Loaded Down From Speaking Up: A Resource-Based Examination of Voicer Regret Following Supervisor Delegation

A Conservation of Resources Perspective of Enactment on Voicer Regret

Conservation of resources theory positions resource preservation as a fundamental personal goal (Hobfoll, 1989). Indeed, employees “strive to obtain, retain, protect, and foster” resources such as time, energy, or objects valued by employees (Hobfoll, 2001: 341). Because employees possess a strong aversion to resource loss, they experience a negative psychological reaction when they are unable to preserve resources (Hobfoll, 1989, 2001). For example, an employee who is asked to take on more work will likely respond negatively to the obtrusion that consumes their time and energy (Halbesleben, Neveu, Paustian-Underdahl, & Westman, 2014).

Given that changing the status quo in response to voice requires someone's effort (Detert et al., 2013), we expect that the extent to which employees are delegated enactment duties after speaking up will instigate a negative reaction in terms of greater overload (e.g., Anderson & Bolino, 2023; Bolino & Turnley, 2005). That is, supervisor delegation following employee voice creates ongoing obligations (Van Dyne & Ellis, 2004) that sap personal resources (Bolino et al., 2015). Since talk is cheap (Van Dyne, Ang, & Botero, 2003) and enactment is expensive (Hirschman, 1970), voicers may find that extra-role suggestions increase their in-role demands as supervisors delegate voice enactment. For example, enacting change may require navigating complex interdependencies that employees do not fully grasp (Detert et al., 2013). Further, voice involves extra-role challenges to the status quo (Van Dyne & LePine, 1998), which may mean that enactment lies beyond voicer's expertise, thereby requiring extra time and effort to deliver (e.g., Sherf, Parke, & Isaakyan, 2021). Voicers may therefore spend valued time and energy initiating change, seeking others’ support and involvement, and overcoming resistance to change (e.g., Axtell et al., 2000; Hon, Bloom, & Crant, 2014), all of which overloads voicers. Extending prior voice research that emphasizes perceived safety and fear in speaking up (Morrison, 2014), we suggest that employees are also likely to consider future workload in their voice calculus.

Employee overload is a hindrance demand that holds employees back from their desire to preserve resources (Crawford, LePine, & Rich, 2010). The negative experience of overload should prompt counterfactual thoughts related to goal failure. Theory on counterfactual thinking reveals how negative events tied to goal failure introduce counterfactuals about how things could have been better had an employee behaved differently (Epstude & Roese, 2008; Roese, 1997; Roese & Epstude, 2017). In the case of voice, voicers may consider how things could have been better (i.e., less overload) had they not spoken up in the first place. These counterfactual thoughts experienced in the wake of receiving resource-depleting demands that overload voicers signal their failed goal to preserve valued resources of time and energy (Hobfoll, 2001; Roese, 1997).

Theory on counterfactual thinking in the citizenship domain (Anderson & Bolino, 2023), other topic areas (e.g., Obodaru, 2012; Verbruggen & De Vos, 2020), and more generally (Epstude & Roese, 2008; Roese, 1997; Roese & Epstude, 2017) present regret as the primary psychological response associated with counterfactual thinking. Indeed, regret is a signpost of counterfactual thoughts (Mazei, Zerres, & Hüffmeier, 2021) because it arises as self-initiated actions produce negative but avoidable outcomes (e.g., “I wish I had kept my mouth shut”). Thus, different from other emotions—such as fear—that are future oriented (Kish-Gephart, Detert, Treviño, & Edmondson, 2009), regret is tethered to past actions (Roese, 1997). Importantly, regret is particularly salient in planned and intentional acts as employees blame themselves for having proactively chosen to carry out a behavior (Anderson & Bolino, 2023; Kahneman, 1995). Thus, regret is theoretically well-suited to our model given that the construct captures a proximal outcome of counterfactual thinking that voicers experience as they reflect on the overload associated with supervisor delegation.

Combining these arguments, we propose that supervisor delegation following employee voice causes voicers to feel overloaded as they take on more work, which leads to counterfactual thoughts that manifest as employee regret. Because voice is a planned and intentional behavior (Liang et al., 2012), voicers should feel that being overloaded with enactment responsibilities is their own fault because they went out of their way to speak up and thereby invited supervisor delegation. In turn, counterfactual thinking about the negative experience of overload will foster regret for speaking up as demands could have been avoided had voicers not spoken up in the first place (e.g., Anderson & Bolino, 2023; Roese, 1997).

Hypothesis 1: Supervisor delegation following employee voice positively relates to employee regret through employee overload.

