contributions to the literature. From a theoretical perspective, since deontological and teleological evaluations are key variables in the ethical decision-making literature, assessing their differential impact significantly contributes to our understanding of consumer behavior in ethical contexts. Findings allow for an empirical assessment of the distinct effects of deontological and teleological evaluations and the contingency factors related to these effects. Our research also has practical contributions in relation to preventative measures that can curtail unethical consumer behavior. Two distinct categories of preventative measures have been proposed in prior studies. These include educational approaches, which impose moral constraints on consumers, which are consistent with a deontological approach to ethics (Fullerton & Punj, 2004). Further, deterrence strategies, which communicate sanctions and consequences of harmful actions, are consistent with a teleological approach to ethics (Dootson et al., 2014). Thus, understanding the contexts and circumstances under which deontological or teleological evaluations are more/less impactful will allow managers and policymakers to effectively implement preventative measures to reduce unethical consumer behavior. Finally, our findings provide valuable insights for future research, as well as important managerial implications.

Meta-Analytical Framework and Conceptual Development

Deontology and teleology are the two mainstream ethical theories that have predominated Western philosophical thought over the last three centuries. Deontology, or morality based on rules and obligations, is in the tradition of the philosophical works of Immanuel Kant (Kant, 1785/1993), whereas teleology, or morality based on the consequences of actions, is in the tradition of the philosophical works of Jeremy Bentham (1789/1996). Both theories have been embedded in ethical decision-making theories. Specifically, Hunt and Vitell (1986, 2006) utilize these theories in their H-V theory of marketing ethics, which outlines the individual ethical decision-making process and provides the conceptual foundation for our investigation. At the core of this process model, ethical judgments are determined by the joint impact of two types of moral evaluations derived from these theories: deontological and teleological evaluations (Hunt & Vitell, 1986, 2006).

Figure 1 represents the key relationships between the variables in the ethical decision-making process adapted from Hunt and Vitell (1986, 2006). An individual enters the ethical decision-making process upon recognition of an ethical issue (Rest, 1986). Subsequently, they may form a deontological evaluation, which relates to the inherent rightness/wrongness of each perceived alternative course of action. Concurrently, they may formulate a teleological evaluation, which refers to the goodness/badness relative to the perceived consequences of each alternative. For example, digital piracy is an evolving global issue that constitutes an example of unethical consumer behavior, which results in drastic losses for entertainment industries (Jugović Spajić, 2022). Thus, when a consumer is confronted with the option to consume or download pirated digital content, their ethical judgment may be formed based on how right/wrong the perceived courses of actions are (i.e., to pirate vs. not pirate), as well as how good/bad the consequences of those actions are, potentially encompassing losses and gains to all relevant stakeholders.

Fig. 1
figure 1

Key variable framework

Various interdisciplinary critics argue for the mutual exclusivity of deontological and teleological evaluations and often investigate individuals’ adoption of deontologically or teleologically favorable outcomes (e.g., Friesdorf et al., 2015). However, we focus on the ethical decision-making process for which the combined effect of deontological and teleological evaluations has received considerable conceptual and empirical validation (Conway & Gawronski, 2013; Love et al., 2020; Macdonald & Beck-Dudley, 1994). Accordingly, we treat them as two independent variables that have distinct effects on ethical judgments and intentions.

The relative influence of the two types of moral evaluations remains unexplored, which is the purpose our meta-analysis. Inconsistencies in previous findings could be attributed to circumstances where one type of moral evaluation has a greater impact than the other, as well as circumstances where one might have no effect at all. Regardless, both types of moral evaluations are relevant to consumer ethics and have important implications for marketing practitioners (Bateman & Valentine, 2010; Vitell et al., 2001). Understanding the effect and magnitude of each moral evaluation and their contingencies provides insights and allows for the effective implementation of interventions to limit unethical behavior and associated negative consequences. In order to provide theoretical underpinnings for the meta-analysis, we first propose hypotheses regarding the direct relationships between each type of moral evaluation with ethical judgments and intentions. We then focus on the development of a framework that identifies several contingency factors that may moderate these relationships. These factors relate to (1) the contextual elements of the ethical issue, (2) the stakeholders, and (3) the methodological characteristics of primary studies. The conceptual models of this research are in Fig. 2A and B.

Fig. 2
figure 2

A Conceptual model. B Conceptual model

Deontological Evaluations and Ethical Judgments/Intentions

Deontological evaluations of an ethical issue are determined by the perceived alternative course(s) of actions available and deontological norms (Hunt & Vitell, 1986), which are relative to one’s personal values or perceived behavioral norms (Hunt & Vitell, 2006). For example, if a consumer receives too much change after a service encounter, they may adopt a “finders’ keepers” approach based on their personal value system. Conversely, a different consumer might feel constrained by their personal values or the norms of their surroundings, which emphasizes the value of honesty from which they would form a different evaluation. Both these distinct deontological evaluations are likely to influence consumers’ ethical judgments and behavioral intentions.

The effect of deontological evaluations on ethical judgments is well established in previous empirical work (Hunt & Vásquez-Párraga, 1993; Mayo & Marks, 1990). According to the H-V theory of marketing ethics, deontological evaluations are proposed to directly impact only ethical judgments, which are posited to subsequently affect ethical intentions (Hunt & Vitell, 1986, 2006). However, several studies suggest that there is a significant direct relationship between deontological evaluations and ethical intentions (e.g., Chan et al., 2008; DeConinck & Lewis, 1997). We propose that the effect is warranted and worth investigating. We thus propose the following hypothesis:

H1

Deontological evaluations will have a positive effect on (a) ethical judgments and (b) ethical intentions.

