The Authentic Moral Self: Dynamic Interplay between Perceived Authenticity and Moral Behaviors in the Workplace

Regardless of whether true selves are ontologically real, there is a widespread belief in true selves that appears to be psychologically consequential (Schlegel & Hicks, 2011). The way people think about “true selves” seems to mirror the ways that people conceive of the “essences” of other object categories such as “lions” or “trees”. Specifically, people believe both true selves and the essences to be composed of a set of immutable attributes that give rise to observable features (Christy, Schlegel, & Cimpian, 2017; De Frietas et al., 2017; Strohminger, Knobe & Newman, 2017). Importantly, when the essence of an object is changed, people no longer perceived it as the same thing it once was. For example, if a lion loses its essence, it is no longer, a “lion”.

Emerging research suggests that a change in a person’s moral character might lead other people to judge him or her as no longer the same person (Chen, Urminsky, & Bartels, 2016; Heiphetz, Strohminger, & Young, 2017; Strohminger & Nichols, 2014, 2015), particularly if it is a change for the worse (Molouki & Bartels, 2017; Tobia, 2016). For example, people are more likely to infer identity discontinuity when they perceive a change in moral character as opposed to changes in other mental faculties, such as memory, or other individual differences, such as personality (Strohminger & Nichols, 2014).

This suggests that there is a shared belief that the essence of an individual is his or her positive moral character. Consistent with this suggestion, people widely believe that the true selves are morally good and are reluctant to attribute negative moral behaviors to someone’s true self (Newman et al., 2014; Newman et al., 2015; Strohminger et al., 2017; see also Gino, Kouchaki, & Galinsky, 2015; Gino, Norton, & Ariely, 2010). This “good true self bias” is so strong and robust that it has even been observed among people who self-identify as misanthropes (De Freitas, Sarkissian et al., 2017).

Importantly, this belief in morally good true selves has downstream consequences on people’s self-perceptions. If being morally good is integral to individuals’ perceptions of the true self, then behaviors deviating from one’s moral code should shake the basis for perceptions of authenticity. Christy and his colleagues (2016) provided direct support for this contention. In a series of four studies, participants were asked to a) report their daily (im)moral behaviors for five consecutive days (Study 1), b) recall past (im)moral behaviors (Studies 2 and 3), or c) engage in hypothetical donation decisions on which they received false feedback that they were generous (versus selfish, Study 4). Across the studies, thinking about behaving immorally negatively affected perceived knowledge of their true selves.

While research has provided support that perceived (im)moral behaviors engender feelings of (in)authenticity, there is evidence showing that the reverse is also true—trying to be authentic reduces people’s tendency of moral transgression (Kim, et al., 2017; Knoll et al., 2016). We contend that the belief in true selves, as a lay theory, may function like an inner compass, directing people away from immoral behaviors.

The idea that the self directs judgments and behaviors is prevalent in both classical theories (e.g., Higgins, 1987; Swann, Stein-Seroussi, & Giesler, 1992) and more recent empirical research (e.g., Schlegel, Hicks, Arndt & King, 2009; Schlegel, Hicks, Davis, Hirsch, & Smith, 2013; Schlegel, Hicks, King, & Arndt, 2011). The expression of true self, according to many researchers (Deci & Ryan, 1985, 2000; Kernis & Goldman, 2006; Sheldon & Elliot, 1999), is the motivational core of human behaviors. For moral behaviors in particular, the important role of self-concepts has been well-articulated (Jennings, Mitchell & Hannah, 2015). Research on moral identity, for instance, suggests that to the extent individuals’ self-concepts are defined by moral traits, they are more likely to engage in moral behaviors (Aquino & Reed, 2002; Reed & Aquino, 2003; Shao, Aquino, & Freeman, 2008; Winterich, Aquino, Mittal & Swartz, 2013). Similarly, under the belief that the true selves are morally good, individuals should be less likely to engage in morally questionable behaviors.

Indeed, researchers of Authentic Leadership Theory (Walumbwa, et al., 2008) had argued that people cannot be “authentic”, without “complying with high standards of ethical conduct” (p. 94). Recent research (Kim, et al., 2017; Knoll et al., 2016) provided more empirical evidence that efforts to be authentic (i.e., follow one’s true selves) reduce people’s tendency to conduct immoral behaviors. For instance, Kim and his colleagues (2017) manipulated people’s decision-making strategies (i.e., trying to be authentic vs. thinking rationally or realistically) and found that people were less willing to commit moral transgressions when trying to be authentic relative to when they were using other decision-making strategies. Knoll and colleagues (2016) demonstrated that individual differences in trait authenticity negatively predicted moral disengagement (i.e., the propensity to disengagement from self-regulation of moral behaviors, Bandura, 1999), which in turn predicted a reduced tendency to engage in immoral behaviors. These findings suggest that individuals believe they are likely to refrain from engaging in morally questionable behaviors when they are acting authentic.

