Learning goal orientation and promotive voice: A moderated mediation model.

Drawing on the broaden-and-build theory and trait-activation theory, this study investigates the mediating effect of thriving at work on the relationship between learning goal orientation (LGO) and promotive voice behavior, as well as the moderating effect of intrinsic career growth (ICG) on the relationship between employees' LGO and thriving at work. Using the two-wave design with a 4-month time lag involving 279 employees, the results demonstrate that employees' LGO is positively associated with promotive voice behavior by thriving at work. Furthermore, ICG moderates the relationship between LGO and thriving at work. ICG also moderates the mediating effect of thriving at work on the relationship between LGO and promotive voice behavior, such that the mediating effect is only significant when employees perceive high ICG.

The COVID-19 pandemic has dramatically changed the way we live and work; thus, organizations are encountering significant new challenges. In today’s unpredictable and insecure business environment, employees’ proactive inputs are highly beneficial for organizational competitiveness (Bindl & Parker, 2010). One way of obtaining such input from employees is through promotive voice behavior, defined as their expression of new ideas or solutions for improving organizational functioning (Morrison, 2014). That is because employees’ constructive ideas and opinions result in organizational benefits by improving organizational functions, as well as personal benefits, such as increased visibility (Stamper & Van Dyne, 2001) and favorable performance evaluations for speakers (Thompson, 2005). However, presenting one’s opinions and suggestions also entails the risk of being misunderstood by peer colleagues and supervisors. For example, voice behavior can be misinterpreted as “bossiness, unsolicited interferences, and an effort to undermine the credibility” (Liang et al., 2012, p. 72), which can lead to undesirable social consequences for employees. Therefore, despite the benefits of the constructive voice, employees experience reluctance to speak up their ideas, opinions, and suggestions in the workplace.

Given the importance of voice behavior to employees and organizations, numerous studies have examined various organizational and situational contexts as determinants of employee voice behavior. Contextual factors include supervisor and leader behavior (e.g., Detert & Burris, 2007; Zhang et al., 2014); support from coworkers and the organization (e.g., Tucker et al., 2008); fair work environments (Bies & Shapiro, 1988; Zhang et al., 2014); job autonomy (Lam & Mayer, 2014); employees’ personal understanding of voice friendly environments, such as safety (e.g., Burris et al., 2008; Liang et al., 2012; Liu et al., 2015); and workplace climate (e.g., Lee et al., 2014; Zhou & George, 2001). Adding to contextual factors, individual dispositions are also considered the key factor for voice behaviors. A recent meta-study has shown that individual dispositions are critical elements that influence employees’ voice behavior (Chamberlin et al., 2017). This is because the people’s attributes (e.g., internal characteristics or personality), not the nature of contexts, are the fundamental determinants of behavior (Schneider, 1987). In particular, discretionary behaviors like voice behavior are better predicted by individual dispositions (Motowidlo et al., 1997). Some individuals tend to engage in voice behaviors despite situational obstacles because they are inherently more capable or willing (Chamberlin et al., 2017; Van Dyne et al., 1995). Empirical studies have found that voice is related to Big Five personality (Lee et al., 2014; LePine & Van Dyne, 2001) and other dispositional constructs such as core self-evaluation (Aryee et al., 2017), proactive personality (Crant et al., 2011; Liang & Gong, 2013), and approach and avoidance orientation (Kakkar et al., 2016). However, despite the advanced literature on voice behavior, there is a research gap on individual differences and voice behavior.

First, previous studies have mostly focused on trait-like individual differences, which are described as personality characteristics that are stable over time and involve consistent behaviors across situations (Chen et al., 2000b). However, it is argued that motivational elements should be considered in relation to voice behavior because speaking up at work is highly dependent on the motivation to benefit individuals, the organization, or both (Zhou et al., 2017). Thus, we suggest goal orientation, which is a robust individual disposition that influences how people interpret and respond to achievement situations (Dweck & Leggett, 1988), as an antecedent of the voice behavior. Individuals’ goal orientation plays a significant role in the motivation process by organizing their affective, behavioral, and cognitive processes (Gong & Fan, 2006). Unlike general traits such as Big Five personality traits, goal orientation is a composite personality characteristic composed of personality variables and motivational elements, which is more proximal to specific behaviors (Beaubien & Payne, 1999; Vandewalle et al., 2019).