Teleological Evaluations and Ethical Judgments/Intentions

Teleological evaluations are determined by three psychological elements that in combination influence how good/bad one perceives the consequences of an action (Cole et al., 2000). Firstly, the probability of consequences poses important considerations. For example, if one is presented with an ethical issue and perceives that harm is more/less likely to happen, the worse/better their teleological evaluation will be, and the overall ethical judgment and intention to act ethically will be higher/lower. Secondly, the desirability of consequences is also considered, such that one will judge an action to be more ethical and intend to behave ethically if the consequences are more favorable for them doing so. Thirdly, they would consider the importance of stakeholders or namely those affected by the ethical issue. One’s judgment and intention might alter based on whether the individual cares for those who are negatively affected by the action. For example, if a consumer is confronted with the issue of receiving too much change, they might assess a negative consequence, such as getting caught after not owning up, relative to receiving a surplus of change. The consumer may also assess the possible negative consequences to the store or cashier in case they do not return the excess change. The teleological evaluation of the consumer would be determined by (1) the probability of these outcomes occurring, (2) the desirability of these outcomes occurring, and (3) how important the cashier/store is to the consumer and how negative the consequence of getting caught would be to the consumer. This assessment would likely affect the consumer’s judgment of which is the most ethical alternative (leaving without owning up to receiving excess change or returning the surplus change) and their behavioral intention. Empirical testing provides considerable support for the role of teleological evaluations in relation to both ethical judgments and intentions (Chan et al., 2008; Mayo & Marks, 1990). Thus, we propose the following hypothesis:

H2

Teleological evaluations will have a positive effect on (a) ethical judgments and (b) ethical intentions.

Factors Moderating the Effects of Deontological and Teleological Evaluations on Ethical Judgments/Intentions

Table 1 provides the definitions and some examples of the variables. We select theoretical moderators relative to the contextual elements of the ethical issue and stakeholders.

Table 1 Definitions and operationalizations of dependent variable, moderator variables and control variables for meta-analysis

The contextual elements of the ethical issue that we consider important include the specific ethical context (offline consumer ethics, online consumer ethics, or environmentally sustainable consumer ethics), the presence/absence of financial implications in the ethical issue, and cultural values. Furthermore, we take into consideration that theoretical models, which include deontological and teleological evaluations as key variables, originate in organizational ethics and have been developed and applied to consumer ethics (Hassan et al., 2021; Vitell & Muncy, 2005). The majority of business ethics research has also focused on organizational contexts, often neglecting consumer ethics (Casali & Perano, 2021). Organizational ethics present a distinct ethical environment compared to those found in consumer contexts. Organizations generally prescribe codes of ethics that dictate situational deontological norms and include rewards and sanctions for compliance and violations of such codes. Hence, as noted earlier, the role of deontological and teleological evaluations in organizational contexts provide a useful reference point to compare these effects with those in consumer contexts. We incorporate original manuscripts in our study that offer effect sizes from organizational contexts to primarily serve as a reference model for our empirical findings in consumer contexts.

However, relative to specific consumer contexts, Vitell et al. (2001) find different patterns in the effect of moral evaluations on ethical judgments in consumer ethical issues that happen offline, mostly in retail contexts, such as switching price tags, compared to those that occur online, such as copying software. Given that our focus in the meta-analysis is on consumer ethics, these potential differences between offline and online consumer ethics are deemed a worthy consideration and relevant for our investigation.

Further, an additional consideration pertains to the fact that business exchanges frequently require monetary exchanges, which can be the source of many consumer and organizational ethical issues. For example, ethical issues such as consumers lying for discounts, not paying for software and employees padding expense accounts all have financial implications. Previous studies have identified relevant trends in individual decision-making in the presence/absence of financial implications (Vohs, 2015). Thus, it is deemed a relevant variable to include in our meta-analytical framework.

In addition, previous studies demonstrate that cultural differences moderate the ethical decision-making process through the lens of various cultural frameworks (e.g., Mitchell et al., 2009; Vitell et al., 1993). Studies that examine the effect of deontological and teleological evaluations take place across various national cultures over a span of three decades. Hence, cultural values are an important variable for us to consider. We utilize Inglehart and Baker’s (2000) traditional versus secular-rational values framework as it provides a relevant conceptual explanation for the effects of deontological and teleological evaluations on ethical responses and accounts for temporal shifts in cultural values.

An additional likely source of moderation is the role of stakeholders. According to stakeholder theory, there are various important people and entities that are affected by business activities that need to be accounted for (Freeman et al., 2018; Phillips et al., 2003). Thus, when ethical issues arise, stakeholders are a key consideration, which is also emulated in many ethical decision-making models (e.g., Ferrell and Gresham, 1985; Hunt & Vitell, 1986; Jones, 1991). In our analysis, stakeholders are categorized in accordance with common conceptualizations of ethicality. Two important stakeholders—the agent and the patient—can be identified when an ethical issue arises, which corresponds to a common conceptualization of morality, the moral dyad (Gray et al., 2012). The moral dyad refers to individuals’ tendency to conceive ethical issues in a bidirectional dyadic fashion where one person/entity is harmed and falls victim while the other benefits (Gray et al., 2012). For example, in the case of not engaging in recycling, the environment is the victim, or the entity that is harmed as a result of the (in)action, whereas the individual or corporation is the beneficiary as they save time, resources or any effort associated with recycling. Thus, in our meta-analysis, we categorize stakeholders as the perceived victim and beneficiary of the ethical issue as a likely source of moderation of the effect of deontological and teleological evaluations on ethical judgments and intentions.

Ferrell and Gresham (1985) posit that individuals do not learn (un)ethical behavior from society but rather from the norms of and interpersonal exchanges with disparate groups. Relative to business exchanges, there is a clear distinction between the norms of exchanges with personal versus corporate entities. Therefore, in our meta-analysis, victims and beneficiaries are categorized based on the relational level they have with scenario protagonists relative to whether it is corporate or personal.

However, a recent surge in environmentally sustainable consumer ethics research (Nova-Reyes et al., 2020) means that various studies are unable to meet the above-mentioned relational criteria as the victim is neither corporate nor personal. Environmentally sustainable consumer ethics is an important subset of offline consumer ethics, and a key commonality of these ethical issues is that the natural/ecological environment is the victim. Thus, we introduce environmentally sustainable consumer ethics as an additional ethical context that is operationalized as the environment being the victim.

In addition to the theoretically relevant moderators related to the contextual elements and stakeholders, we also consider two other exploratory moderators (see Fig. 2A and B). Consistent with the moral dyad, the scenario actor (i.e., the agent) offers a source of potential moderation. In order to avoid potential social desirability bias, vignettes are often adopted in ethics research and include third-person protagonists (Wason et al., 2002). Other studies differ in their research design and use first-person protagonists to heighten ecological validity (e.g., Yoon, 2012). Previous studies have suggested that third-person perspective taking often increases ethicality (e.g., Eyal et al., 2008; Žeželj & Jokić, 2014), yet it is not clear how it may moderate the effect of deontological and teleological evaluations on ethical responses. Thus, we consider this to be an important exploratory variable and include it in our meta-analytical framework.