In sum, previous research suggests a bidirectional relationship between moral behaviors and authenticity. However, this line of empirical research is in its infancy and has almost exclusively focused on college students in laboratory settings judging hypothetical targets or making hypothetical decisions. Relatively little attention has been dedicated to more applied settings, such as the workplace. The workplace, in particular, presents a unique case where the link between moral behaviors and authenticity may not necessarily hold. Unlike the rest of daily routines, behaviors at work are regulated by highly job-specific contextual norms. Violation of norms can incur great cost (e.g., on one’s career, or social relationship). While these norms may or may not deviate from people’s own moral code, people may associate their behaviors at workplace with the externally imposed norm rather than their beliefs about the true self. A manager choosing to protect the environment at the cost of corporate interests may seem virtuous and have made a decision in align with his or her true self. However, she may think this is done merely to conform to social norm and to protect the corporation’s interests in the long run. It is thus an empirical question whether the link between moral behaviors and authenticity hold true in workplace.

The current work thereby examines whether moral behaviors and authenticity are bidirectionally related among samples of people with full-time jobs. Our focus is the generalizability of this relationship in work-related contexts. Therefore, we did not examine or compare to non-work-related contexts where the relationship between moral behaviors and authenticity has been established (Christy et al., 2016; Kim, et al., 2017). We examined both sides of this potentially bidirectional relationship. On the one side, the lay belief in morally good true selves implies that people should experience lower levels of authenticity at work to the extent they perceive their workplace behaviors as deviating from being morally good. Unlike Knoll and colleagues (2016), our focus here is on authenticity as a state of subjective experience (van den Bosch & Taris, 2014a, 2014b) as opposed to authenticity as dispositional trait. On the other side, if following the true self is to be morally good, then people trying to be authentic should avoid behaving immorally at workplace.

We conducted three studies to examine the proposed bidirectional relationship between moral behaviors and authenticity at work. All studies were conducted under IRB approval (Studies 1 and 2 from Texas A&M University, Stuy 3 from both Texas A&M university and Nanjing University). We have also obtained informed consent from all participants. The first two studies made use of online samples from North America. Study 1 employed a between-subject experimental design; we manipulated participants’ perceived moral level (high vs. low) of their work-related behaviors and examined whether such perceptions would alter their feelings of authenticity at work. Study 2 examined the reversed causal relationship using a within-subject experimental design. Specifically, we adapted the procedure of previous research (Kim et al., 2017) and had participants use different decision-making strategies (i.e., trying to be authentic vs. thinking rationally vs. thinking realistically) to judge their willingness to engage in immoral behaviors at work. We predicted that those who trying to be authentic would be less likely to engage in immoral behaviors at workplace. Finally, Study 3 used a daily diary method to demonstrate the association between daily fluctuations in authenticity at work and moral behaviors among a sample of Chinese employees. Although classic theories posit that people from Eastern cultures may view the self differently than those from Western cultures (e.g., independent from vs. interdependent with others, Markus & Kitayama, 1991), recent research has demonstrated that ideas about true self and authentic experiences are quite similar across cultures (English & Chen, 2007, 2011; Slabu, Lenton, Sedikides, & Bruder, 2014). Hence, we expected to see the bidirectional link between authenticity and moral behaviors being replicated among the Chinese sample.

In sum, we hypothesize that moral behaviors and authenticity would reinforce each other in work-specific contexts. We state that we have reported all measures, manipulations, and participant exclusions for the three studies reported here. All sample sizes were determined before any data analysis.

Study 1 aimed to examine the impact of perceived morality on employees’ feelings of authenticity at work. We asked participants to recall aspects of their work that varied in degrees of moral valence. To capture the broad complexity in defining moral behaviors within organizational contexts (e.g., Clegg et al., 2007; Sorbet, 1955; Vallini, 2007), we examined two types of moral behavior relevant to work contexts. Specifically, we borrowed Bolino and Grant’s (2016) distinction between prosocial behaviors (i.e., behaviors adhering to moral norms in work settings, such as being honest with one’s supervisors) and prosocial impacts (i.e., behaviors bringing about positive impacts on the society such as contributing to the greater good), and considered both moral-related behaviors and moral-related impacts. The two factors were crossed, resulting in a two moral-type (specific behavior vs. contribution to society) by two perceived-moral-level (high vs. low) between-subjects design. Based on previous findings (Christy et al., 2016), we expected to see a significant main effect of moral level on authenticity at work such that participants should report being more authentic, to the extent they perceived their behaviors at work as moral.

Two hundred and thirteen participants from the U.S. and Canada participated in the online survey on Amazon’s Mechanical Turk platform. Among them, one hundred and twenty-eight (51 women, 75 men, 2 gender nonconforming) self-identified as current full-time employee or running full-time business. These participants were allowed to complete the rest of the survey; the remaining participants were automatically directed to the end of the study and were informed that they were not eligible. The eligible participants received $1.00 in exchange for their participation. They were mostly Caucasian (71.9%). Their ages ranged from 19 to 66 years (M = 33.83,SD = 9.69). The highest level of education they received ranged from high school graduate diploma or the equivalent (e.g., GED, 8.6%) to doctorate degree (1.6%, mean fell between associate degree and bachelor’s degree). The sample size was determined by power analysis based on pre-specified analytical plans following the recommendations of Pocock (1977) and Lakens (2014). Specifically, we had originally planned to collect 256 participants. To save the labor and resources of the laboratory, we had planned to conduct a sequential interim analysis when half of the planned sample was collected. The results of interim analysis reached preset critical value so that we terminated data collection with the current sample. More detailed information on the analytical plans, materials, exclusion criteria and rules for terminating data collection are pre-registered and can be found at https://osf.io/bzcrq/. A sensitivity analysis revealed that a sample size of 128 would be able to detect an effect size of η² = .059 or similar (with α = .05, power = .80).