Second, as previous studies on dispositional antecedents of employee voice have particularly concentrated on the main effects of one or more dispositional factors, possible interactive and combined effects of situational factors have been largely neglected. However, the influence of one disposition on an employee’s behavior can be different depending on the context. More specifically, although goal orientation is described as a trait, a central tenet of goal orientation is that engaging in different response behaviors (e.g., adaptive or maladaptive behavior patterns) depends on the situation in an achievement setting (Dweck, 1986). Therefore, it would be imperative to understand the interactions between individual factors and situational cues. Lastly, most prior studies regarding dispositions on voice behavior have used a cross-sectional design. Although earlier research has identified that dispositional factors are a significant predictor of voice behavior, owing to the cross-sectional design, a firm conclusion between the two variables cannot be drawn yet.

Based on the broaden-and-build theory (Fredrickson, 2001) and trait-activation theory (Tett & Burnett, 2003), we propose learning goal orientation (LGO) as a crucial determinant for promotive voice behavior. Learning goal-oriented people are inclined to seek self-development opportunities and perceive challenging tasks as learning opportunities; thus, they are more likely to value the potential benefits of speaking up and exert increased effort to offer suggestions, ideas, and opinions. Furthermore, we suggest that thriving at work plays the role of a mediator in the relationship between employees’ LGO and promotive voice behavior. Studies have shown that individual factors influence employees’ thriving at work because it is a subjective experience and may thus differ among employees even in the same organization (Spreitzer & Porath, 2013). Employees with LGO attempt to develop their competence by acquiring new skills and mastering tasks, and thus experience more energy at work and seek more learning opportunities in the organization. Furthermore, when employees feel energetic and have learning experiences at work, they are increasingly likely to engage in extra roles, such as promotive voice behavior, for improving the current work situation. Therefore, we suggest that learning goal-oriented employees are more likely to experience thriving at work, thereby facilitating them to speak up in their organization.

We further examined one boundary condition for the relationship between employees’ LGO and thriving at work. To derive an improved understanding of the effect of dispositional factors on employees’ attitudes and behavior, it is important to discern the conditions that enhance or mitigate the effects of dispositional factors. Previous studies have indicated that organizational support for employee development is a key contextual factor that influences how employees interpret their work environment (Kraimer et al., 2011). In this study, we suggest intrinsic career growth (ICG), which is a sub-dimension of organizational career growth (OCG), as a key conditional factor that enhances the relation between LGO and thriving at work. OCG refers to employees’ experience of career growth within their current organization (Weng et al., 2010) and ICG indicates the degree to which employees perceive their organization’s support and gratification of intrinsic growth (i.e., career goal progress, professional development). Based on the trait-activation theory (Tett & Burnett, 2003), we propose that the effect of LGO on thriving at work would be enhanced when employees perceive a high level of ICG.

This study provides the following contributions to the voice literature. First, although individual characteristics are considered as antecedents linked to promotive voice behavior (Chamberlin et al., 2017), only few empirical studies have explored the role of employees’ LGO with relation to their voice behavior. Thus, the current study extends our knowledge by examining LGO as a key predictor of promotive voice. Second, our study may offer an extensive explanation of how LGO promotes employees’ voice behavior by investigating the underlying mechanism of thriving at work in the relationship between LGO and voice behavior. Third, by examining the moderating effect of intrinsic OCG, our study provides key information about how the effect of LGO may be improved in certain work environments. Figure 1 presents the proposed research model.