Implications for Subsequent Employee Voice

We next consider the effects of overload -induced regret as a barrier to future employee voice. According to conservation of resources theory, resource loss is associated with a negative psychological response as employees “adopt a defensive posture to conserve their resources” (Hobfoll, 2001: 356). Providing an explanation for why this adaptation occurs, we integrate theory on counterfactual thinking to explain why regret motivates behavioral change (Roese & Epstude, 2017). Given the unsavory nature of regret, employees are likely to prioritize actions that minimize regret (Zeelenberg, 1999). Specifically, regret suggests that employees realign their cognitive frame to learn from mistakes following resource loss by better preserving resources and improving their future performance (Morris & Moore, 2000).

Applied to our model, because voice is a discretionary behavior (Van Dyne & LePine, 1998), employees may speak up less in order to avoid receiving additional delegated assignments that further drain their time and energy. Indeed, prior research has shown that feelings of citizenship fatigue are strongly related to subsequently reduced voice (Bolino et al., 2015). Thus, employees should respond to voice-induced regret by withholding input and decreasing their future voice. Employees who anticipate negative outcomes from speaking up—such as continued overload and regret—should be expected to modify their behavior (Roese & Epstude, 2017). Supervisor delegation following employee voice should therefore provoke voicers to withhold subsequent input because employees will come to view voice as volunteering for more work, which would reinforce their failure to protect resources (e.g., Hobfoll, 1989, 2001). This logic—that regret will impede voice—is consistent with theory that regret will reduce subsequent citizenship behavior (Anderson & Bolino, 2023) and research showing that negative emotions diminish voice (Chamberlin et al., 2017; Grant, Parker, & Collins, 2009; Kish-Gephart et al., 2009; Milliken, Morrison, & Hewlin, 2003). In short, employees who regret speaking up will adopt a defensive posture in which they reduce voice to realign with the aim to conserve resources.

Hypothesis 2: Supervisor delegation following employee voice has a negative serial indirect effect on future employee voice through employee overload and employee regret.

The Moderating Role of Supervisor Consultation

We have argued that supervisor delegation following employee voice leads to overload, evoking regret for speaking up that ultimately inhibits voice. However, the extent to which delegation overloads voicers may depend on accessible resources in enacting the suggestion. Conservation of resources theory describes the value of social support from supervisors in reducing the negative effects of demands (Hobfoll et al., 2018). Indeed, when employees are tasked with enacting a suggestion, they may benefit from supervisors’ guidance, consideration, and encouragement to enact change since supervisors are key suppliers of resources for voice enactment (Axtell et al., 2000; Burris, 2012). The voice literature refers to this guidance, consideration, and encouragement, as supervisor consultation (Tangirala & Ramanujam, 2012).

As conduits of key organizational resources (Morrison, 2014), supervisors can provide instrumental resources lacked by subordinates to facilitate enactment of a delegated idea. Supervisor consultation conveys resources targeted to what voicers need in the moment (e.g., House, 1971)—such as guidance regarding how best to enact their idea. Supervisors can guide voicers about how to navigate complex interdependencies and challenges visible from a managerial view but perhaps obscured to voicers (Burris et al., 2017). This guidance should weaken how overloaded employees feel. Moreover, consultation should be taken as a sign of consideration, wherein supervisors communicate concern and respect alongside delegation (e.g., Judge, Piccolo, & Ilies, 2004). The consideration that accompanies supervisor delegation of enactment responsibilities should increase employee beliefs in the value of enacting their suggestions—that the benefits of doing extra work are worthy of the extra costs—which diminishes perceptions that employees are overloaded (Bandura, 2000; Jex & Bliese, 1999). Relatedly, consultation also entails encouragement that enables voicers to proceed more confidently through their delegated responsibilities (Cheong et al., 2016). Thus, supervisor consultation may heighten voicers’ efficacy in enacting ideas, thereby curtailing voicers’ sense of overload. Taken together, as supervisors get involved through consultation related to voice enactment, they offer guidance, consideration, and encouragement that should weaken the effect of supervisor delegation following employee voice on employee overload.

In contrast, lower supervisor consultation should strengthen the relationship between supervisor delegation following employee voice and employee overload. When supervisors delegate enactment responsibilities to voicers and provide little consultation, voicers should experience more overload because they do not receive clear guidance. The lack of guidance from supervisors may leave voicers feeling deserted as they navigate complexities on their own. As a result, voicers may be more likely to view supervisor delegation following employee voice as a hindrance stressor because the supervisor does not help with goal progress during enactment (e.g., Crawford et al., 2010). Moreover, low consultation signals less consideration for voicers, which conveys an absence of appreciation and support (Bass, 1990) that may strengthen voicer perceptions of excessive hindrance demands (Schaubroeck & Fink, 1998; Van Yperen & Hagedoorn, 2003). Low supervisor consultation may also reflect reduced encouragement for voicers who carry out delegated responsibilities by themselves, thereby leaving employees with a sense of an uphill battle to make changes (e.g., Jex & Bliese, 1999). Thus, low supervisor consultation should leave employees with diminished resources and associated guidance, consideration, and encouragement, ultimately strengthening the effect of supervisor delegation following employee voice on overload.