The whole procedure was delivered via Qualtrics software. Participants were randomly assigned to one of the four conditions, and were asked to describe three ways they engaged in (a) “morally good behaviors at work (e.g., being honest with your supervisor or clients)” (the high-moral-behavior condition), (b) “morally questionable behaviors at work (e.g., being somewhat dishonest with your supervisors or clients)” (the low-moral-behavior condition), (c)“behaviors at work that contribute to the greater good (e.g., promoting the welfare of society)” (the high-moral-contribution condition), or (d) “behaviors at work that do not contribute to the greater good (e.g., not promoting the welfare of society)” (the low-moral-contribution condition).¹ All participants then completed a measure of perceived authenticity at work. At the end of the survey, participants completed a seriousness check item.

We assessed authenticity at work with the 12-item authenticity measure developed by van den Bosch and Taris (2014a). The measure consists of three subscales: authentic living subscale assesses feelings of being authentic (e.g., “At work, I always stand by what I believe in”), self-alienation subscale assesses feelings of inauthenticity at workplace (e.g., “At work, I feel alienated”), and accepting external influence subscale measures the extent to which individuals feel that their behaviors are influenced by others (e.g., “At work, I feel the need to do what others expect me to do”). Participants indicated on a 7-point Likert scale to what extent each statement described them at work (1 = does not describe me at all, 7 = describes me very well). The subscale of accepting external influence indicated low reliability (Cronbach’s α = .46) which could not be fully remedies by dropping poorly performing items. In addition, it was only weakly correlated with the other two subscales (rs = .09 and .18). Given the seriousness of the concerns regarding the reliability and validity of this subscale, we deviated from our pre-registration and created a composite using only the items from the other two subscales (M = 5.28,SD = 1.06; Cronbach’s α = .86).

Following the recommendation of Aust and his colleagues (Aust, Diedenhofen, Ullrich, & Musch, 2013), we included a single item seriousness check at the end of the entire survey (i.e., “It is important for our scientific analysis that we only use data from people who took the study seriously. It would be very helpful if you could tell us at this point whether you have taken the study seriously or not. You will be compensated for your time either way”). Participants were asked to indicate whether they had taken the test seriously or just clicked through. All participants indicated that they had taken the study seriously.

We conducted a two-way ANOVA to examine the effects of perceived moral level and moral type on authenticity at work (Figure 1). As predicted, the analysis revealed a significant main effect of moral level,F (1, 124) = 10.77, p < .001, η_p² = .080. People who were instructed to recall highly moral aspects of their job reported a higher level of authenticity at work (M = 5.57, SD = 1.04) than those recalled aspects of their job that lower in morality (M = 5.00,SD = 1.01). In addition, there was also a significant main effect of moral type, F (1, 124) = 5.44, p = .021, η_p² = .042. People who were instructed to recall contributions to society reported a higher level of authenticity at work (M = 5.49, SD = .92) than those who recalled specific behaviors (M = 5.09, SD = 1.15). The two-way interaction between moral type and perceived moral level was not significant, F (1, 124) = .020, p = .89, η_p² < .001.

It should be noted that the analyses we present here deviated from our preregistration because we dropped the accepting external influence subscale of van den Bosch and Taris’ (2014a) measure due to low reliability. The pre-registered analyses using the full scale revealed no significant main effect of morality on the authenticity composite. While we hesitate to make strong conclusions from that scale given the psychometric concerns, details of this analysis as well as exploratory analyses investigating accepting external influence separately from the other dimensions, are presented in the supplemental materials.

Study 1 indicated that employees who recalled highly moral aspects of their behaviors at the workplace reported to be more authentic than those recalled less moral aspects of their behaviors. There was also a main effect of perceived moral type. However, this main effect was not of interest and appeared to be driven by the fact that the lack-of-good impact condition was more neutral than the low-moral-behavior condition. The interaction effect was non-significant. However, this null effect should be interpreted with caution given that the sample size was likely underpowered for detecting an interaction effect (Giner-Sorolla, 2019).

Study 1 thereby established the causal role of perceived moral behaviors in feelings of authenticity at work. We then proceeded to test the reverse relationship in Study 2.

The goal of Study 2 was to test whether intending to be authentic would attenuate the tendency for people to engage in immoral behaviors. Following Kim and colleagues’ paradigm (2017), we manipulated decision-making strategies (i.e., trying to be authentic versus thinking rationally and thinking realistically) within participants and had them engage in hypothetical decision-making scenarios involving unethical behaviors at workplace (e.g., wasting, mismanaging, or abusing organizational resources). Rational thinking and realistic thinking were selected for comparison because these two strategies are two of the most common decision-making strategies people use and have been employed in previous studies (e.g., Schlegel et al., 2013). We predicted that people trying to be authentic (relative to thinking rationally or realistically) should be less willing to engage in morally questionable behaviors at work in the hypothetical scenarios.