Fig. 1

Proposed model

Theoretical Background and Hypotheses

LGO and Promotive Voice

Dweck and her colleagues (Dweck, 1986; Elliott & Dweck, 1988) originally conceptualized two different goal orientations: (a) learning goal orientation (LGO) which focuses on developing ones’ competence by acquiring new skills and mastering tasks and (b) performance goal orientation (PGO) that emphasizes demonstrating their abilities by gaining favorable judgments from others. Later, VandeWalle (1997) found that performance goal orientation can be partitioned into two dimensions, namely, performance prove goal orientation (PPGO) which focuses on demonstrating one’s competence and gaining favorable judgments from others, and performance avoid goal orientation (PAGO) that is, a focus on avoiding unfavorable judgments about one’s competence.

Learning goal-oriented individuals tend to seek self-development opportunities and take challenging tasks as learning opportunities since they are intrinsically motivated to learn and grow (Payne et al., 2007; VandeWalle et al., 1999). Therefore, individuals high in LGO may have more opportunities for acquiring new knowledge and skills (i.e., learning) and are more likely to feel the energy through self-improvement and growth (i.e., vitality). Furthermore, high LGO people tend to have a “growth mindset (Dweck, 2006)” in which ability is changeable and can be developed so that, they are more likely to set challenging goals and persist in achieving them (Brett & VandeWalle, 1999). Achieving goals and valuing self-improvement may provide more opportunities that reinforce thriving at work. However, performance goal-oriented individuals have a “fixed mindset” in which intelligence and abilities are innate and less changeable; thus, they tend to be more threatened by challenging tasks and less engaged in self-development and learning activities (Dragoni et al., 2009). In addition, different from individuals with LGO—who are more attentive to intrinsically motivating factors, such as the task itself—people with PPGO and PAGO tend to be motivated by extrinsic factors, such as extrinsic rewards (VandeWalle, 1997). In this study, as we are primarily interested in why and how people are more engaged in promotive voice behavior, we focus on LGO.

Learning goal-oriented individuals tend to perceive challenges as learning opportunities and are thus more likely to have favorable attitudes toward proactive behavior to alter the work environment (Chiaburu et al., 2007). Learning goal-oriented individuals are motivated to expand their skills and knowledge through a growth mindset and are thus more likely to focus on challenging and changing the current methods used in their organization. In addition, learning-oriented individuals can adopt diverse learning strategies, such as deep processing (i.e., integrating novel information with prior experience), and can thus easily detect the problems in work units and organizations, thereby leading them to present their opinions and suggestions with the goal to improve the current situation. Learning goal-oriented people are inclined to seek self-development opportunities and perceive challenging tasks as learning opportunities; thus, they are more likely to value the potential benefits of speaking up and exert increased effort to offer suggestions, ideas, and opinions. Some scholars have noted that learning goal-oriented individuals are more likely to be engaged in extra-role behavior, such as voice behavior (Kensbock & Stöckmann, 2021; Ng & Lucianetti, 2018). Recently, Kensbock and Stöckmann (2021) also found that learning goal-oriented individuals are more engaged in voice behavior within organizations.

The Mediating Role of Thriving at Work Between Learning Goal Orientation and Voice Behavior

Drawing from Fredrickson’s (2001) broaden-and-build theory of positive emotions, we suggest thriving at work as the underlying mechanism of the effect of LGO on employees’ promotive voice. The broaden-and-build theory posits that the experiences of positive emotions extend the scope of cognition and actions, which in turn builds beneficial resources for people. According to the theory, positive emotions such as joy, interest, contentment, and love can broaden individuals’ mindsets which allows individuals to explore new ideas, actions, and social bonds, while negative emotions narrow people’s ideas about action. The broadened mindsets lead them toward building enduring intellectual, socioemotional, and psychological resources beneficial for personal growth. Thus, when people experience positive emotions, it may expand the scope of attention, cognition, and the ways of acting, and give them more opportunities to build valuable resources for their development.

