Keywords
ethical leadership, morality-as-cooperation, value congruence, moral values, moral behavior, perceived ethicality
This study investigates the influence of perceived follower–leader congruence in basic moral behaviors, as defined by the morality-as-cooperation theory (MAC), on the perceived level of ethical leadership. Although important for the leadership theory and practice, this association is scarcely covered by the literature.
The study employed a mixed-methods approach, consisting of a one-site quantitative survey in Phase 1 carried out in a Russophone logistics company, and a follow-up series of semi-structured interviews within the same population in Phase 2.
The results of Phase 1 showed a moderate positive correlation (r(54) = [.34;.64], p < .05) between perceived congruence in each MAC element and the perceived level of ethical leadership, whereas the findings from Phase 2 demonstrated that this association is causal. Moreover, the research unexpectedly revealed that the influence of the congruence in care for kin, heroism, and deference is mediated by the congruence in loyalty to the group and reciprocity, with the significance level at p < .05 for all indirect effects. The study also found a moderating role of duration of leader’s moral behavior observation by followers.
This research advances ethical leadership literature by applying the MAC theory to studying the antecedents of the perceived leaders’ ethicality and revealing the mediating role of some basic moral values in assessing a leader’s ethicality. Besides, the study contributes to the cross-cultural validity of both MAC and ethical leadership theories by testing the respective instruments in a Russophone organization.
ethical leadership, morality-as-cooperation, value congruence, moral values, moral behavior, perceived ethicality
The emphasis on ethical behavior in organizational circumstances and its specific relevance to leadership practice was fueled by corporate scandals at the beginning of this century (Hoch, Bommer, Dulebohn, & Wu, 2018; Northouse, 2019). However, researchers differ substantially on their understanding of what exactly should be considered as ethical or moral (Westra & Andrews, 2022). Despite the claim that no universal ethics for humanity exists due to the dependence of ethics on cultural contexts (Wong, 1984), the recent developments in evolutionary moral psychology have offered various accounts of the moral basics that underlie the whole variety of ethics and morality that emerge in various circumstances (Curry, 2016; DeScioli, 2016; Haidt & Joseph 2004; Rai & Fiske, 2011; Schein & Gray, 2018).
Recent literature stresses the importance of knowledge about antecedents of followers’ perception of a leader as ethical, and points to the lack of empirical research in this field (Banks, Fischer, Gooty, & Stock, 2021; Brown & Treviño, 2014). Given that a positive influence of such a perception on the group’s outcomes has been covered by numerous papers (Bedi, Alpaslan, & Green, 2016), the knowledge of factors influencing the perceived level of ethical leadership turns out to be a valuable tool for ethical leaders aiming to enhance the outcomes.
Paraphrasing Edwards and Cable’s (2009) definition of values, personal moral values may be specified as general beliefs about morally acceptable and desirable behaviors or end states. Hayes, Pankey and Gregg (2002) argue that personal value congruence occurs when someone’s behavior corresponds with self-attributed values. This entails the importance of moral value congruence for ethical assessment of leaders’ behaviors. Banks et al. (2021) posit that the congruence of leaders’ behavioral signals, such as prosocial values, with their followers’ moral identities, which is defined by Wang and Hackett (2020) as sets of self-associated moral values, determines their evaluation as ethical or unethical. However, while the association of value congruence with ethical leadership has been addressed by recent publications, there is still a scarcity of empirical studies on this subject (Peng & Lin, 2017).
More specifically, few studies focus on follower–leader moral value congruence as a factor influencing followers’ ethical evaluation, including Egorov, Kalshoven, Pircher Verdorfer, and Peus (2020) and Pircher Verdorfer and Peus (2020), and neither provide empirical evidence of causality. This research aims to provide such evidence by a one-site mixed-method inquiry to the follower–leader congruence in moral values as a causal antecedent of followers’ moral evaluation of ethical leaders’ behaviors.
Moral judgment on a leader’s behavior employs various criteria, either rational or irrational, and conscious or unconscious (Alfano, 2022; Haidt, 2012). Curry (2016) proposed the morality-as-cooperation (MAC) theory, which posits that morality represents a set of typical solutions for the cooperation problems that humanity has faced in our species’ history. Intra-group interactions and gene–culture co-evolution have led to the adoption of such solutions and their conceptualization as morally acceptable behaviors (Tomasello, 2016; Buchanan, 2020). Curry (2016) argues that they correspond to the cooperative games studied by game theory, which involve such behavioral strategies as care for kin, loyalty to the group, reciprocal exchange, dominance and heroism, deference and submission, fairness in dividing disputed resources, and respect to prior possession ( Table 1). These behaviors are thus widely considered moral (Curry, 2016), and turn out to be universally morally relevant to all cultures across the globe (Alfano, Cheong, & Curry, 2022; Curry, Mullins, & Whiterhouse, 2019b).
According to Curry, Alfano, Brandt, and Pelican’s (2021) suggestion, the MAC basic moral behaviors, or MAC values, represent basic elements of human morality, that, taken in various combinations, can possibly constitute any known moral value. Assuming this hypothesis on morality as a combinatorial system as true, the appreciation of basic moral behaviors should be present in any moral assessment, and if a leader’s behavior demonstrates adherence to those basic moral elements, then it should be evaluated as ethical.
Summarizing the above, Curry and his colleagues suggest that basic cooperative moral values represent the culturally universal basis of morality, in whatever values it may be explicitly expressed or implicitly perceived. Hence, the MAC basic moral values can serve as culturally universal criteria for moral evaluation of a leader. Therefore, the aim of this study is to explore whether congruence in MAC basic moral elements causally influences the evaluation of a leader as ethical, which, to the knowledge of the authors, has not been covered by empirical research so far.
The two main studies on cross-cultural universality of MAC values (Alfano et al., 2022; Curry, Mullins, & Whiterhouse, 2019b) were based on anthropological sources written in English. They noted this as a limitation, and thus suggested testing the MAC in various languages and cultural contexts as an avenue for further research. Russian-speaking organizations is one of such non-English cultural contexts, the application to which can provide a valuable contribution to the cross-cultural validity of the theory.
The Russophone diaspora is one of the world’s largest (Ryazantsev, 2015), which spreads beyond the post-Soviet area. In some Western countries such as the United States, Germany, and Israel, the Russophones have formed extensive well-established communities hundreds of thousands or millions strong that continue to use Russian for informal and formal business communications (Zabrodskaya & Ivanova, 2021). However, notwithstanding the topic’s importance, a literature search returns no studies on the MAC applied to the Russophone groups. Similarly, there is a scarcity of studies on ethical leadership in the Russophone cultural context. Therefore, the application of both these theories in a Russian-speaking organization was expected to offer important contributions to both fields.
Therefore, to close the gap identified above, the following research question (RQ) has been posed: Does the perceived follower–leader congruence in MAC basic moral elements causally influence the perceived level of ethical leadership in the Russophone language context?
The expected answer to the RQ depends on the results of testing for two research hypotheses. First, the association hypothesis RH1: The relationship between perceived follower–leader congruence in any of the MAC elements (Xn) and perceived level of ethical leadership (Y) is statistically significant. If confirmed, the causality hypothesis RH2 should be tested: The influence of perceived follower–leader congruence in any of the MAC elements on the perceived level of ethical leadership (Xn→Y) is causal. The conceptual model is depicted in Figure 1.
The figure depicts the conceptual framework to be tested on association and causality. Hn stand for hypotheses alternative to the null hypotheses outlined in Table 9. a Each Xn signifies perceived congruence in respective MAC elements drawn from Curry (2016). b Y stands for perceived ethical leadership construct suggested by Brown, Treviño, and Harrison (2005).
Discussing the methodological issues of causal inference in empirical research, Maxwell (2004) advocates a realist approach, which recognizes the validity of both positivist and interpretivist methods, and suggests ways of allowing for causality inference in a single-setting design, such as, for example, triangulation of quantitative and comparative qualitative data. Thus, this research was structured as a two-phase mixed-method study: the quantitative cross-sectional testing of the relationships’ significance and differences in means, and the follow-up qualitative cross-sectional inquiry into causality of findings from the first phase.
As the cross-sectional design was chosen due to the limited time frame for sampling provided by the sampled company, a survey with a subsequent quantitative analysis was conducted for testing RH1. Given the limited accuracy of possible causality inferences in a quantitative cross-sectional inquiry, a follow-up qualitative study is required for the exploration of causal relationships stated by RH2 (Gelfand, Mensinger, & Tenhave, 2009; MacKinnon, Fairchild, & Fritz, 2007). Thus, as a Phase 2, a series of semi-structured interviews were performed.
For Quantitative Phase 1, the study addressed confidentiality and anonymity by obtaining informed consent from all participants. Participants consented by checking a relevant box in the web form prior to starting answering questions. The consent was recorded and documented as per ethical guidelines. No data that could identify specific participants was collected, and the company received results only in aggregated form to ensure that individual identities were protected. The consent procedure and data handling process received prior approval from the SBS Swiss Business School Ethical Committee, in compliance with the Declaration of Helsinki’s principles. The date of approval is September 2, 2022, while the data collection was performed between 16 and 19 of September, 2022.