Furthermore, the importance of stakeholders poses additional considerations, and researchers and scholars frequently distinguish between, and exclusively measure, two different branches of teleology: egoism and utilitarianism. Egoism is concerned with maximizing one’s self-interest, such that an egoist perspective of teleological evaluations focuses on the perceived goodness/badness of consequences relative to the decision-maker’s interests. In the above example of receiving excess change, a potential gain for the consumer would be through not returning this financial surplus. A potential loss for the consumer would be the possibility of getting caught and being penalized. Conversely, utilitarianism centers on maximizing good for the greatest number of people. Thus, a utilitarian perspective of teleological evaluations would assess the magnitude and valence of consequences relative to greater good, and not solely oneself. In the above example of receiving excess change, a utilitarian perspective would include the potential loss to the store and any negative consequences to the cashier when making evaluations of the acceptability of the act of not returning the excess change.

Another example can be provided in the context of digital piracy. In one’s teleological evaluation, an egoist perspective would assess the consequential outcomes relative to their own interests, such as saving money versus the reduced quality of the product. Alternatively, one could take a utilitarian perspective and give weight to the destructive impact of their actions on the entertainment industry and related stakeholders. We argue that this consistent distinction, combined with the clear conceptual differences between egoism and utilitarianism, poses important considerations and is an interesting exploratory distinction worthy of incorporating into our meta-analytical framework.

The following section offers a discussion of the hypothesized effects that our theoretical moderators (ethical contexts: offline consumer ethics, online consumer ethics, environmentally sustainable ethics; financial implications; cultural values; stakeholders) are expected to have on the effects of deontological and teleological evaluations on ethical judgments and intentions.

Offline Consumer Ethics

Offline consumer ethics refer to the ethical issues like those arising in physical retail stores and servicescapes, such as receiving too much change (Bateman & Valentine, 2010) and switching price tags (Vitell et al., 2001). In many offline consumption environments, consumers are less bound to a code of ethics and the potential consequences of ethical actions have differing levels of clarity as compared to an organizational context. Individuals in organizations (see Table 1, e.g.) are generally bound to an ethical code which dictates deontological norms. Further, these ethical codes often prescribe rewards and sanctions, which communicate respective positive and negative consequential information associated with ethical conduct. Conversely, in a retail environment, which is commonly featured in offline consumer ethics research, and is susceptible to great losses due to unethical consumer behavior, customer satisfaction is a higher priority (Gomez et al., 2004). Consequently, these codes of ethics are not pursued in the same vein as what is typically seen in organizational contexts to not to be seen as alienating for consumers (Fullerton & Punj, 2004). However, previous studies have consistently uncovered the positive effect that these codes of ethics and rewards/sanctions systems have on ethical decision-making (Craft, 2013). Thus, due to the established effect of these codes and systems that determine ethical climates and the differences between how they are applied in organizational and offline consumer contexts, we propose the following hypothesis:

H3

Both the effect of (a) deontological evaluations and (b) teleological evaluations on ethical judgments and ethical intentions will be weaker in offline consumption contexts relative to organizational contexts.

Online Consumer Ethics

Online consumer ethics relate to ethical issues that consumers encounter during online consumption activities. Examples of online consumer ethical issues include piracy (Yoon, 2012) and consumer cyber aggression (Han and Vasquez, 2020). Research on online consumer ethics is a particularly relevant subset of ethics research due to the increase in global Internet penetration rates (International Telecommunication Union, 2020) and the unique, dynamic unethical behaviors that are present in online interactions (Saban et al., 2002).

Johnson (1997) outlines three unique features of the Internet that pose different moral considerations. The internet’s extended scope, offered anonymity, and possibility of reproducibility allow for more immediacy, a broader reach, faceless crimes, endurance of information, and possibility of permanence unlike offline behaviors (Freestone & Mitchell, 2004; Johnson, 1997). Chatzidakis and Mitussis (2007) propose that all three dimensions heighten unethical piracy intentions due to decreased negative consequences that are typically associated with such behaviors. Vitell and Muncy (2005) also categorize online unethical consumer actions like software piracy as “no harm, no foul” actions, implying that many consumers do not find these actions to be unethical as they are not viewed as having negative consequences. The lack or absence of perceived consequences will likely reduce the weight of teleological evaluations in an online consumption context compared to an organizational context where there are policies and procedures that outline what happens if there is a violation. Additionally, the uniqueness and dynamic nature of online behaviors, such as digital piracy, implies a lack of deontological norms, as there are less situational constraints and less familiarity with the moral stringency of the act. Thereby, we argue that online consumer ethics have distinct considerations for the role of deontological and teleological evaluations in relation to ethical judgments and intentions. Thus, we propose the following hypothesis:

H4

Both the effect of (a) deontological evaluations and (b) teleological evaluations on ethical judgments and ethical intentions will be weaker in online consumption contexts relative to organizational contexts.

Environmentally Sustainable Consumer Ethics

Environmentally sustainable consumer ethics pertain to ethical concerns regarding the relationship between consumers and the natural ecological environment. Examples include recycling, sustainable transport, or pro-environmental shopping practices, etc. (Chan et al., 2008; Yin et al., 2018). These ethical issues appeal to the greater good of humankind (and other species), whereas ethical issues in organizational and consumer contexts are often relative to a self-serving organization as opposed to all humanity. Therefore, environmentally sustainable consumer ethics relate directly to consequentialism and thus, teleological evaluations. Consequentialist ethics are flexible, as noted by Palmer et al., (2014, p.431) “environmental ethicists have defended different forms of consequentialism, in particular, by proposing varying views of good outcomes.” Regarding the relationship between teleological evaluations and ethical judgment, Culiberg and Bajde (2013) find that the consequence-focused dimensions of Jones’s (1991) moral intensity theory serve as significant predictors of ethical judgment. In a specific environmentally sustainable consumer ethics context (i.e., bringing your own bag when shopping), Chan et al. (2008) find that teleological evaluations are a significant predictor of ethical judgments and intentions. Hence, we argue that an environmentally sustainable consumer ethical context will lead to significantly positive moderation of the relationship between teleological evaluations and ethical judgments and intentions. Thus, we propose the following hypothesis:

H5

The effect of teleological evaluations on ethical judgments and ethical intentions will be stronger when the environment is the victim relative to an organization.