Hypothesis 3: Supervisor consultation moderates the positive relationship between supervisor delegation following employee voice and employee overload, such that employees experience less (more) overload when supervisor consultation is high (low).

Applying these arguments to our overall model, we expect that supervisor consultation will moderate the serial indirect effects of supervisor delegation following employee voice on employee outcomes. More precisely, when supervisor consultation is high, supervisor delegation following employee voice should have a weaker indirect effect on employee regret through employee overload, resulting in increased voice. In contrast, when supervisor consultation is low, supervisor delegation following employee voice should have a stronger indirect effect on employee regret through employee overload, resulting in decreased voice.

Hypothesis 4a: Supervisor consultation moderates the positive indirect relationship between supervisor delegation following employee voice and employee regret through employee overload such that employees experience less (more) regret when supervisor consultation is high (low).

Method

Study 1: Measures

All measures were rated using a scale from 1 = strongly disagree to 5 = strongly agree.

Supervisor Delegation Following Employee Voice. Supervisors rated their delegation following employee voice using a 5-item measure developed for this study (see the Appendix in the online supplemental materials for validation evidence). The lead-in to the items was “During the past month …” The items are “I provided assignments to this employee to carry out their suggestions,” “I delegated tasks to this employee to enact their suggestions,” I gave this employee additional responsibilities related to their suggestions, “I assigned this employee extra work duties related to implementation of their suggestions,” and “I asked this employee to take on new obligations related to their suggestions” (α = .94).

Employee Overload. Employees rated their overload with the 3-item measure adapted from Bolino and Turnley (2005) to the delegated voice context. The lead-in to the items was “During the past month …” The items are “The amount of work I received as a result of speaking up is too great,” “It often seemed like I receive too much work related to my suggestions,” and “I never had enough time to complete follow-up assignments related to my suggestions” (α = .94).

Employee Regret. Employees rated their regret using a 5-item measure expanded from Tsiros and Mittal's (2000) 3-item measure of regret and adapted to the delegated voice context. We added additional items consistent with the spirit of the construct to ensure that we would capture the entirety of the construct domain. The lead-in to the items was “During the past month …” The items are “I feel sorry for speaking up with new suggestions,” “I regretted speaking up with my suggestions,” “I should not have shared my suggestions,” “I wished that I had not expressed my suggestions,” and “I feel that I made a mistake in voicing my suggestions” (α = .98).

Supervisor Consultation. Employees rated supervisor consultation using the 3-item measure adapted from Tangirala and Ramanujam (2012). Whereas the original measure included a voice solicitation item, we adapted that item to focus on supervisor consultation after employees had spoken up. The items are “My supervisor gives me clear instructions about how to implement my suggestions,” “My supervisor consults with me to get my reactions and input about major changes that will affect me,” and “My supervisor encourages me to express any concerns or doubts I may have about a proposal that is under consideration” (α = .87).

Employee Voice. Supervisors rated employee voice using the 6-item measure of voice from Van Dyne and LePine (1998). Although scholars have noted potential contamination with two items in this measure (Grant et al., 2009; Liang et al., 2012; Tangirala & Ramanujam, 2012), our results are consistent with or without those items, and we maintained the original 6-item validated scale. The lead-in to the items was “During the past month, this employee …” Sample items are “Spoke up in this group with ideas for new projects or changes in procedures,” and “Developed and made recommendations concerning issues that affect this work group” (α = .93).

Control Variables. We used control variables to account for alternative explanations and assess the robustness of our findings. Given the prevailing explanation that employees do not voice due to fear of negative consequences (i.e., psychological safety; Morrison, 2014), we controlled for psychological safety in our primary analysis. We captured psychological safety using the 3-item measure from Burris, Detert, and Chiaburu (2008; e.g., “It is safe for me to speak up around here”). Since psychological safety represents an alternative explanation to employee regret—that is, the negative effect of supervisor delegation following employee voice on subsequent voice could be transmitted through psychological safety—we specified psychological safety as a parallel mediator to employee regret. This approach enabled us to test whether our theorizing about the role of regret goes beyond the existing theory about psychological safety.