For Qualitative Phase 2, participants provided written consent, confirmed orally at the beginning of each interview. Data were anonymized personally by the researchers, and only anonymized data were used in the report. To ensure the company did not incur any damage and to maintain transparency, the approval of the CEO and HR Director was obtained before engaging with any employees. The results were presented to the company in an aggregated format, so that participants’ identities cannot be revealed in any circumstances. The procedures and the use of consent were reviewed and approved by the SBS Swiss Business School Ethical Committee. The date of approval is December 7, 2022, while the data collection spanned from 07 till 26 of December 2022.
No material incentive was offered or provided to any participant. All participants were informed of the study’s purpose, methods, and potential risks and benefits. They were made aware that participation was voluntary, and they could withdraw at any time without any consequences.
All data collected during this study were handled confidentially and stored securely on encrypted password-protected hard-disk of the researcher’s computer, and will be retained for five years before being destroyed.
This process ensured adherence to the principles of autonomy and respect for participants as stated in the Declaration of Helsinki.
The questionnaire was composed of items used in previous research. Two relevant instruments, presented below in detail, were drawn from the literature.
For the measurement of the seven basic moral behaviors suggested by MAC theory, Curry, Chesters, and Van Lissa (2019a) designed and validated the 42-item Morality-as-Cooperation Questionnaire (MAC-Q). Whereas MAC-Q measures perceived values themselves, the present research is aimed to measure perceived congruence on these values. Thus, the instrument’s introductory messages were modified in order to measure followers’ perceived congruence in basic moral elements with those of the leader’s, while the items themselves were left identical to those from the original questionnaire.
The literature suggests numerous approaches to measuring ethical leadership, which vary depending on the philosophical angle on ethics, focal respondents, and differences in the construct aspects that are intended to be measured (see for example Langlois, Lapointe, Valois, & de Leeuw, 2014; Kalshoven, Den Hartog, & De Hoogh, 2011; Riggio, Zhu, Reina, & Maroosis, 2010; Spangenberg & Theron, 2005; Yukl, Mahsud, Hassan, & Prussia, 2013). The majority of those instruments are in line with Brown et al.’s (2005) approach in assessing ethical leadership as a specific behavioral practice, though enhancing it in certain aspects. Meanwhile, the Ethical Leadership Scale (ELS) suggested by Brown and colleagues represents one of the most frequently used in empirical research of ethical leadership (Bedi et al., 2016). For this reason, this study considers the ELS as the most appropriate instrument for the empirical advancement of Brown et al.’s (2005) conceptualization of ethical leadership.
The questionnaire employed in the survey was therefore composed of two instruments: the ELS and the modified versions of the MAC-Q Relevance and Judgment scales, denoted as mMAC-Q(R) and mMAC-Q(J) scales in this study ( Table 2). Besides this, such demographic variables as gender, age and organizational level were collected in order to test between-group differences in responses. All non-demographic questionnaire’s items were measured using a 5-point Likert-type response range. Although numerous papers insist on better reliability of 7-point scales compared to 5-point scales (see for review Taherdoost, 2019), some others argue that the difference in reliability is insufficient, and 5-point scales are less confusing and easier to respond (Aybek & Toraman, 2022; Babakus & Mangold, 1992). Using this assumption, this research employed 5-point scale to be easier to read and understand from mobile screens by respondents, thereby helping them to avoid being overwhelmed when picking answers, which can more probably occur with a wider response scale.
Var # | Variable name | Short name | Scale | Var. label | Items |
---|---|---|---|---|---|
X1 | Perceived follower–leader congruence in “Care for kin”a | Congruence in kin care | mMAC-Q(R) | rK | rk1, rk2, rk3 |
mMAC-Q (J) | jK | jk1, jk2, jk3 | |||
X2 | Perceived follower–leader congruence in “Loyalty to the group”a | Congruence in group loyalty | mMAC-Q(R) | rG | rg1, rg2, rg3 |
mMAC-Q(J) | jG | jg1, jg2, jg3 | |||
X3 | Perceived follower–leader congruence in “Reciprocal exchange”a | Congruence in reciprocity | mMAC-Q(R) | rR | rr1, rr2, rr3 |
mMAC-Q(J) | jR | jr1, jr2, jr3 | |||
X4 | Perceived follower–leader congruence in “Display of dominance”a | Congruence in heroism | mMAC-Q(R) | rH | rh1, rh2, rh3 |
mMAC-Q(J) | jH | jh1, jh2, jh3 | |||
X5 | Perceived follower–leader congruence in “Display of deference”a | Congruence in deference | mMAC-Q(R) | rD | rd1, rd2, rd3 |
mMAC-Q(J) | jD | jd1, jd2, jd3 | |||
X6 | Perceived follower–leader congruence in “Fairness”a | Congruence in fairness | mMAC-Q(R) | rF | rf1, rf2, rf3 |
mMAC-Q(J) | jF | jf1, jf2, jf3 | |||
X7 | Perceived follower–leader congruence in “Respect to prior possession”a | Congruence in respect to property | mMAC-Q(R) | rP | rp1, rp2, rp3 |
mMAC-Q(J) | jP | jp1, jp2, jp3 | |||
Y | Perceived level of ethical leadershipb | Ethical leadership | Ethical Leadership Scale | pELS | el1, el2, el3, el4, el5, el6, el7, el8, el9, el10 |
a Names within quotation marks are drawn from Curry (2016).
b The term is drawn from Brown et al. (2005).
The composed questionnaire was translated into Russian. The correctness of the items’ translation was checked by reverse-translation involving third-party translators. Then, a review of the instrument’s form and content was carried out by running a questionnaire among a private Russian HR community in order to ensure that the overall composition of the questionnaire was easy to use, the Russian wording of the items was clear to the respondents, and there was no misperception of the constructs which were supposed to be measured. After final wording amendments, the resulting instrument was exhibited at Google Forms platform for sampling.
The survey was conducted in a Russian logistics company (hereinafter the company). The total number of employees is 1319 located in various regions of Russia, among which 263 work in the Moscow headquarters, which was chosen as a sampling site following the company’s management requirements. However, although the behavioral patterns of the regional branches’ employees theoretically may differ from those of the Moscow HQ due to possible cultural differences between different regions of Russia, this limitation was considered negligible for the generalization of the results to the whole company, and wider to the people employed in the industrial sector of the Russian economy, as the substantial part of headquarters’ employees moved to Moscow from various regions and industries, according to the company’s management statement. No respondents were deliberately chosen for the participation by management or the researcher, and the entire process was voluntary and open to any participant. To that extent, the selection was random, although the choice of the company and its headquarters utilized convenience sampling.
The emails contained an introductory message motivating employees to participate in the survey were sent to all HQ employees, and provided a link to the questionnaire’s web-form. Confidentiality and anonymity issues were addressed thoroughly, as outlined below in the Ethics and Consent section.
Fifty six responses were received (Belinskiy & Olsen, 2024). The response rate was therefore 56/263, or 21 percent. Participation was voluntary, responses were anonymous, and there were no repetitive patterns of answers on any scales. Therefore, all responses (n = 56) were considered to have been made sincerely and consciously, and thus valid. Following Fritz and MacKinnon’s (2007) recommendations, a minimal required sample size should not be less 33 for valid regression analysis, so the actual sample size (n = 56) was found sufficient.
The amendment of the original MAC-Q in introductory messages changed the constructs being measured, and hence the validity of the instrument had to be verified. The initial measurement model was tested for a model fit using SPSS Amos 29 package in order to assess whether it fitted the dataset and was good in predictive capacity. CFA initially resulted in poor fit of all models combined from the MAC-Q items ( Table 3), in contrast to good fit demonstrated by the initial MAC-Q Relevance and Judgment scales (Curry, Chesters, & Van Lissa, 2019a). After removal of lower-loading factors and factors demonstrating the most salient levels of standardized residual covariance with those of other items, the best model fit was achieved when each of seven MAC constructs had been reduced to two items from the Relevance scale ( Figure 2): All factor loadings were above .70 and all model-fit values (CMIN Sig. .156, CFI .985, TLI .975, RMSEA .059, RMSEA 90%CI [.000, .107], PCLOSE .380) are close to levels commonly accepted in the literature as good-fit measures (Schreiber, Nora, Stage, Barlow, & King, 2006; Ullman, 2006); only the upper bound of RMSEA CI .107 exceeds Schreiber et al.’s threshold .08 ( Table 4). This model (hereinafter Model A’, mMAC-Q, or the MAC section of the instrument), was employed for the following reliability and validity analysis.
Model fit criteria | Good fit reference valuesa | Model A 7 Relevance 3-item scales | Model B 7 Judgment 3-item scales | Model C 7 combined 6-item scales | Model D 14 separate 3-item scales |
---|---|---|---|---|---|
CMIN | 391 | 417 | 2386 | 1977 | |
df | 168 | 168 | 798 | 728 | |
χ2/df | ≤2 | 2.327 | 2.481 | 2.990 | 2.716 |
CMIN Sig. | >0.05 | .000 | .000 | .000 | .000 |
CFI | ≥.95 | .813 | .712 | .459 | .574 |
TLI rho2 | ≥.95 | .767 | .640 | .416 | .496 |
RMSEA | <0.06 | .155 | .164 | .190 | .177 |
RMSEA low 90% CI | .000 | .135 | .144 | .181 | .167 |
RMSEA high 90% CI | <0.08 | .175 | .184 | .199 | .186 |
RMSEA Sig. (PCLOSE) | >0.05 | .000 | .000 | .000 | .000 |
a Good fit reference values drawn from Schreiber et al. (2006), and Ullman (2006).