Financial Implications

Financial gain or loss is a pivotal element in many ethical issues. For example, various ethical issues such a piracy and gray market procurement arise for individuals to undercut industry prices (e.g., Wan et al., 2009; Zhuang et al., 2014). A universal moral value is respecting property rights by honoring another’s possessions (Curry et al., 2019). Specifically, honoring one’s territory and possessions is a universal teaching that is integral to conceptualizations of deontology (Sandberg, 2013). For example, various religious teachings and judiciaries condemn stealing and coveting another’s possessions. Although, studies suggest that money has a negative effect on ethical responses (e.g., Gino & Mogilner, 2014; Kouchaki et al., 2013; Vohs, 2015), they also reveal moderating conditions that directly relate to deontological fundamentals including property rights. These conditions include the legality of how it was obtained (Yang et al., 2013) and who it belongs to (Polman et al., 2018). Thus, the deontological norm of respecting property rights is likely prompted when for consumers there are financial implications.

Further, given the nature of some of the industries included for organizational effect sizes in our meta-analysis, such as: accountancy (Burns & Kiecker, 1995; Shapeero et al., 2003), market research (Mayo & Marks, 1990), and sales (Cherry & Fraedrich, 2000; Lund, 2001), it is likely that individuals that represent an organization act as surrogate shoppers who are empowered to act with others’ money. Systems that designate the role of surrogate shoppers rely heavily on reputation/trust and are susceptible to abuse; thus, in such scenarios individuals are bound to a structure that hosts a code of ethics and strong deontological norms (Hollander & Rassuli, 1999). We suggest that individuals’ deontological evaluations would be more influential in forming ethical responses when there are financial implications. Thus, we propose the following hypothesis:

H6

The effect of deontological evaluations on ethical judgments and intentions will be stronger when there are financial implications.

Cultural Values

Empirical testing of the effect of deontological and teleological evaluations has been conducted across various national cultures, but most consistently the United States. Inglehart and Baker (2000) propose a cultural framework related to ethical decision-making. They state that nations differ on cultural values, which correlates with economic development. This suggests that traditional values, which are more prominent in less economically developed countries, are highly correlated with, thus reflected through, idealist responses to moral absolutes, which are rigid ethical beliefs regarding the universal rightness/wrongness of an action (Forsyth et al., 2008). Conversely, individuals from more secular-rational societies are more flexible and less punitive in response to such moral absolutes (Inglehart, 2006). For example, in traditional societies, individuals are more likely to attend religious institutions and shun those who do not abide by their teachings, whereas the opposite is typically the norm in secular-rational societies (Inglehart, 2006). Thus, individuals with stronger traditional values are more likely to be influenced by rigid beliefs when forming ethical judgments and intentions, which correspond with deontological evaluations (Mitchell & Chan, 2002). Alternatively, those with higher secular-rational values are less likely to have inflexible moral beliefs and are more likely to consider the relative consequences of actions when forming ethical judgments and intentions. This distinction in the moral foundations of ethical decision-making between traditional values-based societies and secular-rational values-based societies (i.e., western, educated, industrialized, rich, and democratic societies or WEIRD) was proposed by Haidt (2012). Thus, we propose the following hypotheses:

H7a

The effect of deontological evaluations on ethical judgments and intentions will be stronger when traditional values are higher.

H7b

The effect of teleological evaluations on ethical judgments and intentions will be stronger when secular-rational values are higher.

Stakeholders

The importance of stakeholders is often related to teleological evaluations but is not posited to be related to deontological evaluations (Hunt & Vitell, 1986, 2006). However, based on research in stakeholder theory (e.g., Gibson, 2000; Mansell, 2013), we argue that this is a significant oversight. Stakeholders play an important role in the effects of deontological evaluations, particularly when they are victims of unethical actions. Gibson (2000) and Mansell (2013) argue that deontological claims offer strong moral bases for stakeholder theory. Businesses often stress moral obligations that internal actors have relative to external stakeholders (Gibson, 2000). For example, moral duties that employees have to consumers to provide effective and safe products/services form a prevalent part of many organizational ethical environments. Thus, stakeholders play a crucial part in forming deontological norms in organizations. Deontological norms are more likely to be enforced in exchanges with professional entities compared to individual consumers. We anticipate that this will strengthen the impact of deontological evaluations when the victim or beneficiary of an unethical act is an organization related entity rather than a personal acquaintance. In other words, the relationship between deontological evaluations and ethical judgments/intentions will be weaker when the victim or beneficiary is a personal relation as opposed to an organizational relation. Thus, we propose the following hypothesis:

H8

The effect of deontological evaluations on ethical judgments and ethical intentions will be weaker when individuals have a personal relationship with (a) the victim and (b) the beneficiary relative to an organizational relationship.

Additionally, stakeholders pose important considerations for the formation of teleological evaluations and their effects on ethical responses. Since stakeholders are those affected by, or those who bear the consequences of the (un)ethical act, they are important from the perspective of both organizational and personal relations. It is likely there are distinct effects for affective responses (such as empathy) to stakeholder outcomes depending on whether there is an organizational or personal relationship with the affected stakeholder. Empathy is triggered when the victim of an unethical act is identifiable (Bloom, 2013) and is defined as “feeling what others feel” (Jordan et al., 2016, p. 1107). Findings suggest that empathy is subject to in-group bias, and higher empathetic responses are exhibited when an in-group member falls victim to an unfavorable consequence (Bloom, 2017). We posit that one’s personal relationship to the victim is consistent with this view (i.e., being a member of a person’s in-group). Empathy has been consistently correlated with individuals’ ethical responses (e.g., Chowdhury & Fernando, 2014; Mencl & May, 2009). Thus, we expect that when the victim is a personal relation, a higher empathetic response will be triggered, which will strengthen the effect of the teleological evaluation. Thus, we propose the following hypothesis:

H9

The effect of teleological evaluations on ethical judgments and ethical intentions will be stronger when individuals have a personal relationship with the victim relative to an organizational relationship.