We also tested demographic control variables that could relate to voice and employee–supervisor dynamics: gender, education, organization tenure, and dyadic tenure. Including these controls did not change the pattern or significance level of our results. In accordance with best-practice recommendations to exclude controls when they do not change results (Becker, 2005; Carlson & Wu, 2012), we did not retain these variables in our final model. We do, however, present correlations of these control variables in Table 1.

OPEN IN VIEWER

Table 1

Descriptive Statistics, Correlations, and Reliabilities (Study 1)

Variable	Mean	SD	1	2	3	4	5	6	7	8	9	10
1. Supervisor Delegation Following Employee Voice	3.34	1.04	(.94)
2. Employee Overload	2.32	1.15	.35*	(.94)
3. Employee Regret	1.91	1.10	.24*	.56*	(.98)
4. Supervisor Consultation	3.72	0.85	−.25*	−.53*	−.46*	(.87)
5. Employee Voice	3.67	0.82	−.19*	−.48*	−.38*	.44*	(.93)
6. Psychological Safety	3.83	0.87	−.20*	−.43*	−.45*	.27*	.30*	(.95)
7. Gender	−	−	−.01	−.02	−.01	.05	.09*	−.02	−
8. Education	−	−	.00	.05	.02	−.01	−.01	−.01	−.03	−
9. Organization Tenure	21.13	9.94	.03	.06	.02	.01	−.04	−.03	.05	.09*	−
10. Dyadic Tenure	8.67	3.70	.05	.06	.03	−.01	−.05	−.04	.06	.07	.69*	−

Note: N = 587. Coefficient alpha is provided along the diagonal. Education was coded as 0 = no college degree, 1 = college degree or higher. Controls for gender, education, organization tenure, and dyadic tenure did not change our model results and were removed from our final model in accordance with recommendations to exclude controls when they do not alter results (Becker, 2005; Carlson & Wu, 2012). We include them in the table for informational purposes.

*p < .05; two-tailed.

Study 1 Analysis and Results

Table 1 presents descriptive statistics, correlations, and coefficient alphas. We examined the factor structure of our model variables using a multilevel confirmatory factor analysis in Mplus 7.4 (Muthén & Muthén, 2015). The 6-factor measurement model exhibited good fit to the data: χ ² (260) = 354.92, p < .001; CFI = .99; SRMR = .02. All factor loadings were significant.

Given that supervisors in our sample managed multiple employees, there is the potential for non-independence of observations due to systematic differences in supervisor ratings (Bliese, 2000). To account for these differences, we designated “twolevel” as our model type, clustered on supervisor identification numbers, and specified all variables at the within level (Muthén & Muthén, 2015). This approach provides a more accurate and conservative test of predictions than single-level modeling because it utilizes clustered standard errors (Muthén & Muthén, 2015).

To test our hypothesized interactions, we grand-mean centered our exogenous predictors to improve interpretability and remove nonessential multicollinearity between variables and product terms (Cohen, Cohen, West, & Aiken, 2003; Cortina, Chen, & Dunlap, 2001). To test our hypothesized indirect effects, we modeled the direct effects of supervisor delegation following employee voice on mediators and outcome variables (MacKinnon, Lockwood, Hoffman, West, & Sheets, 2002). Because the product of path coefficients is not necessarily normally distributed, we examined indirect effects using a Monte Carlo resampling approach with the RMediation software package (Tofighi & MacKinnon, 2011) and evaluated our conditional indirect effects by testing for a significant difference of the indirect effects at high and low levels of supervisor consultation using 20,000 simulations. Our model exhibited acceptable fit to the data: χ ² (6) = 66.50, p < .001; CFI = .91; SRMR = .08.

Hypothesis 1 predicted that supervisor delegation following employee voice would positively relate to employee regret through employee overload. As shown in Table 2 and Figure 1, and as a pre-condition to this indirect effect, supervisor delegation following employee voice had a positive relationship with employee overload (γ = .29; SE = .04; p < .001) and employee overload had a positive relationship with employee regret (γ = .53; SE = .04; p < .001). Using Monte Carlo simulation as a re-sampling technique to create sampling distributions and 95% bias-corrected confidence intervals (Selig & Preacher, 2008), we looked for confidence intervals excluding zero to indicate a significant indirect effect. Supporting Hypothesis 1, supervisor delegation following employee voice had a positive indirect relationship with employee regret through employee overload (IND = .15; 95% CI [101, .213]).