The figure represents the structural equation model for the best fitting model A’. It is generated using IBM SPSS Amos 29 software when performing CFA. Circles “X1” – “X7” represent latent variables standing for basic moral behaviors (Curry, 2016). Squares like “rk3” represent observed variables correspondent to the respective questionnaire items ( Table 2). Ellipses denoted like “e3” represent errors, or the variance in the observed variable that is not explained by the latent variable.
The single-headed arrows indicate regression weights or factor loadings from one variable to another. The double-headed arrows represent correlations or covariances between latent variables. The numbers next to these arrows show the correlation values between different latent variables.
Model fit criteria | Reference valuesa | Model A’ 2-item R-scales | Model B’ 2-item J-scales | Model C’ 2-item mixed |
---|---|---|---|---|
CMIN (χ2) | 66.65 | 105.76 | 72.07 | |
df | 56 | 56 | 56 | |
χ2/df | ≤2 | 1.190 | 1.889 | 1.287 |
CMIN significance | >0.05 | .156 | .000 | .073 |
CFI | ≥.95 | .985 | .903 | .976 |
TLI rho2 | ≥.95 | .975 | .842 | .961 |
RMSEA | <0.06 | .059 | .127 | .072 |
RMSEA low 90% CI | .000 | .000 | .089 | .000 |
RMSEA high 90% CI | <0.08 | .107 | .164 | .117 |
RMSEA Sig. (PCLOSE) | >0.05 | .380 | .001 | .232 |
a Reference values drawn from Schreiber et al. (2006), and Ullman (2006).
Each scale showed a Cronbach’s alpha within a range between .83 and .90, which represents a good level of internal consistency according to Taber’s (2018) review. The composite reliability value for each scale ranged from .83 to .91, exceeding the .70 threshold suggested by Hair, Ortinau, and Harrison (2010). Therefore, each construct in the reduced model was found reliable.
Pairwise correlations within each construct are positive and statistically significant (r > .70, p < .001) for all Xn1–Xn2 within-construct pairs ( Table 5). Each scale’s average variance-extracted value (AVE) exceeds the threshold .50 ( Table 6) suggested by Fornell and Larcker (1981). Thus, the scales of the reduced instrument demonstrate sufficient convergent validity.
Item | rk2 | rk3 | rg2 | rg3 | rr1 | rr3 | rh2 | rh3 | rd1 | rd3 | rf1 | rf2 | rp1 | rp2 | |
---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|
rk2 | r | 1 | .74a | .72 | .55 | .66 | .66 | .63 | .66 | .60 | .68 | .33 | .38 | .30 | .27 |
Sig. | .000 | .000 | .000 | .000 | .000 | .000 | .000 | .000 | .000 | .013 | .004 | .027 | .045 | ||
rk3 | r | .74a | 1 | .68 | .48 | .63 | .70 | .59 | .60 | .63 | .64 | .36 | .30 | .29 | .20 |
Sig. | .000 | .000 | .000 | .000 | .000 | .000 | .000 | .000 | .000 | .006 | .024 | .028 | .145 | ||
rg2 | r | .72 | .68 | 1 | .71a | .87 | .85 | .71 | .71 | .71 | .70 | .26 | .33 | .15 | .11 |
Sig. | .000 | .000 | .000 | .000 | .000 | .000 | .000 | .000 | .000 | .051 | .012 | .267 | .404 | ||
rg3 | r | .55 | .48 | .71a | 1 | .77 | .75 | .69 | .58 | .57 | .50 | .18 | .27 | .27 | .19 |
Sig. | .000 | .000 | .000 | .000 | .000 | .000 | .000 | .000 | .000 | .178 | .043 | .048 | .165 | ||
rr1 | r | .66 | .63 | .87 | .77 | 1 | .80a | .68 | .63 | .68 | .64 | .27 | .31 | .10 | .18 |
Sig. | .000 | .000 | .000 | .000 | .000 | .000 | .000 | .000 | .000 | .046 | .019 | .476 | .186 | ||
rr3 | r | .66 | .70 | .85 | .75 | .80a | 1 | .71 | .71 | .73 | .67 | .32 | .35 | .24 | .22 |
Sig. | .000 | .000 | .000 | .000 | .000 | .000 | .000 | .000 | .000 | .015 | .009 | .075 | .099 | ||
rh2 | r | .63 | .59 | .71 | .69 | .68 | .71 | 1 | .78a | .64 | .65 | .51 | .49 | .43 | .48 |
Sig. | .000 | .000 | .000 | .000 | .000 | .000 | .000 | .000 | .000 | .000 | .000 | .001 | .000 | ||
rh3 | r | .66 | .60 | .71 | .58 | .63 | .71 | .78a | 1 | .64 | .62 | .44 | .41 | .27 | .32 |
Sig. | .000 | .000 | .000 | .000 | .000 | .000 | .000 | .000 | .000 | .001 | .002 | .047 | .017 | ||
rd1 | r | .60 | .63 | .71 | .57 | .68 | .73 | .64 | .64 | 1 | .76a | .36 | .38 | .28 | .31 |
Sig. | .000 | .000 | .000 | .000 | .000 | .000 | .000 | .000 | .000 | .007 | .004 | .039 | .019 | ||
rd3 | r | .68 | .64 | .70 | .50 | .64 | .67 | .65 | .62 | .76a | 1 | .43 | .38 | .29 | .33 |
Sig. | .000 | .000 | .000 | .000 | .000 | .000 | .000 | .000 | .000 | .001 | .004 | .029 | .013 | ||
rf1 | r | .33 | .36 | .26 | .18 | .27 | .32 | .51 | .44 | .36 | .43 | 1 | .82a | .62 | .83 |
Sig. | .013 | .006 | .051 | .178 | .046 | .015 | .000 | .001 | .007 | .001 | .000 | .000 | .000 | ||
rf2 | r | .38 | .30 | .33 | .27 | .31 | .35 | .49 | .41 | .38 | .38 | .82a | 1 | .56 | .69 |
Sig. | .004 | .024 | .012 | .043 | .019 | .009 | .000 | .002 | .004 | .004 | .000 | .000 | .000 | ||
rp1 | r | .30 | .29 | .15 | .27 | .10 | .24 | .43 | .27 | .28 | .29 | .62 | .56 | 1 | .74a |
Sig. | .027 | .028 | .267 | .048 | .476 | .075 | .001 | .047 | .039 | .029 | .000 | .000 | .000 | ||
rp2 | r | .27 | .20 | .11 | .19 | .18 | .22 | .48 | .32 | .31 | .33 | .83 | .69 | .74a | 1 |
Sig. | .045 | .145 | .404 | .165 | .186 | .099 | .000 | .017 | .019 | .013 | .000 | .000 | .000 |
rK | rG | rR | rH | rD | rF | rP | |
---|---|---|---|---|---|---|---|
AVE | .74 | .72 | .80 | .78 | .76 | .83 | .79 |
Sqrt | .86 | .85 | .90 | .89 | .87 | .91 | .89 |
The discriminant validity was assessed according to Fornell and Larcker’s (1981) criterion, implying that the square root of the scale’s AVE should exceed its correlation with other items in the model. This criterion was met for all items except the second item in “loyalty to the group” scale, which returned slightly negative differences between AVE square root and correlations with first and second items in “reciprocity” scale: -.02 and -.01 respectively ( Table 7).
X1:rK | X2:rG | X3:rR | X4:rH | X5:rD | X6:rF | |
---|---|---|---|---|---|---|
X1:rK | ||||||
X2:rG | .98 | |||||
X3:rR | 1.00 | 1.23 | ||||
X4:rH | .96 | 1.04 | 1.05 | |||
X5:rD | 1.00 | .97 | 1.06 | .99 | ||
X6:rF | .51a | .39a | .47a | .70a | .58a | |
X7:rP | .42a | .28a | .29a | .59a | .48a | 1.07 |
Therefore, the modified MAC-Q Relevance scale reduced to two items per each MAC-related construct, or mMAC-Q, was considered a valid instrument for measuring follower–leader perceived congruence in MAC elements. All data related to skipped or removed items was not taken into account in further analysis. Therefore, while the unchanged MAC-Q section of the instrument failed to demonstrate good model fit, the modified mMAC-Q section represents a valid and coherent model for measuring perceived congruence in MAC elements in Russian.
Similarly to the mMAC-Q section, the ELS scale demonstrated a high rate of internal consistency (α = .92). The EFA and CFA analyses confirmed Brown et al.’s (2005) argument that perceived ethical leadership measured by the 10-item ELS represents a coherent construct. The EFA returned a one-factor model with an eigenvalue of 5.96, so that 59.6 percent of variance is explained by one factor with the factor loading ranged from .483 to .894 per item. The CFA returned good model fit for the scale (CMIN Sig .326, CFI .990, TLI .987, RMSEA .041, RMSEA 90%CI [.000, .108], PCLOSE .538); hence, the original Brown et al.’s (2005) ELS was employed for further inferential analysis as a uni-dimensional measure of the Y variable.
The content validity of the Russian-worded items was qualitatively addressed in Phase 2, resulting in support for claims of Curry, Chesters, and Van Lissa (2019a) and Brown et al. (2005). Hence, the instrument is viewed as a valid instrument to measure the perceived level of ethical leadership in Russian-language contexts.