One hundred and thirty-three participants from the U.S. and Canada participated in an online survey posted on Amazon’s Mechanical Turk platform. Among them, one hundred and five (46 women, 59 men) were self-identified as current full-time employee or running full-time business and were allowed to complete the rest of the survey; the remaining participants were again automatically directed to the end of the study where they were informed that they were not eligible. Eligible participants received $2.50 in exchange for their participation. Most of them were White (70.5%). Their ages ranged from 18 to 59 years (M = 35, SD = 9.64). The highest level of education participants received ranged from high school graduate diploma or the equivalent (e.g. GED, 12.4%) to doctorate degree (19%, mean fell between trade/technical/vocational training and associate degree). A power analysis through G-power software (Erdfelder, Faul, & Buchner, 1996) suggested that for an adequately powered (i.e., of .90) within-subject experiment, a minimum sample size of 105 participants were needed to detect an effect of η_p² = .02, atp =.05 (two-tailed). Therefore, we aimed to recruit 105 participants. A sensitivity analysis revealed that with a sample size of 106 we had 80% power to detect an effect size of η² = .015 or similar (with r set at .50 and α = .05).

We included the same seriousness check item as used in Study 1. One participant indicated that he or she had just clicked through and were removed from data analysis, This resulted in a final sample of 104 participants.

The whole procedure was delivered via Qualtrics software. All participants completed a monetary decision task and a business decision task in random order. Both tasks contained a number of decision situations that could potentially occur in workplace. Within each task, participants made each decision three different times once for each of the three decision-making strategies (i.e., being authentic vs. being rational vs. being realistic). Full scripts, materials, and data are available upon request or online at https://osf.io/47eyb/.

The basic structure of this task was adapted from the Moral Foundations Sacredness Scale (Graham, Haidt, & Nosek, 2009) which asks participants how much money it would take for them to violate the principles of various moral foundations. We used this structure but replaced the moral foundations violations with 17 items describing unethical behaviors relevant to the workplace. The behaviors were selected from Kaptein (2008). Examples include “falsifying or manipulating financial reporting information” and “Wasting, mismanaging, or abusing organizational resources”. Participants indicated how much money they thought it would take for them to engage in each unethical behavior on a scale from 1 to 8 (1 = $0 (I’d do it for free), 2 = $10, 3 = $100, 4 = $1,000, 5 = 10,000, 6 = 100,000, 7 = 1,000,000, 8 = never for any amount of money). This served as a measure for the unwillingness to engage in unethical behaviors.

Participants were presented with 6 scenarios from Ashton and Lee (2008). Each scenario placed participants in a hypothetical business setting and presented them with an unethical behavior (e.g., “voting for profitable mining operations that cause serious long-term pollution of water and soil”). Participants indicated whether they would engage in that behavior on 4-point Likert scales (e.g., “Would you invest your pension fund‘s money in the company that runs this sport”, 1 = Definitely Not, 2 = Probably Not, 3 = Probably Yes, 4 =Definitely Yes). We reversely coded all the items so that a higher score indicated a lower tendency to engage in the unethical behaviors.

For each item in the two decision tasks, we asked participants to indicate their decision three times, each time using a different decision-making strategy. One strategy was to be as authentic as possible (i.e., “think about who you are ‘deep inside’, and make your decisions based on how they align with your ‘real’ self”). Another strategy was to think rationally (i.e., “being logical and objectively considering the pros and cons of the available options”). The third strategy was to think realistically about how one would actually respond if faced with the situation (i.e., “select the option that best reflects how you would actually respond if you faced these decisions in real life”). Since participants’ responses were highly inter-correlated across the 17 unethical behaviors in the monetary decision task (αs = .95, in both the authentic and the realistic conditions; α = .94, in the rational condition) and across 6 scenarios in the business decision task (αs = .80, in both the authentic and the realistic conditions; α = .73, in the rational condition), within each task, we collapsed participants’ responses under each decision-making strategy into one composite.

Table 1 presents the means (SDs), and correlations among the dependent variables. We conducted a repeated measures MANOVA test to examine the effect of decision strategies on people’s willingness to engage in the hypothetical unethical behavior at work in the two tasks. The analysis revealed a significant overall effect of decision-making strategies, Wilks’ λ = 0.60, F (4, 100) = 16.68, p < .001, η_p² = .40. Planned Helmert contrast (comparing authentic condition against realistic and rational thinking conditions) revealed the effect was in our predicted direction, F (1, 103) = 16.87,p < .001, η_p² = .14 for monetary decision task, and F (1, 103) = 33.65, p < .001, η_p² = .25 for business decision task (see Figure2a and 2b).

We then conducted follow up analyses to examine the differences between the conditions. For the monetary decision task, pairwise comparisons adjusted with Bonferroni procedure revealed a significant difference between authenticity and realistic thinking as we predicted, t(103) = 5.72,p < .001, d = .56, 95% CI [.23, .57]. However, the difference between authenticity and rational thinking strategies, though trending towards our predicted direction, was not significant,t(103) = 2.19, p = .091, d = .22, 95% CI [–.02, .42]. Compared to rational thinking, thinking realistically prompt significantly more willingness to make an immoral decision in the hypothetical scenarios, t(103) = –2.81, p = .02, d = .28, 95% CI [.03, .37]. For the business decision task, the analyses revealed significant differences in the hypothesized direction between authenticity and the other two strategies (for realistic thinking,t(103) = 4.34, p < .001, d = .45, 95% CI [09, .31], and for rational thinking, t(103) = 5.35,p < .001, d = 50, 95% CI [.19, .51]). Relative to rational thinking, thinking realistically prompted less willingness to engage in the hypothetical moral transgressions, t(103) = 2.43,p = .05, d = 24, 95% CI [–.30, .00]. Overall, the results of Study 2 supported our prediction; people who tried to be authentic (compared to be rational or realistic) were less willing to engage in hypothetical immoral working behaviors.