OPEN IN VIEWER

Table 2

Results of Multilevel Path Analysis (Study 1)

	Employee Overload		Psychological Safety		Employee Regret		Employee Voice
Variables	γ	p Value	γ	p Value	γ	p Value	γ	p Value
Supervisor Delegation Following Employee Voice	0.29*	< .001	−0.05	.16	0.05	.18	−0.01	.88
Employee Overload			−0.31*	< .001	0.53*	< .001	−0.20*	< .001
Psychological Safety							0.22*	< .001
Employee Regret							−0.11*	.004
Interaction of Supervisor Delegation and Supervisor Consultation	−0.16*	< .001
Supervisor Consultation	−0.57*	< .001
R2	0.34*	< .001	0.32*	< .001	0.19*	< .001	0.31*	< .001

Note: N = 587. Hypothesized coefficients are bolded.

* p < .05; two-tailed.

OPEN IN VIEWER

Figure 1

Results of Multilevel Path Analysis for the Effects of Supervisor Delegation Following Employee Voice (Study 1)

Hypothesis 2 predicted that supervisor delegation following employee voice would have a negative serial indirect effect on future employee voice through overload and regret. As reported in Table 2 and Figure 1, employee regret had a negative relationship with future employee voice (γ = −.11; SE = .04; p = .004). Supporting Hypothesis 2, the negative indirect relationship between supervisor delegation following employee voice and future employee voice through overload and regret was also significant (IND = −.02; 95% CI [−.025, −.006]).

Hypothesis 3 predicted that supervisor consultation moderates the positive relationship between supervisor delegation following employee voice and employee overload, such that employees experience less (more) overload when supervisor consultation is high (low). As reported in Table 2 and Figure 1, the interaction between supervisor consultation and supervisor delegation following employee voice on employee overload was significant (γ = −.16; SE = .04; p < .001). As depicted in Figure 2 and supporting Hypothesis 3, this relationship was weaker when supervisor consultation was high (simple slope = .15, SE = .05, p = .001) and stronger when supervisor consultation was low (simple slope = .42, SE = .07, p < .001).

OPEN IN VIEWER

Figure 2

Interaction of Supervisor Delegation Following Employee Voice and Supervisor Consultation on Employee Overload (Study 1)

Hypothesis 4a predicted that supervisor consultation moderates the positive indirect effect between supervisor delegation following employee voice and employee regret through overload such that employees experience less (more) regret when supervisor consultation is high (low). The indirect relationship of supervisor delegation following employee voice on employee regret through overload was weaker at high levels of supervisor consultation (IND = .08; 95% CI [032, .131]) and stronger at low levels of supervisor consultation (IND = .23; 95% CI [ .145, .310]). Importantly, the difference between high and low levels of supervisor consultation was significant (Δ IND = −.15; 95% CI [−.255, −.061]), providing support for Hypothesis 4a.

Finally, Hypothesis 4b predicted that supervisor consultation moderates the negative serial indirect effect of supervisor delegation following employee voice on future employee voice. In applying the first- and second-stage moderation effects of supervisor consultation to the full mediation chain outlined by our model, the indirect effect was weaker at high levels of supervisor consultation (IND = −.01; 95% CI [−.019, −.002]) and stronger at low levels of supervisor consultation (IND = −.03; 95% CI [−.045, −.007]). Providing support for Hypothesis 4b, the difference between the high and low levels of supervisor consultation was significant (Δ IND = .02; 95% CI [.003, .045]).

Study 1 Supplemental Analysis

Based on theory regarding the effect of psychological safety and voice (Morrison, 2014), we controlled for psychological safety as an alternative mechanism for our effects. When positioned as a parallel mediator to regret, we found that employee overload was negatively related to psychological safety (γ = −.31; SE = .04; p < .001), consistent with Kahn's (1990) view that safety is a proximal antecedent to fully investing oneself at work. To evaluate the relative strength of effects on regret and psychological safety, we used a Wald test to compare the absolute value of the indirect effects of supervisor delegation following employee voice and employee overload on both employee regret and psychological safety. We found that supervisor delegation following employee voice had a significantly stronger effect on regret compared to the indirect effect on psychological safety (Δ IND = .06; ω = 11.43, p < .001), and the difference in these indirect effects on voice was significant (Δ IND = .001; ω = 4.83, p = .028). Moreover, when psychological safety was positioned as a parallel mediator to overload—with no other changes to the model from our primary analysis—our hypothesized results were also unchanged. Thus, although psychological safety may be relevant to our model, these findings—anchored in theory on counterfactual thinking which highlights the role of regret (Epstude & Roese, 2008; Roese & Epstude, 2017)—support our focus on regret as a novel and more significant predictor of future voice behavior following supervisor delegation than psychological safety.