Among 56 responses, 34 participants were female and 22 were male. All the participants were adults, eight of them between 18 and 30 years old, 40 between 31 and 45, and the remaining eight within the 46 and 60 age range. A total of 18 responses were collected from those who received link one, 14 followed link two, and 24 received link three, which implies each response allocated to the corresponding organizational layer ( Table 8).
Raw Likert integer values were transformed to continuous values by calculating the mean of the array of values corresponding to each scale of the instrument, thus forming the data-set for the statistical tests. Initial responses were coded as follows: strictly negative responses were coded as 1, strictly positive as 5, with a neutral point at 3. The means above 3, therefore, demonstrate the evidence of the perceived congruence in MAC elements in the case of independent variables, and perceived ethical leadership in the case of dependent one. The results demonstrate the perception of all MAC elements as congruent with the leader (3.31 < M < 3.89), and perception of the leadership as ethical (M = 3.86). The data were considered as normally distributed with skewness [-.10, .53] and excess kurtosis [-1.35, .75] intervals close to those proposed by Gravetter, Wallnau, Forzano, and Witnauer (2020) for normal distributions ( Table 8).
Seven null hypotheses were posed for the quantitative test ( Table 9). The results demonstrated that all means of follower–leader perceived congruence, in each MAC element, correlate with perceived ethical leadership, with Pearson r(54) values ranging from .34 to .59 and significance level at p < .001, except those for “fairness” and “property” scales, which are significant at p < .05 and < .01 respectively ( Table 10). Regression analyses stress the argument for the rejection of all null hypotheses ( Table 11). Thus, the results were considered to show a statistically significant moderate positive linear correlation for each Xn–Y pair using Ratner’s (2009) criteria, which supports the RH1, meaning that there is a statistically significant relationship between perceived follower–leader congruence in each MAC element and the perceived level of ethical leadership ( Table 22).
It was expected that this association may be demographic sensitive, which means that age, gender, or organizational level may interfere between congruence in some MAC values and the perceived level of ethical leadership. To control for demographics, the difference in means between various population groups was assessed. However, tests for the equality of means between gender ( Table 12) and age ( Table 14) groups failed to demonstrate statistically significant differences between means of these groups ( Table 13; Table 15).
In comparison, a one-way ANOVA demonstrated that the difference in means between the top-management team (n = 18), middle management (n = 14), and lower-level staff (n = 24) is statistically significant for any variable at p < .001 ( Table 16; Table 17). A post-hoc pairwise Tukey HSD test showed that each variable demonstrated difference in means between top-management and two other organizational levels at least at p < .01 in all cases, except the difference in congruence in “fairness” significant at p = .021 ( Table 18). Meanwhile, neither comparison of variable means between mid- and lower-layers provided sufficient evidence of such a difference (p > 0.05). The mean differences between top-management and two other groups were all positive and fell into the range between .61 and 1.29, which signifies that each variable demonstrated a higher mean rate when assessed by top managers than when assessed by employees from other organizational layers.
Var. | Group (I) | Group (J) | Mean diff. (I-J) | Std. Error | Sig. | 95% Confidence Interval | |
---|---|---|---|---|---|---|---|
Lower bound | Upper bound | ||||||
pELS | Top | Mid | .74* | .16 | .000 | .37 | 1.12 |
Low | .81* | .14 | .000 | .48 | 1.14 | ||
Mid | Top | -.74* | .16 | .000 | -1.12 | -.37 | |
Low | .07 | .15 | .895 | -.29 | .42 | ||
Low | Top | -.81* | .14 | .000 | -1.14 | -.48 | |
Mid | -.07 | .15 | .895 | -.42 | .29 | ||
rK | Top | Mid | 1.15* | .24 | .000 | .58 | 1.73 |
Low | .75* | .21 | .002 | .25 | 1.25 | ||
Mid | Top | -1.15* | .24 | .000 | -1.73 | -.58 | |
Low | -.40 | .22 | .179 | -.95 | .14 | ||
Low | Top | -.75* | .21 | .002 | -1.25 | -.25 | |
Mid | .40 | .22 | .179 | -.14 | .95 | ||
rG | Top | Mid | 1.08* | .23 | .000 | .53 | 1.63 |
Low | .81* | .20 | .000 | .32 | 1.29 | ||
Mid | Top | -1.08* | .23 | .000 | -1.63 | -.53 | |
Low | -.27 | .21 | .416 | -.79 | .24 | ||
Low | Top | -.81* | .20 | .000 | -1.29 | -.32 | |
Mid | .27 | .21 | .416 | -.24 | .79 | ||
rR | Top | Mid | 1.26* | .23 | .000 | .70 | 1.82 |
Low | .88* | .20 | .000 | .39 | 1.36 | ||
Mid | Top | -1.26* | .23 | .000 | -1.82 | -.70 | |
Low | -.39 | .22 | .189 | -.91 | .14 | ||
Low | Top | -.88* | .20 | .000 | -1.36 | -.39 | |
Mid | .39 | .22 | .189 | -.14 | .91 | ||
rH | Top | Mid | 1.23* | .20 | .000 | .76 | 1.70 |
Low | .97* | .17 | .000 | .55 | 1.38 | ||
Mid | Top | -1.23* | .20 | .000 | -1.70 | -.76 | |
Low | -.26 | .19 | .332 | -.71 | .18 | ||
Low | Top | -.97* | .17 | .000 | -1.38 | -.55 | |
Mid | .26 | .19 | .332 | -.18 | .71 | ||
rD | Top | Mid | .97* | .22 | .000 | .44 | 1.50 |
Low | .74* | .19 | .001 | .27 | 1.20 | ||
Mid | Top | -.97* | .22 | .000 | -1.50 | -.44 | |
Low | -.23 | .21 | .506 | -.73 | .27 | ||
Low | Top | -.74* | .19 | .001 | -1.20 | -.27 | |
Mid | .23 | .21 | .506 | -.27 | .73 | ||
rF | Top | Mid | 1.15* | .25 | .000 | .54 | 1.76 |
Low | .61* | .22 | .021 | .08 | 1.15 | ||
Mid | Top | -1.15* | .25 | .000 | -1.76 | -.54 | |
Low | -.54 | .24 | .073 | -1.11 | .04 | ||
Low | Top | -.61* | .22 | .021 | -1.15 | -.08 | |
Mid | .54 | .24 | .073 | -.04 | 1.11 | ||
rP | Top | Mid | 1.29* | .32 | .001 | .51 | 2.07 |
Low | .94* | .28 | .004 | .26 | 1.62 | ||
Mid | Top | -1.29* | .32 | .001 | -2.07 | -.51 | |
Low | -.350 | .30 | .498 | -1.08 | .39 | ||
Low | Top | -.94* | .28 | .004 | -1.62 | -.26 | |
Mid | .35 | .30 | .498 | -.39 | 1.08 |
This was assumed as a sign of possible moderation by the organizational layer parameter, and tested by a follow-up multiple regression ( Table 19). Baron and Kenny (1986) recommended comparing the unstandardized B for two scenarios: with and without the influence of a potential moderator. The test showed that adding an organizational layer as a moderator lowers levels of B in all seven Xn–Y relationships. Hence, the data quantitatively supported a moderating effect of the organizational layer. However, as shown below, the results from Phase 2 demonstrate that it is not the actual moderator, thought the moderation effect remains valid with another parameter.
The positive correlation itself does not entail the causality of Xn→Y. To be supported, the relationship should fulfill three conditions, namely (1) a significant association between variables, (2) a directional or temporal precedence of variance in Xn to variance in Y, and (3) an unconfoundedness of this relationship (Gelfand et al., 2009; Kenny, 1979). While the first condition is met, as reported above, the second is impossible to establish quantitatively in a cross-sectional design (Gelfand et al., 2009; MacKinnon et al., 2007), so it has been examined qualitatively in Phase 2, as recommended by Maxwell (2004). The third condition, unconfoundedness, appears to be the most questionable, as shown below.
When planning the research, it was hypothesized (RH2) that each perceived congruence in MAC value causally impacts the perceived level of ethical leadership, and no possible interplay or interdependence of independent variables was assumed. The data show that all Xn–Xm pairwise correlations are significant and positive (.21 < r < .92, n = 56, p < .05), except for the pairs “property”–“group” and “property”–“reciprocity” which correlate at an insignificant p > .05 ( Table 20). This may signal that some XM variables might confound or mediate relationships of other Xn variables with Y.
According to Kenny (1979), the unconfoundedness of the Xn–Y association implies that there is no controlling variable that turns their correlation insufficient or insignificant. Partial correlation test of Xn–Y associations controlled by any other XC variable returned insignificant (p > .05) partial correlations for 18 Xn–Y pairs out of 42 possible ( Table 21). The congruence in “kin”, “group”, “reciprocity” and “heroism” impacts some other constructs’ relationships with “perceived EL”, so that the unconfoundedness condition in this respect may not be fulfilled, hence questioning RH2. However, the quantitative data without theoretical explanation are not sufficient to infer a confounding influence (Pieters, 2017). Therefore, the qualitative inquiry was expected to reveal these variables’ capacity to determine the importance of congruence in other values for the perception of a leader as ethical, which is reported below.