Study 2 demonstrated that trying to be authentic undermined the tendency to behave unethically when facing hypothetical moral transgression scenarios at the workplace. Taken Studies 1 and 2 together, our data provided support for the bidirectional relationship between moral behaviors and authenticity at workplace. On the one hand, perceiving oneself as behaving in a moral way engenders feelings of authenticity at work. On the other hand, trying to be authentic reduces people’s tendency to engage in morally questionable behaviors in hypothetical scenarios.

Although Studies 1 and 2 provided experimental evidences for our hypothesis, both studies are controlled experiments and could suffer from low external validity. Our reliance on instructed recall in Study 1 and the hypothetical nature of the unethical behaviors in Study 2 might produce artificial findings not generalizable to real work settings. In addition, samples of Studies 1 and 2 were all from relatively Westernized cultural backgrounds. It is important to examine the generalizability of these findings to other cultures. To address these issues, in Study 3, we used a daily diary method to examine the covariation between authenticity and ethical behaviors at work.

The aim of Study 3 was to examine the external validity of our findings by studying the relationship between morality and authenticity in real-life working settings. In addition, we aimed to examine whether the relationships would generalize to people from a different cultural background. To these ends, we conducted a 5-day daily study among a sample of full-time workers in China. Previous studies have shown that like people from Western cultures, people from interdependent cultures also believe that the true self is morally good (De Freitas, Sarkissian, et al., 2017). Hence, we hypothesized that perceived morality and authenticity at work would covary with each other in the current investigation, both within and across individuals.

Moreover, we also explored whether authenticity at work and self-reported (im)moral behaviors would have downstream consequences on meaning in work and job satisfaction, two important indices of well-being in organizational settings (Judge & Klinger, 2008; Schnell, Höge, & Pollet, 2013; Sousa-Poza & Sousa-Poza, 2000; Steger, Dik & Duffy, 2012). Previous research has suggested that authenticity and moral behaviors could both be potential predictors of well-being in work settings (Allan, Duffy, & Collisson, 2017; Grant, 2008; Grant & Campbell, 2007;Grant & Sonnentag, 2010; Ménard & Brunet, 2011). We thereby predicted that authenticity and moral behavior at work should contribute to meaning in work and job satisfaction simultaneously.

A total of 189 employees (50 men, 119 women, 20 did not indicate gender, ages ranged from 20–54 years, M_age = 30.29,SD = 6.22) voluntarily participated in this study. We determined this sample size via the software PinT (Snijders, Bosker, & Guldemond, 2007). Parameters were estimated based on previous findings about the association between morality and authenticity (i.e., around 70% of the variance was at level-2; Christy et al., 2016). The analysis suggested that with 5 data points for each individual, a minimum of 150 participants should be included to detect a small effect (r = .10, with α = .05, power = .80, residual variance = .50, and 70% of the variance locating at level-2). Expecting attrition, we attempted to recruit about 200 participants.

Participants were recruited through a web course in introductory psychology. As in Studies 1 and 2, participants were eligible only if they had a full-time job. The majority of participants (58%) had a bachelor’s degree, and their mean monthly income was around CNY8000 (US$1200). Their jobs ranged from secretaries, engineers, managers to physicians, and job tenures ranged from 0.3 to 36 years (M = 7.24, SD = 6.47).

Participants completed the daily surveys right after work over 5 consecutive working days.² Before taking part in this study, participants were given information about the structure and content of the survey, as well as the exact time of day they should complete the questionnaires. Each day at 4:00pm, the questionnaire was activated and a reminder including a link to the survey was sent to each participant through their smartphone. Participants were asked to complete the questionnaire right before or after leaving their workplace. Considering that participants might work late, we sent a second reminder at 8:00pm every day and closed each survey the next morning at 8:00am. In total, participants provided 780 valid data entries. A summary of this study was preregistered online (study materials, datasets are also available at https://osf.io/hz9qu/). A sensitivity analysis revealed that a sample size of 189 with 4 data points for each participant on average would be able to detect an effect size ofr = .096 or similar (with α = .05, power = .80, residual variance = .50, and 70% of the variance locating at level-2).

The daily surveys included measures of perceived authenticity at work, moral (immoral) behaviors, and prosocial (i.e., morally good) impacts. In addition, we collected daily job satisfaction and meaning in work as potential downstream variables. To examine whether the relationship between morality and authenticity at work held independent of other factors, we also measured several variables that may influence judgments of morality or authenticity (i.e., emotional status, job involvement, job performance, relationships with coworkers, and workload). All scales originally in English were translated into Chinese with the standardized back-translation procedure (Brislin, 1970).

As in Study 1, we measured moral behaviors both in terms of specific behaviors and in terms of positive impact on society. To assess morally-related behavior, we used the same five moral (i.e., fair, courageous, honest, helpful, and hard-working) and immoral (i.e., cruel, malicious, uncooperative, untrustworthy, and selfish) words as those used in Christy et al. (2016) to assess moral (immoral) behaviors. Participants indicated to what extent each adjective described their behaviors on each day along a 7-point scale (1 –not at all to 7 – very much). Their responses to these items were aggregated to form a moral (M = 5.52, SD = .96, α = .82) and an immoral (M = 5.52, SD = .96, α = .80) index respectively.