Study 1: Discussion

In a three-wave study of employees and supervisors, we found that supervisor delegation following employee voice elicited employee regret that impeded subsequent voice. Study 1 enabled us to pinpoint overload as the theoretical mechanism that connects delegation-induced resource loss and regret for speaking up. Importantly, these findings supported our theoretical integration of conservation of resources and counterfactual thinking: Supervisor delegation following employee voice diminishes employee resources in the form of increased overload, prompting voicers to regret speaking up and to safeguard resources by withholding voice. Further, the indirect effects of supervisor delegation following employee voice on future voice were stronger through overload and regret than through overload and psychological safety. We also found that employee overload and subsequent regret was attenuated when employees received supervisor consultation—the indirect effects of delegation following employee voice on employee overload, regret, and future voice were weaker when supervisor consultation was high.

To establish causality, mitigate situational confounds, and address individual differences through random assignment, we next employed an experimental causal chain design in Study 2 (Spencer et al., 2005). We conducted an experiment (Study 2a) that examines the first stages of our model—the effect of supervisor delegation following employee voice on employee regret through overload—and an additional experiment (Study 2b) to examine the last stage of our model—the effect of employee regret on employee voice. This design enabled us to draw conclusions about causality from one stage of the model to the next.

Method

Study 2a: Analysis and Results

Results from our analysis of variance (ANOVA) demonstrated the effectiveness of our delegation manipulation. There was a significant main effect of the high delegation condition on our measure of supervisor delegation following employee voice (F = 168.53, M = 5.91 vs. 2.54, p < .001). We then used ANOVA to examine our predictions about supervisor delegation following employee voice and employee overload. As shown in Table 3, employee overload was higher in the high delegation condition than in the low delegation condition (F = 37.55, M = 3.94 vs. 2.24, p < .001), offering causal support for the theoretical link between supervisor delegation following employee voice and employee overload. Moreover, an ANOVA revealed that employee regret was also higher in the high delegation condition than in the low delegation condition (F = 23.58, M = 3.20 vs. 1.87, p < .001).

OPEN IN VIEWER

Table 3

Results of Analysis of Variance (Studies 2a and 2b)

Study 2a
Supervisor Delegation Following Employee Voice → Employee Overload → Employee Regret
Mean for overload in high delegation condition (n = 68)	3.94 (1.96)	F(1, 139) = 37.55*
Mean for overload in low delegation condition (n = 72)	2.24 (1.26)
Mean for regret in high delegation condition (n = 68)	3.20 (1.95)	F(1, 139) = 23.58*
Mean for regret in low delegation condition (n = 72)	1.87 (1.23)
Study 2b
Employee Regret → Employee Voice
Mean for voice in high regret condition (n = 75)	3.89 (1.57)	F(1, 139) = 57.37*
Mean for voice in low regret condition (n = 65)	5.50 (0.90)

Note: Each study represents an independent sample. Standard deviations are listed in parentheses.

* p < .05; two-tailed.

Hypothesis 1 predicted a positive indirect effect of supervisor delegation following employee voice on employee regret through overload. In line with Study 1, we tested this prediction using the product of coefficients approach in Mplus (Version 7.4; Muthén & Muthén, 2015). We again used Monte Carlo resampling in the RMediation software package to produce bias-corrected confidence intervals (Tofighi & MacKinnon, 2011). In accordance with our ANOVA findings, supervisor delegation following employee voice was positively associated with employee overload (b = 1.32; SE = .27; p < .001) and employee overload was positively associated with employee regret (b = .61; SE = .06; p < .001). In support of Hypothesis 1, supervisor delegation following employee voice had a positive indirect effect on employee regret through employee overload (IND = .81; 95% CI [.628, .989]).

Hypothesis 3 predicted that supervisor consultation moderates the relationship between supervisor delegation following employee voice and employee overload, such that employees experience less (more) overload when supervisor consultation is high (low). The interaction of supervisor delegation following employee voice and supervisor consultation was a significant predictor of employee overload (b = −.36, SE = .16, p = .02). As depicted in Figure 3 and supporting Hypothesis 3, this relationship was weaker when supervisor consultation was high (simple slope = .72, SE = .37, p = .054) and stronger when supervisor consultation was low (simple slope = 1.92, SE = .38, p < .001).