Method

Data Collection

To conduct the meta-analysis, we conducted extensive literature searches of published and unpublished studies in March 2021 and February 2022. Data retrieval methods were consistent with previous meta-analyses published in consumer ethics research (see Eisend, 2019; Pan & Sparks, 2012). First, we searched electronic databases [i.e., Scopus, Google Scholar, ScienceDirect, SSRN, ResearchGate, and OATD (Open Access Theses and Dissertations)]—with main keywords such as “teleological,” “deontological,” “ethical,” and “moral.” Later searches also included words that are synonymous with deontology and teleology in the ethics literature such as: “egoism,” “utilitarianism,” consequentialism,” “formalism,” and “contractualism.” Second, using the same keywords, we manually searched through archives of academic journals, including: Journal of Marketing Research, Journal of the Academy of Marketing Science, Journal of Business Ethics, Journal of Business Research, Journal of Macromarketing, etc. Third, conceptual articles and review papers were identified by relevance and citation count (e.g., Hunt & Vitell, 1986; Pan & Sparks, 2012; Reidenbach et al., 1991). A backward and forward reference search of these papers was conducted to screen for potential studies. Fourth, references for all obtained manuscripts were examined to identify other articles for potential inclusion in our meta-analysis.

In Fig. 3, we outline the PRISMA search flow (see Appendix 1 for full reference list). A study was eligible for inclusion in our meta-analysis if it met the following criteria: (1) the independent variables (deontological evaluations and evaluations) are mutually exclusive constructs; (2) the dependent variable captures a measure of the participants’ ethical judgments and/or ethical intentions; (3) the scenario took place in a business context (either organizational or consumer); and (4) the study was empirical in nature. In total, we screened 250 papers, 124 of which were excluded due to not fulfilling our criteria, leaving 74 eligible for inclusion. Although we contacted first authors whose papers did not contain the correct statistical information, a further 21 papers were still excluded due to lack of usable data. Thus, we coded the 53 eligible papers published between 1990 and 2021 that were available during February 2022 from which we obtained 316 effect sizes. Thirty-three of the papers were related to organizational ethics and 20 were related to consumer ethics, which allowed for a relatively even split of organizational (k = 135) and consumer (k = 181) effect sizes in the final dataset.

Fig. 3
figure 3

Adapted from Moher et al. (2009). 1. 60 articles had a dependent variable which did not meet the criteria, 64 had the incorrect operationalization of the independent variable. 2. 21 articles were excluded due to lack of statistical information

PRISMA search flow.

Effect Size Computation

We chose Pearson’s r as our effect size metric. The size of r provides an estimate of the strength of association between two variables and its direction (Eisend, 2019). If r was not available, t-values and Cohen’s d were obtained and converted following common meta-analysis guidelines (see Appendix 2 for effect size calculations).Footnote 1 Various studies employed 2 × 2 experimental designs (e.g., Burns & Kiecker, 1995; Vitell et al., 2001). In such cases, we obtained mean differences and standard deviations, and effect sizes were re-calculated when the study presented no alternative methods of interpreting the data. Once we obtained all correlations, the coefficients were disattenuated. Disattenuation considers the internal reliability of measurement items and accounts for the weakened effect of effect sizes due to measurement errors, thus it is necessary to obtain the true effect size (Zimmerman & Williams, 1997). It is not uncommon for disattenuated coefficients to result in r > 1.00; effect sizes outside of −1.00 < r < 1.00 were determined as outliers and omitted from the analysis. We converted r into Fisher’s z-scores to avoid potential biases that could arise from the non-normal distribution of r. Fisher’s z-scores were then transformed back to r for predictive values, which report mean effect sizes.

Moderator Coding

We developed a coding system to identify any moderators (see Table 1). These moderators relate to: (1) the contextual elements of the ethical issue, (2) the stakeholders, and (3) the methodological characteristics of primary studies. Two exploratory moderators were also included.

First, for the contextual elements of the ethical issue, we considered the ethical context, which delineated whether the ethical scenario was in an organizational context, offline consumer context, online consumer context, or environmentally sustainable consumer context. Second, we coded the financial implications of the scenario for whether monetary exchanges or consequences were included in the scenario. Third, we included cultural indexes to account for sample differences that could be related to national culture. We obtained a continuous measure of each relevant nation’s traditional versus secular-rational values from the World Values Survey (Inglehart et al., 2014). The data were obtained from the most recent data collection relative to the year of presumed data collection (2 years before the year of publication) and mean centered to provide an accurate metric. We assumed that the metrics from when the study’s data were collected was a more accurate representation of a nation’s economy, which like cultural values, fluctuates. Thus, the data do not reflect a fixed cultural trait, but rather a dynamic reflection on how cultural values evolve and fluctuate with national economies. Fourth, for stakeholders, we distinguished between the victim and beneficiary, which was operationalized at the relational level with the study participant (i.e., organizational, personal, or the environment).

As mentioned above, we also investigated two exploratory moderators. The first was related to whether the teleological measurement was egoistic or utilitarian. The second was whether the actor (the scenario protagonist) was the self (first-person perspective) or other (third-person perspective). In addition, various variables that relate to methodological characteristics of primary studies were identified. These related to both study-level characteristics and publication bias controls. We controlled for study-level characteristics that related to two contrasting approaches in methodologies that reflected the measurement of deontological and teleological evaluations. First, we accounted for whether researchers directly manipulated or measured the independent variable. Second, we coded for whether researchers measured for, or considered, potential social desirability bias in their research design. Third, we coded to control for the outcome measurement or type of ethical response. Thus, we noted whether the ethical response was a judgment or an intention. We introduced publication bias controls to account for potential selective publication of studies (Ferguson & Brannick, 2012), which could be due to sample sizes, contradictory results, or null effects (Duval & Tweedie, 2000). Thus, we included the status of publication (unpublished vs. published), year of publication [to ensure that there was no inflation of novel effects in earlier papers (Camerer et al., 2018)], and a precision measure. The precision measure was computed as the inverse of the standard error to account for the potentially negative relation between the effect size and the study’s sample size (Stanley & Doucouliagos, 2012). All the categorical variables were dummy-coded and the continuous variables were mean centered. To increase the confidence in our coding, a sample of effect sizes were double coded by an external researcher who was not part of the research team. The intercoder reliability was r = 0.72, and any disagreements were settled through discussion.