The qualitative Phase 2 was intentionally limited in scope and focused only on themes which could confirm or reject the quantitative findings of the preceding survey. It aimed to triangulate the quantitative phase results by revealing qualitative evidence supporting or questioning RH2 to test whether two remaining conditions of causality are fulfilled, namely directionality and unconfoundedness. In addition, Phase 2 was aimed at verifying the Phase 1 assumption on the content validity of the Russian wording of the final instrument items.
According to Maxwell (2004), a comparative multicase design represents a qualitative strategy allowing for causality testing of relationships. He argues that such an approach addresses concerns related to qualitative case studies inability to infer what would have happened if a presumed cause was not present (Cook, Campbell, & Shadish, 2002). Comparing cases here was expected to provide various perspectives on predetermined themes, so as to smooth the impact of one personal opinion on central tendency estimation. Hence, this phase represents the deductive case study, which is explanatory in Yin’ (2013) terminology, or instrumental in the Stake’s (1995) one, conducted in a multiple-case design for examination of causality following Maxwell’s (2004) recommendations, meaning by multiple cases the interviews made in a single organizational context.
Semi-structured interviews were chosen as a methodology for Phase 2 to ensure the collection of comparable statements or judgments by setting particular themes for discussion. The propositions, a term suggested by Yin (2013), are based on the MAC (Curry, 2016) and ethical leadership (Brown et al., 2005) theories, as well as the findings of Phase 1. The following questions were therefore posed for qualitative inquiry. First, have the relevant phenomena been properly measured by the items employed by Phase 1? The answer requires a qualitative examination of the relevance of the Russian wording of the mMAC-Q and the ELS sections’ items to the domain of perceived leaders’ morality, or the content validity of the questionnaire’s items. Second, are these constructs and the conceptual model relevant to the organizational context of the company? This implies examining whether the congruence in MAC values influences and precedes the perceived ethicality of the leader, how some MAC values interfere with others in their influence on the latter, and whether such an influence is unilateral and causal.
The literature suggests that the most common principle for defining the adequate sample size is saturation, which represents the point when the additional data reveals no new properties or insights about a construct examined (Guest, Bunce, & Johnson, 2006). Hennink and Kaiser (2022) show that, in recent empirical studies, the saturation was typically achieved within a range from 9 to 17 interviews, with a mean at 12–13 interviews, if the population was homogeneous, and objects were narrowly defined. Since Phase 2 involved sampling from a culturally homogeneous one-site population, and examined relationships between constructs that were narrowly defined by previous research, the sample size was arbitrarily set at 14 interviews. Applying the Hennink, Kaiser, and Marconi’s (2017) criteria, the code saturation was achieved after coding the first three interviews, while the meaning saturation was reached after the coding of the 10th interview, when no new meanings, nuances, or insights were found by analyzing the remaining transcripts. Therefore, 14 interviews were eventually considered the adequate sample size, which provided the necessary richness and depth of the qualitative data.
Candidates for interviews were selected in collaboration with the company’s HR director based on the following criteria: they worked in the HQ of the company, they did not participate in the Phase 1 survey, and they demonstrated an interest in participation. The HR director had a series of phone conversations with potential participants until seven top-level and seven lower-level candidates had agreed to participate ( Table 23).
For the avoidance of researcher bias, a professional Russian-speaking interviewer was invited to perform all interviews. Her purpose was to motivate interviewees to talk freely about themes taken from the modified instrument and provide justifications or rationalizations for their judgments and evaluations. In addition to this, she was expected to discuss with the participants what would happen to their perception of a leader if a similarity in moral values were to change. Practically, interviewees were asked to express their thoughts about the items used in the preceding survey.
There were six female and eight male participants. All conversations were conducted and transcribed verbatim in Russian, which was the mother tongue of the majority of the interviewees. Those two persons who noted that their mother tongue was not Russian, spoke Russian with nativelike proficiency. The series of 14 interviews resulted in 1053 minutes of records, with each interview’s length lying in the range [57 min, 95 min], mean 75 minutes ( Table 23). Transcript files exceeded 120,000 words in total volume.
The set of verified transcripts was analyzed applying Braun and Clarke’s (2006) six-phase framework for thematic analysis using the Nvivo 1.7 package. The data and propositions were linked by the markers of assertion and explanation. When a phenomenon was named relevant or important, it was coded correspondingly. When the explanation of that importance indicated other markers of the importance, they were also linked to the respective proposition as potentially influencing the phenomenon they were used to explain. The main criterion of the relevance of evidence found in the data was frequency of display across numerous cases. When some aspect was found as coded episodically, it was eventually neglected; only assertion or explanation patterns that were widely repeated in numerous interviews were considered to reflect a tendency in the given population.
After the initial coding, the codes were collated into themes, a term which here denotes data units for further interpretation of the aspects of phenomena under examination (Boyatzis, 1998). The final iteration of review of transcripts, codes, and themes and elimination of the episodically mentioned codes and themes resulted in a final theme list. It needs to be noted that the coding was performed by the researcher, so the consideration of results of Phase 2 should keep in mind potential researcher bias, though the data analysis has been attempted in good faith, without conscious intention to confirm the findings of the previous phase.
Ethical issues were addressed adhering to the Declaration of Helsinki principles, as outlined in detail in Ethics and Consent section below.
Each interview in Phase 2 included discussion about every item of the ELS in order to confirm the validity of Brown et al.’s (2005) interpretation of the ethical leadership concept for the company, which operates in a Russian cultural context within a specific industrial subculture. The way the content was coded resulted in numerous quotes referencing the ELS items’ relevance ( Table 24). All interviewees expressed no doubt when evaluating their CEO as an ethical leader, and that their judgments were unequivocal and unconditional. Despite the fact that four out of 10 items of ELS were considered as irrelevant by some interviewees, the dominant sentiment is that all aspects are important for such an evaluation. Hence, the present study confirms the content validity of the Russian-language ELS instrument when applied to the sample company.
There was no difference found between top- and lower-level interviewees in evaluation of their CEO as an ethical leader. All of them unequivocally assessed the leader as ethical, and all expressed sympathy and general follower-leader congruence in moral values. At the same time, the majority of participants noted a difference between top managers and other employees in evaluation of the leader. This is consistent with the moderating role of the organizational layer parameter revealed in Phase 1, though the data demonstrate its derivative nature.
In responses to the reasoning request, interviewees explained the difference with reference to a proximity to the leader. At the same time, they repeatedly pointed out that proximity to the leader is not sufficient when communication is lacking. Communication frequency was noted by the interviewees as an even more important parameter influencing the evaluation of the leader as ethical. Direct communication was mentioned as an instrument for shaping followers’ perception, and thus one of the main constraints in big companies. Hence, communication deficit turns out to be the main reason for the difference between organizational levels in perceived level of ethical leadership. When communication is sufficient, the difference may disappear. None of the interviews pointed to a different reason.
Therefore, although the perceived level of ethical leadership depends on the organizational level of evaluators in the focal population, the moderating role of the organizational layer parameter revealed in Phase 1 is found to be company specific, and not generalizable to a wider population. Instead, the leader’s proximity, that is behavioral visibility and cumulative communication duration with followers, emerge as factors influencing the perceived level of ethical leadership.
Similarly, there is a difference in moral value congruence between different organizational levels and the leader. The proximity and communication were also noted as main reasons. On the other hand, inability to understand the leader was pointed out as a reason for the difference in value congruence. Hence, the farther the hierarchical distance between the leader and his actual or potential followers, the less their capacity for interaction, and, thus, the bigger the gap in congruence.
While the data set contains multiple indication of the perceived link between moral value congruence and the perception of ethical leadership, the exact mention of directionality is essential for the establishment or refutation of a causality hypothesis, namely the explicit indication that change in perceived value congruence precedes the change in the perceived level of ethical leadership or, vice versa—the mention of the influence of ethical leadership on moral value congruence.
Although there is some mentioning of a congruence itself as a factor influencing ethical leadership, the main body of evidence is centered around the relevance of MAC moral behaviors for the assessment of a leader as ethical. The interviewees were responding to questions about the importance of these moral behaviors in such an evaluation, and, when justifying their opinions, they were predominantly talking about what is important for them personally, which was treated as possessing focal value or indicating appreciation of the corresponding behavior. Therefore, the relevance discussion was understood as a discussion about implied value congruence between participants and either an ideal leader of their dreams or the actual leader of the company.
All MAC basic behaviors were recognized as important for the evaluation of a leader as ethical ( Table 25), though with differences in some nuances.
Care for kin (“kin”, X1) was mainly pointed out as important for a leader mostly in the context of a group. Interviewees perceived such care as proof that their expectations regarding the leader’s care for a group would also be fulfilled. Although some point to a low significance of care for kin for leadership, such a view is rarely represented in the dataset.
The importance of loyalty to the group and care for its members (“group”, X2) is supported unanimously. All reasoning going beyond the intrinsic importance of a group itself is centered on performance indicators and factors leading to a better productivity. Uniting people within a group is seen as a prerequisite for productive work. This is consistent with group dynamics theorists positing that group cannot exist without cohesiveness (Dion, 2000). Hence, care for a group turns out to be a core value, as it directly impacts the group’s well-being, while, on the contrary, lack of care entails lack of confidence.