OPEN IN VIEWER

Figure 3

Interaction of Supervisor Delegation Following Employee Voice and Supervisor Consultation on Employee Overload (Study 2a)

Hypothesis 4a predicted that supervisor consultation moderates the positive indirect effect between supervisor delegation following employee voice and employee regret through overload such that employees experience less (more) regret when supervisor consultation is high (low). The indirect relationship of supervisor delegation following employee voice on employee regret through overload was weaker at high levels of supervisor consultation (IND = .44; 95% CI [ −.006, .910]) and stronger at low levels of supervisor consultation (IND = 1.18; 95% CI [.684, 1.716]). The difference between high and low levels of supervisor consultation was significant (Δ IND = −.74; 95% CI [−1.439, −.080]), providing support for Hypothesis 4a.

Study 2: Discussion

Using an experimental causal chain design (Spencer et al., 2005), we demonstrated the effect of supervisor delegation following employee voice on overload and regret (Study 2a) and regret on future employee voice (Study 2b). We also found converging support for the moderation of supervisor consultation in dampening the negative implications of supervisor delegation following employee voice on overload and regret. Overall, Study 2 offered causal support for our integration of conservation of resources theory and theory on counterfactual thinking in detailing how delegation obstructs employee voice via increased overload and regret.

Theoretical Implications

Organizations need employee input to improve and innovate. However, changing the status quo is not without costs. We advance the voice literature by theorizing that voicers may be obligated to enact change, bringing to light a neglected thread of Hirschman's (1970: 39) foundational theorizing regarding the “cost of voice” incurred during enactment. Indeed, an awareness of such costs reveals how voice layers on responsibilities for the good soldiers in organizations (e.g., Bolino et al., 2015; Organ, 1988), elucidating a puzzle in the voice literature that reveals how voice enactment operates as an unexpected hurdle for future voice. Research implies that voicers respond positively to supervisors who validate their input and initiate organizational change, such that employees are motivated to keep voicing (e.g., Janssen & Gao, 2015; Lam, Rees, Levesque, & Ornstein, 2018; Wei et al., 2015). However, the implications of our research reveal potentially negative implications for voicers. Supervisor approval of an idea in the form of delegating enactment responsibilities shapes voicers’ retrospective view about speaking up in a way that impairs their future voice. Thus, we advance the voice literature by theorizing how delegation complicates getting endorsement for voicers, turning what has been portrayed as a benefit to voicers into an obstacle for sustained speaking up. In building theory about how voicers react to voice-enactment obligations, our work introduces a resource-based perspective of why employees may not speak up. This insight may provide one explanation for why voice has a negative trajectory in employee-supervisor dyads (Li & Tangirala, 2021) as employees learn over time that speaking up can contribute to resource loss and overload.

Importantly, we illuminate the role of voicer regret stemming from overload in discouraging future employee voice. Voice research has highlighted how fear or lack of psychological safety reduces voice (Kish-Gephart et al., 2009; Morrison, 2014). Yet, our theoretical integration of conservation of resources theory and theory on counterfactual thinking positions overload-producing regret as a critical factor of voice beyond previous explanations. Although the protection of personal resources is a core principle of conservation of resources theory, Hobfoll and colleagues (2018: 106) have argued that the causes and conditions of individuals’ entering a defensive mode to protect resources is “the least researched principle of COR theory but one that has high explanatory power.” Our findings support this principle and reveal that employees have a reactive aversion to resource loss informed by regret. This view of seemingly more deliberate resource protection highlights how regretful employees strategically withhold voice to protect resources. Specifically, voicers seek to avoid being volunteered for more work, which expands the literature's discussion of fear and psychological safety as factors that reduce voice (e.g., Kish-Gephart et al., 2009; Morrison, 2014) to also include the calculus of one's future workload spurred by regret. In short, after disadvantageous voice encounters, voicers feel regret and modify their behavior to preserve resources and mitigate ongoing enactment costs.

Finally, our research speaks to supervisors’ role in voice enactment. A tacit assumption in the voice literature is that supervisors drive enactment as supervisors take corrective action based on employee input to enhance the work environment (Burris et al., 2017; Detert et al., 2013; Li & Tangirala, 2021). Rather than enactment representing a way in which supervisors can show employees that their voice makes a difference, our work highlights the ramifications of supervisors’ involving voicers in enactment via delegation. Although research has begun to show some drawbacks to delegation (Cheong et al., 2016), research generally focuses on the benefits of delegation (Chen & Aryee, 2007; Cheong, Yammarino, Dionne, Spain, & Tsai, 2019; Schriesheim et al., 1998). We advance this dialogue by pinpointing how delegation, while reducing supervisors’ overload (Akinola et al., 2018), overloads proactive voicers. To mitigate employee overload, our findings point to the importance of supervisor consultation that provides enactment-related resources to guide and encourage employee efforts—with greater consultation reducing voicer overload. Thus, our work highlights the value of infusing supervisor delegation with supervisor consultation to attenuate the negative effects of supervisor delegation on future employee voice.