Meta-Analytic Procedures

We employed a three-level model in our meta-analytic procedures. A multi-level approach would account for the dependencies and variation within the data that could be assumed from a multitude of effect sizes (Van den Noortgate & Onghena, 2008). Each level was incorporated to account for biases that can occur due to effect sizes being derived from the same paper, the same study, and the same sample. We used a three-level, random-effects, maximum-likelihood model with the “metafor” package in R Studio to generate effects (Viechtbauer, 2010). We ran influential case diagnostics to identify any potential outliers that could distort conclusions taken from the data (Viechtbauer & Cheung, 2010). We found several effect sizes had high standardized residuals (> 2.57; Belli et al., 2021), so they were omitted from further analysis. Among remaining effect sizes, we correlated the variables to detect any potential multicollinearity (see Appendix 3 for correlation matrices). Multicollinearity was not detected, as all significant correlations were −0.70 ≤ r ≤ 0.70.

Due to effect size computational procedures, we next separated effect sizes relevant to each independent variable. We favored random-effects models based on the assumption that true effect sizes vary among participants and treatments (Borenstein et al., 2009). The nature of the data is consistent with this assumption, and a random-effects model also allows for "an explicit generalization beyond the observed studies" (Hedges & Vevea, 1998, p.487). We also analyzed H1-2 to provide an estimate of the Q-stat and its significance to assess the level of heterogeneity that would warrant running additional meta-regression models. We ran two multi-level, random-effects meta-regression models, which incorporated two-tailed significance testing, with the consideration that several papers include single-subject experimental designs (e.g., Burns & Kiecker, 1995; Vitell et al., 2001).

Results

Publication Bias

We used three indicators to assess the susceptibility of our findings to publication bias. First, we performed both Rosenthal (1979) and Rosenberg’s (2005) fail-safe N, which, respectively, reveal how many null effect sizes would be needed to take results below a p < 0.05 significance level and the number of studies needed to support the null hypotheses (Carrillat et al., 2018). Table 2 shows that our results passed both tests, thus exceeding the recommended N > 5 k + 10 (Zlatevska et al., 2014). Second, in Fig. 4, we provide funnel plots. Superficially, the funnel plots reveal multiple effect sizes outside of the funnel, thus we proceeded to apply a trim-and-fill approach (Duval & Tweedie, 2000). Trim and fill tests revealed asymmetry for both the deontological and teleological evaluations analyses (21 and 29 missing studies on the right side, respectively). Third, we conducted an Egger’s regression (Sterne & Egger, 2005) in which a significant z-value suggests the possibility of publication bias. The results reveal a significant score for deontological evaluations (z-value = −5.00, p < 0.01), but not for teleological evaluations (z-value = −0.25, p = 0.81). Various indicators suggest that models for teleological evaluations may be susceptible to publication bias.

Table 2 Summary values/indicators for the two models
Fig. 4
figure 4

Funnel plots

Observed Heterogeneity

For overall summary effect sizes, we used three metrics to assess heterogeneity; I2 for between-effect size variance, τ2 for sample estimates of between-effect size variance, and Q-statistics for effect size dispersion (Carrillat et al., 2018). Tables 2 and 3 provide the results for the heterogeneity indicators. Significant Q-statistics, τ2 > 0, and I2 > 50% highlight the between-study variance in the summary effect models, which warranted the running of conditional models. Heterogeneity indicators for meta-regression models also included R2 as an additional metric for variance explained by the model.

Table 3 Meta-regression models

Main Effects of Deontological Evaluations and Teleological Evaluations

Multi-level random-effect analyses reveal significant summary effects of both deontological evaluations (β = 0.69, p < 0.01) and teleological evaluations (β = 0.38, p < 0.01) on ethical judgments and intentions. Predicted values reveal that deontological evaluations had a stronger effect on ethical judgments (grand mean \(\overline{r}\) = 0.77) than intentions (grand mean \(\overline{r}\) = 0.51). This pattern was also observed for teleological evaluations of judgments (grand mean \(\overline{r}\) = 0.45) and intentions (grand mean \(\overline{r}\) = 0.23). Our analyses of conditional models further strengthened this claim. We found that ethical intentions in the meta-regression models relative to the effect on judgments showed significant negative moderation for both deontological (β = −0.25, p < 0.001) and teleological evaluations’ effect (β = −0.17, p < 0.001). Such findings provide support for H1 and H2 and provide validation for the wider application of the effect of deontological and teleological evaluations on ethical judgments and intentions across various domains.

We found a significant difference between the summary effect of deontological evaluations (r = 0.60, s.e. = 0.05) and teleological evaluations (r = 0.37, s.e. = 0.04) through a Wald-type test (z-value = 3.60, p < 0.05). Thus, our findings suggest that deontological evaluations have a stronger overall effect on ethical responses.

Moderator Analysis

Table 3 depicts the findings from the conditional models that include our moderators and control variables. The moderators account for substantial variance across both deontological and teleological models, which can be identified through moderate residual heterogeneity (R2 = 53.20% and 33.78%, respectively), as well as significant Q-statistics (4148.43, p > 0.001 and 2854.82, p > 0.001) and I2 > 50%.

Deontological Evaluations

We next assessed the moderation effects of the relation between deontological evaluations and ethical responses. The results are in Table 3. We find that ethical context is a strong source of moderation, consistent with H3a but not with H4a. The results reveal that in offline consumer contexts, the effect of deontological evaluations is weaker on ethical responses in comparison to organizational contexts (β = −0.16, p = 0.03). The results also reveal that in online consumer contexts, the effect of deontological evaluations is not significantly weaker on ethical responses in comparison to organizational contexts (β = −0.11, p = 0.08). Thus, H3a is supported, but H4a is not. The presence of financial implications in the ethical scenario positively moderated the effect of deontological evaluations, such that they had a stronger effect (β = 0.30, p < 0.01). Thus, H6 is supported. Consistent with our predictions, there was significant negative moderation when the victim (β = −0.53, p < 0.001) and the beneficiary (β = −0.26, p < 0.01) of the scenario was a personal relation in comparison to an organizational relation. Thus, H8 is fully supported. Our findings also highlight that traditional values exert no significant effect (p = 0.24). Hence, H7a was not supported.