Such a confidence is a main rationale expressed by the interviewees when discussing the importance of a leader’s being reciprocal (“reciprocity”, X3), which is perceived in terms of predictability. Reciprocal behavior is noted to be a facilitator of performance, and, thus, all participants are unanimous in feeling that the promises should be fulfilled. Hence, the consistency in reciprocity is stated as a basis of trust. Some perceive a leader with trust in advance, but a low level of reciprocity, when observed frequently, leads to the decrease in overall trust level.
Hawkish (“heroism”, X4) and dove-ish (“deference”, X5) behaviors are mentioned by the MAC as a conflict-avoidance strategy and are both regarded as valuable in the leadership context. Being assertive and bold was regarded as a positive trait by all interviewees, as they recognize a leader’s responsibility to drive the company. It was noted that such behavior is inevitable if a leader aims to keep the big organization manageable. Participant 8 placed the importance of the ability to make hard decisions in the context of reciprocity, as making them entails making alike decisions by team members. Domination has also been mentioned as a productive behavioral feature in terms of day-to-day management.
Deference was a regarded feature as well, however. Talking about a leader in a dove-ish role didn’t induce a wide emotional response and willingness to talk about it extensively. Meanwhile, some noted that leaders should respect superiors if expecting the same attitude from followers towards them. At the same time, such respect to superiors should be deserved by them, and may take a form of value congruence.
The leader’s unfair behavior (“fairness”, X6) in sharing resources was unanimously noted as morally bad. Favoritism, as a special case of unfair conduct, was sharply condemned by all interviewees. In contrast, the avoidance of favoritism induced a positive followers’ reaction. All unanimously supported the recognition of productive efforts, regardless of the leader’s personal sympathies. The recognition of somebody’s contribution to the result by a leader was perceived as fair and good, whereas unfair behavior was noted as negatively impacting trust.
The leader’s respect to prior possession (“property”, X7), was regarded as morally good by all interviewees. The majority couldn’t find a rational justification for this, and stressed emotionally that taking what does not belong to you is absolutely wrong. Some perceived respect to property as a matter of personal respect.
Such a permanent noting of terms ‘respect’ and ‘trust’ requires attention, as these terms are not an explicit part of basic moral values as they suggested by MAC. The word ‘respect’ was mentioned numerously by every interviewee, or 186 times overall across the entire dataset. Besides the above-noted respect for a person as associated with respect to property (Participant 6), the term is related to respect to the position in hierarchy as recognition of the right to behave in a certain way (Participant 1), respect as admission of personal boundaries of another person (Participant 12), respect to the right of others for having personal opinion (Participant 14). Participant 4 placed respect in a context of reciprocal exchange in a straightforward way:
… this is respect … ; we agreed, and that means we have a contract. You owe me then. And I owe you—to protect you, to defend your rights, etc. But this is a contract, you have to abide by it, you have to do certain things in time with a required quality (Participant 4).
Regardless the context, respect emerges as a derivative of the recognition of the right of somebody else to act in a certain way or to possess something, which in turn comes from the reciprocal attitude that implies behaving with others the way that persons expect others to behave towards them. Therefore, respect has not been considered as an independent construct within the theoretical framework of the present research, as it can be reduced to Curry’s (2016) MAC basic reciprocal behavior.
The word “trust” in various forms was used in the interviews over 300 times, which highlights its importance for the participants. All interviewees agree that ethical leaders can be trusted, and that trust is essential for a leader. Trust was prevalently noted by the interviewees as a product of compliance with ethics. That is, living one’s personal life in an ethical manner was pointed out as an element of a self-sacrifice, and some interviewees perceived it as an important cue of the leader’s readiness to take a similar attitude towards a group, as such consistency in behavior is treated as a prediction for the same behavior in the future.
In all cases trust and confidence were mentioned in the context of behavioral predictability, which is perceived as important for all aspects of social life. Such a feature is consistent with recent literature, which considers the value of moral behavior in evolutionary terms as predictable, and thus saving the organism’s metabolic cost of adaptation in future cooperation (Theriault, Young, & Barrett Feldman, 2021). Given the variety of possible moral behaviors that may be expected in future, trust appears rather as a composite phenomenon which can be explained as a combination of basic moral elements within Curry et al.’s (2021) combinatorial hypothesis. This was frequently noted in interviews, such as “something so systemic and big, but consisting of such small [things]” (Participant 5).
All interviewees point to the employees’ congruence in values with the company’s leader. Such a congruence is mentioned in various contexts or ways 45 times as a prerequisite for the evaluation of a leader as ethical ( Table 26). Meanwhile, the dominant participants’ sentiment is unwillingness to work together with a leader if there is no congruence in values, and expectation of decrease in the leader’s perceived ethicality if the congruence in values disappears.
Hence, the directionality of value congruence influence on perceived level of ethical leadership should be considered established. However, there is evidence of the opposite way of influence, though opinions on that way are controversial. Whereas nine participants asserted that their values cannot change as a response to an ethical leader’s behavior, 10 interviews showed that in certain circumstances such an influence is possible ( Table 26). Therefore, the congruence and perceived level of ethical leadership represent rather a positive feedback loop, than just a uni-dimensional influence of value congruence on evaluation of a leader, though the latter was mentioned substantially more frequently.
Although the directionality hypothesis was supported, and the overall evaluation of a leader as ethical was noted as composite, “they are made up of little things” (Participant 2), the impact of different basic moral values varied. Some moral values turned out to be unimportant for the evaluation, in case the congruence in others is not felt or disappears, representing an asymmetry in perceived and assessed importance of values for the evaluation of a leader. When explaining why one or another MAC moral behavior is important for an ethical leader, most answers eventually reduced that importance to group interests or reciprocity ( Table 27). Other MAC values are either not mentioned in this context, or are mentioned episodically throughout the entire dataset, and thus neglected.
The importance of congruence on a certain value in context of other values may represent a mediation relationship. If the value congruence Xn influences “perceived EL” stronger when the mediating congruence XM is present, that may entail substantial implications for practitioners, as it represents an evident reason to concentrate on the mediator in daily practice. This does not refute RH2, which posits the causal influence of congruence in each MAC value on the perceived ethicality of a leader, though such an influence in some cases appears to be indirect. The results of mediation analysis are provided in the next section.
As reported above, the results from Phase 2 showed the interviewees’ unanimous unequivocal support for the influence of congruence in MAC basic moral values on the perception of a leader as ethical, though a positive feedback loop was also present in some extent. This means that the second condition of causal inference—directionality—is fulfilled. The third condition—unconfoundedness—is fulfilled fully for the influence of congruence in “group”, “reciprocity”, “fairness” and “property”, and partially for congruence in “kin”, “heroism”, and “deference”, on the perceived level of ethical leadership, as outlined below.
Although both research phases demonstrate interdependence between congruence in some MAC values in their influence on the perceived level of ethical leadership, the results of the qualitative phase demonstrate no signs of conflation of the MAC values by the participants. Despite the statistically significant positive pairwise correlations between the MAC variables for some pairs, Phase 2 demonstrates their distinctness and relevance to the moral domain, which is consistent with recent research on the cross-cultural universality of the MAC values (Alfano et al., 2022; Curry, Mullins, & Whiterhouse, 2019b). Partial correlation analysis demonstrates that any of the MAC variables can interfere with the association between some other MAC variables and the evaluation of a leader as ethical. Therefore, given that the directionality condition has been considered established, mediation analysis was employed to identify whether there are MAC variables causally mediating the influence of other MAC variables.
Statistical simple mediation analysis (MacKinnon et al., 2007; Preacher, Rucker, & Hayes, 2007) of the data gathered in Phase 1 suggests that 16 out of 42 possible combinations of MAC values in Xn→XM→Y relationships may be mediating, where XM stands for the MAC variable that mediated the influence of Xn MAC variable on the perceived level of ethical leadership ( Table 28). The data collected by Phase 2 provides sufficient qualitative evidence supporting only six of these potential mediations. Hence, the triangulation of both studies led to the inference on the mediating role of the congruence in “group” and “reciprocity” MAC values in the influence of congruence in “care for kin”, “heroism” and “deference” on the perceived level of ethical leadership ( Table 29). In the meantime, the influence of congruence in “fairness” and “property” MAC values was considered as direct.
IV | Mediating variable | DV | ||||||
---|---|---|---|---|---|---|---|---|
rK | rG | rR | rH | rD | rF | rP | ||
rK | partial | full | partial | pELS | ||||
rG | partial | pELS | ||||||
rR | full | pELS | ||||||
rH | full | full | full | pELS | ||||
rD | full | full | full | full | partial | pELS | ||
rF | pELS | |||||||
rP | partial | full | partial | pELS |
Therefore, the three conditions for causal inference regarding the influence of congruence in MAC values and the perceived level of ethical leadership are fulfilled for “group”, “reciprocity”, “fairness” and “property” MAC values, and their influence is thus considered as direct and causal. The influence of the remaining three MAC values, namely “care for kin”, “heroism”, and “deference” is indirect and causally mediated by “group” and “reciprocity”. This inference resulted in the amended path model depicted by Figure 3.
The figure depicts the resulting causal influence model, based on the research findings. Dashed lines denote mediated relationships. Each Xn signifies perceived congruence in respective MAC basic moral value (Curry, 2016). Y stands for perceived ethical leadership (Brown et al., 2005). The moderating role of the leader’s behavior visibility to the followers and their cumulative communication duration is also depicted.