To assess moral impact, we adapted Grant’s (2008) three-item scale. A sample item is “My work today really made no difference to the world” (reverse scored). Participants indicated how much each item described their behaviors on a 7-point scale from 1 – not at all to 7- very much (M = 4.68, SD = 1.32, α = .77).

We measured the same three dimensions of authenticity at work as in Study 1 (i.e., authentic living, self-alienation and accepting external influence), but each with two items this time, so as to keep the daily survey short. Again, the two items on accepting external influence showed low correlations with each other (r = .38), and with other items (within person rs < .37). Therefore, we removed the items from further analysis. An authenticity index was formed by combining the four items (with the two self-alienation items reversed), M = 5.19, SD = 1.24, α = .85.

Two items asked participants’ emotional status at work: “In general, I am happy at work today”, and “In general, I am not happy at work today”. The two items were combined (r = –.69, M = 5.16, SD = 1.38). Job involvement (M = 5.24, SD = 1.34), job performance (M = 5.24, SD = 1.21), relationships with coworkers (M = 5.95, SD = 1.06), and workload (M = 4.03, SD = 1.84) were each measured by one face-valid item.

Daily job satisfaction was measured with one item: “In general, I am satisfied with my job today”. Meaning in work was measured with two items: “My work today is meaningful”, and “My work today contributes to my life’s meaning”. The two items were combined (r = .68, M = 4.68,SD = 1.40).

Table 2 presents the means (SDs) and correlations among the variables. To investigate the relationships of daily moral and immoral behaviors, prosocial impact to authenticity, we conducted a two-level hierarchical linear modeling (HLM) analysis. We treated daily work authenticity as the dependent variable.³ The unconditional model revealed an intraclass correlation coefficient of .60, meaning that 40% of the variance in authenticity at work occurred within person (at Level 1), which was comparable to previous findings (Christy et al., 2016, Study 1). We then included daily moral (immoral) behaviors, and prosocial impact (all centered within-person) as Level 1 predictors. The personal means of these variables were entered in Level 2. This allowed us to examine how morality and authenticity covaried at both the within- and between-person levels (Fleeson, 2007; Raudenbush & Bryk, 2002). All effects were estimated as random.

Fixed effects of this model are presented in Table 3. As predicted, self-reported moral behaviors and prosocial impact were positively associated with authenticity at work, while immoral behaviors were negatively associated with authenticity, both within and across individuals. That is, individuals experienced higher levels of authenticity at work on days they reported having behaved more morally, less immorally, and on days they believed their work contributed more to the greater good. Meanwhile, across participants, those who behaved in morally good ways and those who had more prosocial impact through their work felt more authentic at work. All the variance components were non-significant (all ps > .21), indicating that the associations between the predictors and authenticity at work did not vary significantly across individuals.

When we included the covariates in the model (i.e., emotional status, relationships with coworkers, job involvement, workload, and job performance), all the main effects remained significant except the effect of moral behaviors at the between-subjects level (b = .072, SE = .049,t = 1.48, p = .14).

To explore how morality and authenticity were associated with job satisfaction and meaning in work, we built a multilevel structural equation model (MSEM) with moral\immoral behaviors, prosocial impacts and authenticity at work serving as predictors, and job satisfaction and meaning in work as outcomes (Figure 3). We found that, except immoral behaviors at the between-subjects level, all the four predictors were significantly associated with job satisfaction, both between-and within-subjects ((bs ranged from –.081 to .54, ps ranged from .048 to < .001). At the within-subjects level, moral behaviors, prosocial impacts and authenticity significantly predicted meaning in work (all bs > .18,ps < .001). At the between-subjects level, prosocial impacts (b = .60, SE = .066, t = 9.18, p < .001) and authenticity (b = .41,SE = .061, t = 6.65, p < .001) also significantly predicted meaning in work. The association between moral behaviors and meaning in work was marginally significant (b = .16,SE = .085, t = 1.90, p = .058). However, a positive association emerged between immoral behaviors and meaning in work (b = .21, SE = .058, t = 3.53, p < .001).

It should be noted that the analyses were deviated from our preregistration because we dropped the accepting external influence subscale due to low reliability. The pre-registered analyses using the full composite were largely consistent with the ones reported here. While we are hesitant to make many conclusions based on the pre-registered analyses given unexpected psychometric concerns, these analyses as well as exploratory analyses investigating accepting external influence separately from the other dimensions, are presented in the supplemental materials.

Results of Study 3 extended the experimental findings in Studies 1 and 2 to real-life workplace among a Chinese sample. The relationship between authenticity and morality was observed both between- and within-individuals. Moreover, authenticity and morality further predicted meaning in work and job satisfaction. These findings corroborated the importance of both authenticity and morality at workplace. The only result that went contrary to our hypothesis is that, after controlling for authenticity and prosocial impacts, immoral behaviors indicated a positive association with meaning in work across participants. This result may not be very surprising, however, given the very high correlations of both prosocial impacts and authenticity with meaning in work (rs = .74 and .62 respectively). These variables may have had suppression effects on immoral behaviors (Friedman & Wall, 2005).