Additionally, moderation from the actor was not significant (β = −0.27, p = 0.08), but various methodological characteristics of primary studies that were incorporated in the model as controls reached a p < 0.05 significance level. The findings indicate that when the independent variable was manipulated and not measured, the effect of deontological evaluations on ethical judgments and intentions was stronger (β = 0.37, p < 0.01). We also observed this pattern when researchers accounted for social desirability bias in their research design (β = 0.38, p = 0.02). The other identified methodological characteristics were not statistically significant.

Teleological Evaluations

Our analysis of the moderation for the effect of teleological evaluations on ethical responses revealed fewer significant effects. No significant moderation was found for ethical context, such that no difference was uncovered between offline consumer ethical issues (p = 0.77) and organizational ethics. Similarly online consumer ethical issues (p = 0.88) did not show any difference compared to organizational ethics. This was further demonstrated by the similarly weighted predicted values across organizational (\(\overline{r}\) = 0.35), offline consumer (\(\overline{r}\) = 0.37), and online consumer ethical issues (\(\overline{r}\) = 0.36). However, we found significant moderation in the case of environmentally sustainable consumer ethics (β = 0.59, p = 0.02). Therefore, neither H3b nor H4b was supported while H5 was supported.

Victim type exerted significant moderation, and teleological evaluations were stronger when the protagonist had a personal relationship with the victim (β = 0.08, p = 0.03). Thus, H9 is supported. Conversely, we also find that when the protagonist had a personal relationship with the beneficiary, there was a significant negative effect in comparison to an organizational beneficiary (β = −0.14, p = 0.03). Further, there was no significant moderation for financial implications (p = 0.32) nor cultural values (p = 0.18). Thus, H7b was not supported.

Finally, our exploratory moderators reveal interesting findings, such that actor did not have a significant effect (p = 0.27), but utilitarianism had a significant effect compared to egoism (β = 0.11, p < 0.001), and the relation between teleological evaluations on ethical responses was stronger when the teleological evaluation measured or manipulated consequences related to the greater good.

The findings from the meta-regression also suggest publication bias for the effect of teleological evaluations on ethical responses, which is consistent with the previous publication bias diagnostics. Specifically, the effect of teleological evaluations was stronger in published paper than from unpublished sources (β = 0.39, p = 0.03) and there was a significant negative moderation of the effect size precision (β = −0.05, p < 0.01). Table 4 provides a summary of the results in relation to the hypotheses that we examined.

Table 4 Summary of hypotheses testing

Discussion

Main Findings and Contributions

We conducted an empirical assessment to validate the effects of deontological and teleological evaluations on ethical judgments and intentions. Our meta-analytical findings provide two key contributions. First, we provide a summary effect that clarifies the impact of two well-established types of moral evaluation in the consumer ethics literature. We add to the ethical decision-making models (Hunt & Vitell, 1986, 2006) by highlighting the strength and weighting of such evaluations. Second, we posit a range of contingency factors that moderate such effects. Our findings pinpoint distinct patterns of effects for deontological and teleological evaluations relative to ethical contexts (offline consumer ethics, online consumer ethics, environmentally sustainable consumer ethics, organizational ethics), the presence of financial implications, and different types of stakeholders (based on their relationships with the decision-maker). All these factors suggest important implications for researchers and practitioners alike.

The Role of Contextual Elements of the Ethical Issue

We identify distinct patterns of the effects of deontological evaluations between organizational and offline consumer contexts. The ethical environment, which in part predetermines deontological norms (Singhapakdi & Vitell, 1991), is likely to be less enforced in offline consumer contexts. Although our study includes data from both organizational and consumer contexts, the ethical environments have been explored less in the context of consumers than in organizations (Craft, 2013). We propose that the difference in ethical environments could be responsible for the weaker effect of deontological evaluations in offline consumption. Conversely, the findings reveal no contextual differences regarding teleological evaluations, apart from a stronger effect when the issue relates to the environment. However, the predicted grand means were still significant, which suggests that teleological evaluations have a general application across the organizational and consumer contexts we explored.

Further, financial implications resulted in stronger deontological evaluations. For consumers, these material consequences likely prompt property rights concerns, which are central to established deontological teachings and judiciaries. On the other hand, for individuals in organizations, this can be due to ethical climates that foster ethical decision-making when money is involved.

The Role of Stakeholders

Deontological evaluations had a weaker effect in personal relationships (i.e., greater social proximity) involving anticipated victims, yet teleological evaluations had a stronger effect. Previous studies that explore the effect of proximity on ethical responses report mixed results, such that some find that increased social proximity results in stronger ethical judgments (e.g., Eyal et al., 2008; Lo et al., 2019), while others find the opposite effect (e.g., Choi et al., 2017). However, our results suggest that this could be due to the differing effect that proximity with the victims may have on the moral evaluations that form ethical responses: stronger (weaker) effects for teleological (deontological) evaluations.

Practical Implications

The effect of deontological evaluations has practical managerial relevance in consumer contexts, as they correspond to preventative measures, namely educational approaches. Educational approaches stress the moral constraints of an act or provide contextual cues on normative ethical behavior expectations (Fullerton & Punj, 2004). For example, the use of visual communications in a servicescape that stress the inherent rightness/wrongness associated with an action appeal directly to a deontological evaluation. Our findings complement extant literature that stresses the role of an ethical environment (e.g., Craft, 2013) by suggesting that in consumer settings, educational approaches that stress deontological norms may help foster stronger ethical climates. This also has important public policy implications in terms of developing and supporting normative standards for online and offline consumer behavior.

On the other hand, teleological evaluations are more likely to relate to deterrence strategies (Dootson et al., 2014). Some previous studies examine the impact and effectiveness of various deterrence methods (Dawson, 1993; Saine et al., 2021). Dootson et al. (2014) show that varying the size of a corporate victim did not influence the perceived acceptability of unethical consumer behavior. Our results advance this finding, as we show that relational levels with the victim have varying effects on both types of moral evaluations that influence ethical responses. Teleological (deontological) evaluations’ effect on ethical responses is stronger (weaker) when there is a personal relationship with the victim. Thus, efforts to accentuate the victim on a relational level in a deterrence strategy could be more effective. For example, when communicating consequential information about fines or procedures following an episode of unethical behavior, management could relate the personal relationship the victim of the unethical behavior has with the perpetuator.