Limitations and Future Research

Despite the strengths of our studies, there are several limitations. Our examination of the resource inflows and outlays voicers encounter in response to supervisor delegation was intentionally focused. We specifically focused on regret as the proximal response that employees experience when they engage in counterfactual thinking regarding their overload and failure to preserve resources (Anderson & Bolino, 2023). However, a broader examination of the resources employees gain or lose in the downstream voice-enactment process could be worthwhile. Such inquiry could consider the resources consumed and received through boundary spanning (Ancona & Caldwell, 1992) or seeking others’ participation (Chamberlin et al., 2018) in voice enactment. Future work may also consider human and social capital resource flows that accrue to voicers (Newton et al., 2022). In short, an examination of the resources potentially required and acquired as voicers enact their suggestions could be an insightful line of inquiry.

Although our theorizing presents regret as a natural outcome of supervisor delegation following employee voice, alternative theoretical frames highlight other potential emotional responses to voice enactment. A justice perspective may suggest that other-focused emotions such as resentment or anger toward supervisors ensue when speaking up elicits overload (Colquitt & Zipay, 2015). A social comparison perspective may suggest that employees experience sympathy or schadenfreude as they view supervisor delegation following coworker voice (Smith, 2000). Attribution theory may suggest that voicers experience guilt or shame for not pitching their idea well enough (Roseman, 2013). Or, cognitive-motivational-relational theory may suggest that voicers experience anxiety when they contemplate the additional enactment responsibilities they could receive (Lazarus, 1991). Examining our model through alternative theoretical frames could lend additional insight into voice enactment to broaden a growing body of research on voice and emotions (Grant, 2013; Kish-Gephart et al., 2009; Milliken et al., 2003; Welsh et al., 2022).

Finally, we encourage research to build on our theoretical model. Although we focused on constructive voice, our model could be deepened or extended to accommodate other ways of speaking up. Going deeper into constructive voice, research could consider voicer reactions in enacting promotive versus prohibitive voice (e.g., Liang et al., 2012). Moreover, more negative types of challenging voice (i.e., defensive voice; Maynes & Podsakoff, 2014) could also be considered. For example, employees who complain about organizational policies and then are asked to resolve such policies themselves may experience heightened feelings of overload and regret. Further, examining “who does the work after voice” could investigate what motivates supervisors to take on voice or delegate it. Future research could consider ratings from supervisors’ and employees’ perspectives (e.g., Matta et al., 2015; Van Dyne & LePine, 1998) to uncover insights regarding the match between a supervisor's and a voicer's preferences for who takes charge of implementing suggestions. Research on enactment could also include coworkers (Satterstrom et al., 2021) to examine how coworkers respond to delegated enactment and whether their relationships with the delegating supervisor or idea-generating voicer are affected.

Practical Implications

Our work has implications for employees, supervisors, and organizations. For employees, our findings expand the cost-benefit calculus related to speaking up (Morrison, 2014) because voice may obligate voicers to enact additional work responsibilities that deplete their personal resources. Thus, voicers would do well to adjust their expectations regarding their involvement in enactment by recognizing that they may be asked to share some of the responsibilities when they speak up with improvement-oriented suggestions. Accordingly, it may be valuable for employees to speak up selectively about issues they deem worthy of their time and energy since more work may accompany their voice.

For supervisors, delegation coupled with consultation appears critical. When an employee pitches an idea, returning it with the directive to “go run with that” may be tempting to reduce a supervisor's workload. However, such delegation may be suboptimal because it leads employees to regret speaking up and, in turn, discourages their future voice. Instead, our findings suggest a supplemental strategy from supervisors to contain employee regret. Specifically, supervisors would do well to offer consultation—guidance, consideration, and encouragement—when they delegate. This consultation offers calculated resources that allow supervisors to facilitate voicer enactment and curtail voicer regret when allocating work. Without consultation, what may otherwise be empowering supervisor delegation is at best ineffective and may even backfire.

Organizations need voice to innovate and avoid pitfalls. Although supervisors may find it convenient to delegate to voicers, organizations should be mindful that doing so could impede the flow of employee input. Since voicers are seen as high performers (Chamberlin et al., 2018), organizations should be sensitive that they do not overload their best workers who frequently voice. Doing so may cause otherwise valuable voicers to become strategically silent. To prevent this from occurring, organizations should provide voicers with resources to avoid provoking regret that results in decreased voice across the organization. In sum, organizations should be aware of how the costs of voice can be minimized to generate the many organizational benefits of voice.