For the population of both research phases, demonstration of fairness and respect to prior possession directly influences the perceived level of ethical leadership. The display of loyalty to the group and reciprocal behavior has the same effect, but they also play an additional role as mediators in influence of congruence in care for kin, heroism and deference on the followers’ moral evaluation of the leader. Being the way in which stimulus leads to response (MacKinnon et al., 2007), mediation in this case implies reasoning behind the hierarchy of importance of the MAC values.
That is, leader’s care for kin is important because it signals to followers that their leader will similarly take care of the group. Being assertive and sometimes dominant within the group is important because it leads to better group performance. Being forceful outwards is important when it protects group members and aims to enhance the group’s prosperity. The leader’s deference to superiors is important because it brings stability to the group’s future. Without referencing the group’s interests, all these three MAC values lose their importance for ethical leader evaluation.
In a similar vein, reciprocity impacts the influence of congruence in care for kin, heroism and deference on the leaders’ perceived ethicality. Reciprocity is referred to as respect in the interviews, where it was frequently mentioned in a context implying the leader’s recognition of a deserved right to expect or receive or possess something. This perceived deserved right means the expectation of reciprocal behavior in exchange for actions undertaken in the past. The display of respect here applies not only to behavior towards followers, but also to all other stakeholders. It appears to be a value in itself, regardless of the personal interests of an interviewee. Hence, a dominant or assertive leader’s behavior is valuable only if respect is genuinely displayed, and respectful actions are undertaken. Deference is valued if it is the manifestation of respect to seniors or superiors. Care for kin is important as respect to a family, which serves as a predictive signal of respectful behavior towards followers.
Hence, the degree of influence on the evaluation of a leader as ethical exerted by congruence in such MAC values as care for kin, heroism, and deference depends on the genuine subjection of such a behavior to the group’s interests, and the degree of reciprocity in wider terms that is displayed by the leader. Practically speaking, congruence with followers on these two mediating MAC values should be the first priority for a leader aiming to earn evaluation as ethical, along with fairness in distributing the disputed resources and respect to prior possession. In other words, to justify any decision without risking losing followers’ ethical evaluation, the leader must operate with care for the group, reciprocity in actions, and fairness in resource distribution, paying full respect to others’ property.
While appeal to heroism, deference, and care for family are not expected to give a substantial effect on the perceived level of ethical leadership, this does not mean that they are not valued by followers or that they have no influence on evaluation at all. In a hypothetical situation when group members were to elect one of two leaders equally perceived as devoted to the group, and congruent with them in feeling the importance of reciprocity, fairness, and respect to prior possession, they will most likely elect that person who is perceived better in family values, heroism and deference.
The interdependence between MAC values appears to be even more complex than the mediation discussed above. As a matter of common sense, moral assessment should be a composite phenomenon, and the qualitative phase provides empirical support for such an intuitive guess. The data collected in Phase 2 provided clear evidence of the insufficiency of congruence in one standalone basic moral value for moral assessment. In the same vein, the data show limited ability of a follower–leader perceived incongruence in one standalone basic moral value to destroy the perception of the leader as ethical, given the inevitable influence of other values. The interviewees point to moral assessment as a complex phenomenon comprising numerous small factors. In the meantime, besides the value hierarchy and moral assessment complexity in ethical leadership, there are more findings unforeseen in the beginning of the research.
The data for Phase 2 demonstrate possible counter-productivity of the moral management factor, which represents one of the pillars of Brown et al.’s (2005) concept of ethical leadership. The interviewees argue that excessive talk on ethics, or inconsistency of that talk with a leader’s behavior, results in lowered evaluation of the ethical leader. This finding is consistent with recent research (Pircher Verdorfer & Peus, 2020). Nevertheless, this study supports the main proposition of ethical leadership theory, which posits that the exemplary role of a leader acts as the main factor in the leader’s influence on followers.
While the causal positive influence of congruence in MAC values on the perceived level of ethical leadership was found established, some specifics of such influence were not evident before Phase 2 was conducted. The qualitative data demonstrate the special role of a leader, who apparently is perceived differently compared to other people in the group. The way the interviewees explained what is right and wrong in the behavior of a leader leaves the impression that in their view, there is the leader and “the others,” two parts of a team with equal weight, no matter how large the “others” part is.
Recent literature suggests that assessing the morality of a leader differs from assessing others (Crockett, Everett, Gill, & Siegel, 2021). For example, instrumental harm attributed to a leader can be viewed as morally good, as it entails benefit for the group, while it is morally unacceptable when caused by peers. Crockett and colleagues suggest that the deontological approach to the moral assessment of peers offers more certainty in terms of future inter-personal comfort in possible cooperation, whereas the peers’ tendency to apply instrumental harm entails future threat to an individual and may thus be condemned, even if it brings the utility for the group.
In contrast, people apply the utilitarian approach more willingly when evaluating leaders (Crockett et al., 2021). Due to the dependence of their actual or envisioned safety and well-being on leaders’ actions, people tend to afford instrumental harm to certain group members if it results in their group’s benefit, and thus to assess such behavior as morally good. Such an explanation gives the basis to understanding the value hierarchy revealed by this study.
Although the sample size of the Phase 1 survey is satisfactory according to criteria suggested by the literature for regression analysis (see Fritz & MacKinnon, 2007), this might not be sufficient for good model fit in different circumstances. Keeping in mind Schreiber et al.’s (2006) arguments, it is recommended for future research to retest model fit on samples exceeding the product of 10 and the number of measured parameters. Moreover, to be generalizable across multiple groups, all the findings on association, moderation and mediation ideally need a retest in a multi-site research design with a substantially larger sample of at least M × N × 33 respondents, meaning M groups multiplied by N categories and 33 valid answers in each category.
The content validity of the MAC items of the resulting modified instrument might represent an issue for future research and practical applications. These items were adopted from Curry, Chesters, and Van Lissa (2019a), who, in turn, drew them from the peer-reviewed literature. However, Curry et al. argue that these items may measure different aspects of morality, even though they are within a designated group of typical solutions for cooperation problems. For better model fit, the original MAC-Q set of items have been reduced from an overall six per each MAC value, combining Judgment and Relevance scales, to just two items per construct, which can narrow substantially the possible aspects of morality covered by the resulting scales. Further research should examine their exhaustiveness and, thereby, the content validity of the entire instrument.
Due to the focus on a company involved in the Russian logistics sector, the findings of both research phases are limited in generalizability to wider cultural and industrial contexts. Tested in different organizational circumstances, the hierarchy of congruence in MAC values might vary. Similarly, as the Russophones are spread across different countries that may culturally influence local diasporas (Zabrodskaja & Ivanova, 2021), the hypotheses of this study would not necessarily be borne out by data from other countries’ Russophones. In the meantime, the qualitative data received affords the expectation of transferability to other groups, as the most participants’ explanations were made in contexts wider than the Company or industry or country. Nevertheless, the results of the present research may be limited in application to a possible wider cultural range of Russophone communities across the globe, and the hypotheses should be re-tested on samples collected in other countries’ diasporas to be applicable to Russophones worldwide.
Although these limitations may prevent the theoretically sound generalization of the results to populations other than the focal one for the present research, this does not mean that these results will necessarily be false for other groups. While further research might examine wider generalization for the sake of scientific rigor, it appears already valuable to employ the findings of this thesis in applied research, management consultancy, and leadership practice.
Banks et al. (2021) point to the lack of research on factors influencing the association between ethical leaders’ behaviors and their perception by followers. Similarly, Brown and Mitchell (2010) point to the lack of research on the influence of followers’ moral identity on the perception of a leader as ethical or unethical. This study contributes to closing this gap by providing empirical support for the influence of followers’ perceived moral value congruence with a leader on the perceived level of ethical leadership. In addition, it contributes to literature on the MAC and ethical leadership by empirical examination of the causal association between the concepts of both theories. Besides that, the discussion of MAC values in terms of perceived value congruence also turns out to be novel to this theory.
Testing of MAC-Q items in a Russian-speaking environment and validating their modified composition widens their cross-linguistic and cross-cultural application to Russophone context. The same is true with the ELS instrument, which had not been applied to Russophones in literature available to the authors. Thus, this study contributes to the MAC and ethical leadership research domains by testing these theories and corresponding instruments in the Russophone context.
The theoretical method employed in this thesis, namely quantitative inquiry with hierarchical regression analysis followed by semi-structured interviews, can be employed by leadership theorists, organizational consultants, or management practitioners for the assessment of the interdependence in the influence of MAC basic moral values on perceived level of ethical leadership within a given organizational context.
Due to established positive group outcomes of the high perceived ethical leadership (Bedi et al., 2016), the enhancement of this perception has an evident practical rationale. Hence, the degree of importance of each MAC element represents valuable information for practitioners aiming to improve the perceived level of ethical leadership. Armed with such knowledge, leaders can adjust their behaviors accordingly, subsequently enhancing followers’ ethical evaluation, which results in enhancement of group’s outcomes.
Moreover, given that congruence in “care for kin”, “heroism” and “deference” moral values is contingent on follower–leader congruence in such MAC elements as “group” and “reciprocity”, the leader’s behavior and narratives reflecting these values should be placed into the context of the safety of and utility for the group, and of respect for others’ efforts and achievements, if the leader aims to attain a better result. If applying the suggested method in different organizational contexts will reveal other value hierarchies, leaders can adjust their behavior accordingly in order to achieve better perceived value congruence.