The results of the three studies strongly supported our proposed bidirectional relationship between authenticity and morality in the workplace. People who were reminded of their own highly moral behaviors in the workplace reported a higher level of authenticity at work than those who recalled less moral behaviors (Study 1). Meanwhile, people who tried to be authentic (vs. think rationally or realistically) reported reduced willingness to engage in morally questionable behaviors (Study 2). The link between moral behaviors and authenticity at work remained robust both across and within individuals, and both variables were associated with positive outcome variables such as meaning in work and job satisfaction (Study 3). Our findings added to a line of previous research demonstrating the bidirectional relationship between morality and feelings of authenticity (Christy, et at., 2016; Christy, Kim, Vess, Schlegel & Hicks, 2017;Gino, et al., 2015; Kim, et al., 2017) and extended the findings to workplace and non-Western cultures. As shown in previous research (English & Chen, 2007, 2011;Slabu, et al., 2014), the core of authentic experiences could be remarkably similar—especially within the same context—across different cultures. Study 3 corroborated this possibility, demonstrating that being morally good is integral to authentic experiences in the workplace in a non-Western culture.

Although we studied authenticity at work as individuals’ subjective sense of being in touch with their true selves, the current research dovetailed with the Authentic Leadership Theory (Gardner et al., 2005; Luthans & Avolio, 2003;Walumbwa, et al., 2008, but see Algera & Lips-Wiersma, 2012). The theory posited that authentic leaders (i.e., leaders whose behaviors are guided by their internalized values or overarching self-concepts as opposed to external rewards or punishments) tend to possess higher levels of moral development and are inherently oriented towards moral behaviors. While Authentic Leadership Theory focused on authentic style of leader behaviors, we examined feelings of authenticity in work-specific contexts (van den Bosch & Taris, 2014a, 2014b); our evidence is thus in parallel to Authentic Leadership Theory’s propositions. However, since we did not distinguish between leaders and followers, our findings have the suggestion that the relationship between morality and authenticity could remain robust, regardless of what positions people might hold in organizations.

The current studies could contribute to the research on subjective experiences of authenticity. Feeling authentic at work bears important implications for work engagement, job satisfaction and a variety of other positive work-related outcomes (e.g., van den Bosch & Taris, 2014b; Ménard & Brunet, 2011; Yagil & Medler-Liraz, 2014). Previous research mostly viewed workplace authenticity is as engendered by the degree to which one “fits” into his or her environment (van den Bosch & Taris, 2014a; Schmader & Sedikides, 2017). We believe that immoral behaviors may be one reason people experience lack of fit and authenticity at workplace at the first place. Work settings encouraging moral transgressions may inevitably generate a mismatch among employees between their beliefs of true self (i.e., to be moral) and such environmental characteristics. Trying to “fit” into such settings may lead people astray and cut them further away from their true self. To cultivate feelings of authenticity in organizational settings, we think job tasks should be designed to emphasize ways that people could behave to abide work norms or to contribute to the greater good.

Our findings could also inform research on the role of self in the origin of moral behaviors at workplace. Indeed, how the self is related to moral behaviors remains a controversial question. While the self is seen as a key component of moral agency (Jennings et al., 2015), it is also sometimes casted in a negative light (e.g., “self-interests”; see Grant, 2009) and seen as a barrier to act prosocially. In addition, even within the literature positing the self as an antecedent of moral behaviors, the degree to which one possesses moral self-concepts or moral identity is often conceptualized as a matter of individual differences (e.g., Aquino & Reed, 2002). Our studies, on the other hand, demonstrated that acting morally at workplace may be a natural consequence of people’s attempts to be authentic and to follow what they believe to be their true selves. Moral behaviors and the true self are so deeply tied that following true selves should foster the tendency to behave morally regardless of the individual differences in moral identity.

Given the role of authenticity in moral behaviors, our findings further shed light on how to develop interventions against immoral behaviors in work settings. Unethical behaviors in the workplace can bring about devastating consequences to employees, corporations, and even the society as a whole (e.g.,Bakir & Vitell, 2010; Orlitzky, Schmidt, & Rynes, 2003). To manage these behaviors, governments and organizations often rely on the imposition of laws and related legislation. Yet, overly relying on rules and restrictions may sometimes backfire and undermine people’s tendency to act according to their true selves (Deci, Koestner, & Ryan, 1999). From our perspective, those external forces (e.g. rewards and punishments) may inhibit workers’ spontaneous tendency to refrain from moral transgressions. To combat immoral behaviors, we tentatively propose an alternative approach, that is, to encourage employees to resort to their inner moral compass, their beliefs about their true selves. Doing so may enhance not only the tendency to behave morally, but also the tendency to embrace the moral behaviors wholeheartedly. Given the increasing interests in fostering benevolence (Arieli, Grant & Sagiv, 2014), our findings call for attentions to be dedicated to understanding the role of true self beliefs in guiding moral behaviors in organizational settings.

Still, if peoples’ true selves are directed to moral behaviors, one may wonder why immoral behaviors (e.g., fabricate false accounts) could occasionally occur in organizational settings. One possibility based on our findings is that in those cases, people may not be actually following their true selves. For example, people may be driven by pragmatic and realistic thinking. As the results of Study 2 suggested, relative to people trying to be authentic, those who thought realistically were more likely to commit moral transgressions at workplace. It is thus likely that transgressors in real work-settings do not wish harm out of their true selves, but simply pursue whatever “good” (e.g., money) that is readily available without an awareness of ethical consequences (Butterfield, Trevin, & Weaver, 2000).