An additional consideration pertains to the stronger effect of teleological evaluations from a utilitarian perspective as opposed to an egoist one. Teleological evaluations are stronger when individuals consider the consequences relative to the greater good rather than to their own self-interest. Thus, managers can signal collective sanctions or the harm/benefit of (un)ethical behavior to the greater good to encourage consumers to behave ethically. For example, to stop people smoking in hospitality venues, visual communications that portray the negative consequences to others (e.g., showing the dangers of second-hand smoke) could be utilized.

In a similar vein, our findings have relevance for sustainability practices and ecological organizations. Given that consumers rely substantially on teleological evaluations in forming ethical responses in environmentally sustainable consumer contexts, marketing efforts can be employed as a means of encouraging environmentally sustainable behavior. Such efforts could focus on deterrence strategies, such as imposing sanctions for behaving unethically in relation to the environment or by presenting beneficial consequences of doing well by it.

Limitations and Future Research Opportunities

Our study adopts the common assumption in a meta-analysis relating to research integrity and relevant disclosure of the original authors (Van Laer et al., 2019). We focused on accounting for methodological differences that could skew/impact the results as control variables.

In addition, due to insufficient statistical information, we were unable to obtain the potential effect sizes of all previous empirical studies. An approximate 30% exclusion is not uncommon in marketing and consumer behavior meta-analyses (Eisend, 2009), to which our 28% exclusion rate corresponds adequately. However, our results should still be interpreted with caution.

It should be noted that due to the nature of meta-analysis, researchers are only able to account for contingencies with information provided in published papers or supplemented by the original authors. Therefore, there are various potential confounding variables that are unable to be accounted for. For example, due to the large range of scenarios included across studies, moderation due to differences in moral seriousness is neglected. Previous studies reveal differences in the ethical responses to different ethical issues (e.g., see Vitell & Muncy, 2005) that may vary in moral seriousness. For example, in our dataset, ethical issues that relate to shoplifting (e.g., Vitell et al., 2001) are likely to be perceived as more serious than issues such as improper bike-sharing (e.g., Yin et al., 2018).

Potential Research on Stakeholders

We argue that categorizing stakeholders on a relational level is a valid means of addressing their role in moderating the effects of ethical evaluations. This approach is consistent with arguments from existing ethical decision-making models that formed the conceptual basis for our study. For example, Hunt and Vitell’s (1986) importance of stakeholders and Ferrell and Gresham’s (1985) contingency framework. Such relational categorizations arguably alter psychological distance. Commonly, there is greater observed social distance between individuals in organizational relationships than in personal ones and differentiating between entities on a relational level is a frequently adopted psychological distance manipulation in experimental research (e.g., Lo et al., 2019). However, relational differences are only one way of establishing social distances (see Karakayali, 2009; Trope et al., 2007) and due to the nature of meta-analysis, relational properties are speculative. For example, one could have a personal relationship with their boss or colleague that is not disclosed in the paper. Thus, future research could further explore the moderating role of social distance on the relationship between moral evaluations and ethical responses in relation to the perceived victim and/or beneficiary.

Potential Research on Offline Ethical Consumer Climates and Preventative Measures

Another key consideration worthy of future examination is the ethical climates in offline consumer contexts. We propose that the differences revealed in the effects of deontological evaluations are due to a weaker ethical climate in consumer contexts as compared to organizational contexts. However, to our knowledge, this has not yet been explored. The findings from such research would uncover the feasibility, effectiveness, and conditional restraints of preventative measures.

Although we posit the relevance that each moral evaluation (deontological and teleological) has regarding Fullerton and Punj’s (2004) and Dootson et al.’s (2014) conceptualization of preventative measures (educational approaches vs. deterrence strategies), there is no empirical evidence to substantiate these claims. Various measures have been studied in isolation, mainly deterrence strategies; for example, the use of CCTV (Dawson, 1993), employee vigilance (Esmark et al., 2017), and manipulating victim size (Dootson et al., 2014). However, previous studies point out contentions in the literature about the effectiveness of preventative measures (Sidebottom et al., 2017) and argue that their implementation can have negative effects (Dawson, 1993). We support Mitchell and Chan’s (2002) argument that preventative measures are necessary, as they prevent associated losses a priori. We contest that for it to be effective, the correct approach must be adopted. Moral evaluations (deontological and teleological) offer a strong exploratory basis that is relevant to selecting the most effective preventative measure for encouraging ethical consumer behavior. Future research could explore the moderating role of moral evaluations regarding the effectiveness of different preventative measures on consumers’ ethical responses.

Potential Research on Actual Behavior

An additional potential future research path pertains to the effect of deontological and teleological evaluations on ethical behavior. The relation between judgments and intentions is theoretically (Fishbein & Ajzen, 1975; Hunt & Vitell, 1986) and empirically established (Albarracín et al., 2001; Madden et al., 1992). These types of measures in ethics research are susceptible to social desirability bias (Randall & Fernandes, 1991). Self-reported data are also a common methodological approach (Lewandowski & Strohmetz, 2009) and have been criticized in terms of validity concerns, from which business ethics research is no exception (Randall & Fernandes, 1991). We focused on ethical judgments and ethical intentions, but not specifically on ethical behavior. Most of the prior studies on the effects of deontological and teleological evaluations have also examined ethical judgments and intentions, rather than behavior as the dependent variable. Although Vitell and Hunt (2015, p. 32) argue that there is a strong consistency between intentions and behavior, others (see Fishbein & Ajzen, 1975) note that there can be discrepancy between behavioral intentions and actual behavior. Future research should examine the link between deontological evaluations and teleological evaluations with actual consumer behavior.

Conclusions

Our meta-analytical study validated the effects of deontological and teleological evaluations on ethical judgments and ethical intentions. We also identify several contingency factors that moderate these relationships. In addition, we identify useful practical implications and fruitful avenues for future research. Overall, the findings further the understanding of consumer ethics.