On the other hand, the above-noted positive feedback loop of the influence of ethical leadership on moral value congruence suggests that followers’ moral values could change in certain conditions. From the practical standpoint, moral value management in a group may appear suitable for pursuing ethical leaders’ purposes. Knowledge of congruence in particular MAC elements may serve as a guide for leaders’ choices regarding moral manifestations to change their followers’ MAC values preferences, which can eventually enhance organizational outcomes in accordance with the ethical leadership theory.
Future research is necessary to expand the theoretical and practical implications of this study. Besides testing its findings beyond the limitations outlined above, some avenues for academic inquiry represent interest in the advancement of knowledge on moral psychology and ethical leadership. This section suggests such avenues from the authors’ perspective, though these suggestions under no circumstances should be considered an exhaustive set of potentially interesting research areas, as the field’s potential scope appears to be substantially wider.
The core of leadership phenomena is a leader’s influence on followers (Northouse, 2019). Brown et al. (2005) build the ethical leadership theory on social learning theory, which involves the role model concept. Gibson’s (2004) concept of a behavior induced by role models may be decomposed to four components, namely attributes of people in social roles, perception of personal attributes as similar, the desire to increase such similarity, and the emulation of that similar behavior. While the perception of similarity was approached by this study to some extent, the phenomenon of social role attribution of MAC elements has not been touched. The knowledge of mechanics underlying the association between MAC values attribution, perceived congruence, and the desire to act similarly turns out to be essential for the understanding of ethical leaders’ influence, so it needs to be addressed in future studies.
Research on the dynamics of the influence of the particular composition of MAC values demonstrated by a leader could make a substantial theoretical contribution given the interdependence of values in moral assessment demonstrated by this research. Such an interplay of MAC values may represent multiple mediation relationships, or depend on the degree of the follower–leader communication frequency, or demonstrate some other specifics that cannot be foreseen based on the findings of this research. More research is thus required for a deeper understanding of the relationship between MAC basic moral values and ethical leadership.
The dynamics of leader’s influence may also depend on the stability of moral evaluation. Recent research shows that moral judgments vary within a week in sacrificial dilemmas (Rehren & Sinnott-Armstrong, 2022). Further research should find an answer to the question of the stability of followers’ MAC value composition. From the practical standpoint, this would provide leaders with knowledge of the moments when the moral accents in their behavior should be re-adjusted to followers’ moral expectations.
Newly formed teams may differ in moral consensus achievement dynamics compared to established groups. Knowledge of the dynamics of belief equilibrium formation can be beneficial for the practice of ethical leadership in early-stage start-up management or in mature corporations’ efforts to expand to new markets. Gavrilets (2021) suggests an approach to modeling group behavior as a function of norms, personal beliefs, and material payoffs. The application of this or a similar approach to group dynamics might entail valuable theoretical and practical implications in the ethical leadership domain. Similarly to newly-established groups with possibly varying moral prioritization across members, mixed-culture teams may face the same specifics and thus represent a promising area for future research.
The differences in moral assessment imposed by either cultural or any other conditions may be reduced to variance in MAC values prioritization, according to the combinatorial hypothesis proposed by Curry et al. (2021). Due to its potentially substantial theoretical and practical implications, it deserves extensive cross-cultural qualitative tests. The qualitative data of Phase 2 provides supporting evidence for such moral values as trust and respect. As the hypothesis claims that any moral value may be reduced to the combination of MAC basic moral elements, qualitative inquiry into wide spectrum of moral values is required.
Besides the prospective research areas mentioned above, there may be others—for example, the influence of the MAC values on narratives that reflect moral evaluation (Kim & Crockett, 2022), the way normative evaluation shapes the behavioral patterns (Theriault et al., 2021) in the context of the MAC and ethical leadership, and so forth. Answers to such questions have practical importance for a group’s behavior engineering, meaning by this the leaders’ fine-tuning of their own narratives and behaviors to achieve the desired followers’ behavior. Such social engineering appears to be a wide and interesting topic for ethical leadership practice.
Keeping in mind the limitations discussed earlier, the data gathered in both phases of this study supports the following statements at least for the population studied.
First, there is a positive association between perceived congruence in MAC basic moral values and the perceived level of ethical leadership, and, in this association, perceived value congruence causally influences the perceived ethicality of a leader, though there is a positive feedback loop to a certain extent. Second, this influence is moderated by the proximity of the leader, understood as the extent to which followers can observe the leader’s behavior, combined with the frequency and duration of the leader’s communication with the followers. Third, the leader’s values observed by the followers are not equal in their influence on the perceived level of ethical leadership; so that influence of congruence in “care for kin”, “hawkish” and “dove-ish” basic moral behaviors on perceived ethical leadership is mediated by congruence in two other basic moral behaviors, namely “loyalty to the group” and “reciprocal exchange”, while congruence in “fairness” and “respect to prior possession” positively influences perceived ethical leadership without mediators.
Besides this, the results of practical application of the morality-as-cooperation and ethical leadership theories in a Russophone organization supports their validity in Russophone contexts. Over and above this theoretical contribution, instruments and method have been offered to ethical leadership practitioners for the assessment of perceived congruence in MAC values, perceived level of ethical leadership, and their interdependence in the Russian language context.
This study was conducted in accordance with the ethical standards of SBS Swiss Business School and adhered to the ethical principles outlined in the Declaration of Helsinki. Approval for conducting both phases of the research was obtained from the SBS Swiss Business School Ethical Committee. The informed consent was received from each participant in both study phases by checking the corresponding box in the Web-form in the Phase 1, and in written and oral forms in the Phase 2.
Alexey Belinskiy: Conceptualization, data curation, formal analysis, funding, investigation, methodology, project administration, resources, visualization, writing — original draft preparation.
Carl Olsen: Supervision, methodology, validation, writing — review and editing.
Open science framework: Influence of MAC Moral Value Congruence on Perceived Level of Ethical Leadership. https://doi.org/10.17605/OSF.IO/42TNC
This project contains following data:
Data are available under the terms of the Creative Commons Attribution 4.0 International license (CC-BY 4.0).
Restricted data: Interview transcripts are available upon request. While anonymized, these transcripts may contain sensitive information about the sample company. As per agreement between the authors and the sample company, qualitative data can only be shared with third parties who demonstrably have no intention or ability to harm the sample company’s business. This determination is at the Corresponding author’s discretion and requires the requesting party to sign a Confidentiality Agreement. The agreement form (Confidentiality agreement for the qualitative data disclosure.pdf ) can be downloaded from https://doi.org/10.17605/OSF.IO/42TNC.
Open science framework: Influence of MAC Moral Value Congruence on Perceived Level of Ethical Leadership [Dataset]. OSF. https://doi.org/10.17605/OSF.IO/42TNC
This project contains following data:
The original, the original interview script and its English translation, and the interview written consent form.
Data are available under the terms of the Creative Commons Attribution 4.0 International license (CC-BY 4.0).
The reporting style adheres to the APA JARS guidelines, namely JARS-Quant (Appelbaum et al., 2018) for the quantitative Phase 1, and JARS-Qual (Levitt et al., 2018) for the qualitative Phase 2.
Referencing style: The Academy of Management Journal http://aom.org/uploadedFiles/Publications/AMJ/AMJ Style Guide(1).pdf style guide (with only italic book and journal titles).
Views | Downloads | |
---|---|---|
F1000Research | - | - |
PubMed Central
Data from PMC are received and updated monthly.
|
- | - |
Is the work clearly and accurately presented and does it cite the current literature?
Partly
Is the study design appropriate and is the work technically sound?
No
Are sufficient details of methods and analysis provided to allow replication by others?
Partly
If applicable, is the statistical analysis and its interpretation appropriate?
Partly
Are all the source data underlying the results available to ensure full reproducibility?
No
Are the conclusions drawn adequately supported by the results?
Partly
References
1. Bartunek J, Rynes S, Ireland R: What Makes Management Research Interesting, And Why Does It Matter?. Academy of Management Journal. 2006; 49 (1): 9-15 Publisher Full TextCompeting Interests: No competing interests were disclosed.
Reviewer Expertise: Work Motivation, Monetary Wisdom, the Love of Money Attitudes, Monetary Aspirations, Business Ethics, Behavioral Economics, Cross-cultural studies
Alongside their report, reviewers assign a status to the article:
Invited Reviewers | |
---|---|
1 | |
Version 1 03 Dec 24 |
read |
Provide sufficient details of any financial or non-financial competing interests to enable users to assess whether your comments might lead a reasonable person to question your impartiality. Consider the following examples, but note that this is not an exhaustive list:
Sign up for content alerts and receive a weekly or monthly email with all newly published articles
Already registered? Sign in
The email address should be the one you originally registered with F1000.
You registered with F1000 via Google, so we cannot reset your password.
To sign in, please click here.
If you still need help with your Google account password, please click here.
You registered with F1000 via Facebook, so we cannot reset your password.
To sign in, please click here.
If you still need help with your Facebook account password, please click here.
If your email address is registered with us, we will email you instructions to reset your password.
If you think you should have received this email but it has not arrived, please check your spam filters and/or contact for further assistance.
Comments on this article Comments (0)