In addition, people could be misinformed about the impact of their behaviors or they could endorse deviant values. The specifics of business ethics are controversial and at times lacking in consensus even among experienced philosophers (see Moriarty, 2017). It is thus not surprising that people may engage in unethical behaviors under the belief that they are not acting in immoral ways. In fact, recent research found that among managers high in Machiavellianism, the relationship between self-report authentic leadership and observed moral actions is somehow reversed—the more authentic they believe they are, the less moral they are seen from the perspective of an observer (Sendjaya, Pekerti, Härtel, Hirst, & Butarbutar, 2016). Machievellianism is a set of values espousing that ends can justify means (Den Hartog & Belschak, 2012). Therefore, provided that people hold deviant values, they may consider it justified to commit transgressions even when trying to be authentic. Future research should therefore examine individual differences in moral values as a potential boundary condition for the effect we observed.

Last, immoral behaviors could also originate from organizations instead of individuals. Unethical corporations (e.g., a firm that illegally logs forests for profits) may have the power to institutionalize and reinforce their members into behaving in morally questionable ways (Ashforth & Anand, 2003; Jackall, 1988; Vaughan, 1992). Employees in unethical corporations could experience pressure and be forced to deviate from what they truly want to do (otherwise, they would be punished). Or, employees may engage in rationalization processes, attending only to aspects of their work that seem moral. For instance, it is likely for employees in unethical corporations to maintain a sense of morality—and thereby authentic feelings—by focusing on the specific behaviors to fulfill their job duties rather than their work’s impact over society (e.g., claiming that “it is my job” even when fully aware that their work may be bad for the society). Future research could extend our findings by examining these possibilities within specific types of organizations (e.g., tobacco industry).

The current research, though having the merit of rigorousness (in terms of pre-registration and adequate power) and diverse sample, is limited in that we failed to include an objective measure of (im)moral behaviors (e.g., behavioral observation or other-reported measures). Rather, we inferred (im)moral behaviors from hypothetical decision-making (Study 2) and self-reported behaviors in terms of particular moral attributes (Study 3). Though the effects based on these measures are statistically robust and consistent with our predictions, such an operationalization of moral behaviors does not rule out the possibility that what people claim may deviate from what they would actually do. It is therefore possible that individuals may believe that they would behave morally when following their true selves, but such a belief does not translate into actual behaviors.

It is also worth noting that our evidence in Study 3 regarding the inter-relationships between authenticity, moral behaviors, meaning in work and job satisfaction are yet correlational. While our interpretation is consistent with the existing literature that enhanced well-being is a downstream consequence of authenticity and moral behaviors (Allan et al., 2017; Ménard & Brunet, 2011; Grant, 2008; Grant & Campbell, 2007; Grant & Sonnentag, 2010), it remains an open possibility that the reverse is true: it is meaning in work and job satisfaction that cause feelings of authenticity and lead to moral behaviors. Future research could explore the opposite direction of this link by manipulating job satisfaction and meaning in work and examining their effects on authenticity and morality.

Last, it remains an open question whether the link between moral behaviors and authenticity we found could be generalized to more general positive behaviors (e.g., being considerate). There is some evidence suggesting that feelings of authenticity is associated with positive over negative behaviors in general (Jongman-Sereno & Leary, 2016). Indeed, Strohminger, Knobe and Newman (2017) argued that lay beliefs about true self are not merelyin essence moral, but “by default” positive in valence (p. 553). The implication is that feelings of authenticity at work should be bidirectionally related to positive behaviors that are not explicitly relevant to morality (e.g. competent behaviors). Nevertheless, test of this extension of our hypotheses is beyond the scope of current research. Future research may manipulate behaviors that are less related to morality and see if it influences authenticity judgments. This should help clarify whether our findings are parts of a more general phenomenon.

Being authentic is intertwined with behaving morally—the current research found consistent evidence in support of a bidirectional causal relationship between authenticity and morality at work. Our findings bear the potential to shed light on the antecedents and consequences of authenticity and ethical behaviors at work. While working ethics can be controversial (Clegg et al., 2007; Sorbet, 1955; Vallini, 2007), we recommend trusting the inner voice from what one believes about one’s true self, and acting accordingly—this enhances both morality and well-being at work.

Materials and data for Study 1 can be retrieved from https://osf.io/bzcrq/. Materials and data for Study 2 can be retrieved from https://osf.io/47eyb/. Materials and data for Study 3 can be retrieved from https://osf.io/hz9qu/.

The open access publishing fees for this article have been covered by the Texas A&M University Open Access to Knowledge Fund (OAKFund), supported by the University Libraries and the Office of the Vice President for Research.

The authors have no competing interests to declare.

Rebecca Schlegel is an editor at Collabra: Psychology. She was not involved in the peer-review of the article.

• Contributed to conception and design: Rebecca Schlegel, Joshua Hicks, Hong Zhang

• Contributed to acquisition of data: Changkai Chen, Kaiyuan Chen, Hong Zhang

• Contributed to analysis and interpretation of data: Rebecca Schlegel, Kaiyuan Chen, Hong Zhang

• Drafted and/or revised the article: Hong Zhang, Kaiyuan Chen, Rebecca Schlegel, Joshua Hicks

• Approved the submitted version for publication: Rebecca Schlegel, Kaiyuan Chen, Hong Zhang, Joshua Hicks, Changkai Chen

The author(s) of this paper chose the Open Review option, and the peer review comments can be downloaded at: http://doi.org/10.1525/collabra.260.pr