Thriving at work (shortened as “thriving”) is defined as a positive psychological state of a joint experience of vitality and learning (Spreitzer et al., 2005). Vitality (as an affective dimension) refers to the positive feeling of energy at work, reflecting a sense of aliveness and high arousal of positive emotions related to the job. Learning (as a cognitive dimension) is the sense that an individual’s work improves through the acquisition and application of newly gained knowledge and skills (Spreitzer et al., 2005). These two dimensions of thriving are independent of each other but should be simultaneously present for employees to thrive. In other words, if an individual feels energetic and alive but does not feel like they have gained new knowledge or skills, they are not thriving. Conversely, if the same individual is learning new things but feels burned out, they are not thriving.

LGO is positively associated with thriving at work. Individuals with high LGO are intrinsically motivated and eager to learn for self-development and personal growth through work (Dweck, 1986; VandeWalle, 1997). More specifically, learning goal-oriented individuals tend to focus more on the task itself and pursue complex tasks without extrinsic rewards; thus, they are more likely to be passionate and energetic in the workplace (vitality). In addition, individuals with high learning orientation constantly seek out new learning opportunities for self-development and engage in adaptive behaviors even when they are confronted with challenging situations; thus, they have more chances to learn and develop skills and knowledge (learning). Accordingly, learning-oriented individuals are more likely to feel a sense of thriving (i.e., vitality, learning). Research has also demonstrated a positive association between LGO and thriving at work (Porath et al., 2012).

Furthermore, thriving at work can result in employees’ promotive voice behavior. When individuals feel alive and energetic at work (vitality), they are able to surpass the formally assigned work roles by putting in additional effort. Therefore, employees thriving at work are more willing to engage in extra roles such as promotive voice behavior, to improve the current work situation. In addition, employees’ learning experiences at work (learning) can lead them to have more ideas and suggestions, thereby allowing them to engage in promotive voice behavior. Some empirical studies have revealed that thriving at work is a significant indicator of employees’ promotive voice behavior (Sheng & Zhou, 2021).

According to the broaden-and-build theory, positive emotions help individuals broaden their physical, intellectual, and social resources, which in turn motivate them to engage in activities that lead them toward enhanced self-development (Fredrickson, 2003; Quinn et al., 2012). Learning goal-oriented individuals are more likely to exert effort to accomplish the task and engage in challenging tasks for personal growth and learning; thus, they feel more satisfied, confident, and hopeful (Pekrun et al., 2006; Pintrich, 2000). In other words, learning goal-oriented individuals focus more on the task itself with growth mindsets and are more likely to feel positive emotions from their work experiences. Such positive emotions may broaden their mental state with more knowledge and energy, which enables employees to engage in more promotive voice behavior. Considering the relationships among employees’ LGO, thriving at work, and voice behavior, we suggest that employees’ LGO will promote thriving at work, thereby increasing their voice behavior. Therefore, we proposed the following hypothesis:

Hypothesis 1: Thriving at work mediates the relationship between employees’ LGO and promotive voice behavior.

The Moderating Role of ICG in the Relationship Between LGO and Thriving

In this study, we propose ICG, which is a sub-dimension of OCG as a moderator in the relationship between LGO and thriving at work. Weng and Hu (2009) proposed that OCG comprises four factors, namely career goal progress, professional ability development, speed of promotions, and compensations. Career goal progress refers to how well the current organization helps employees achieve their career goals, whereas professional ability development refers to how their jobs enable them to acquire new knowledge and experience. In addition, promotion speed refers to the degree to which the organization reinforces accomplishments through promotions, and compensation refers to the likelihood of their salary increase (Weng et al., 2010). OCG reflects both the ICG (career goal progress and professional ability development), and the extrinsic career growth (ECG; promotion speed and compensations) (McElroy & Weng, 2016).

Based on the trait-activation theory (Tett & Burnett, 2003), we propose ICG as a moderator in the relation between LGO and thriving at work. The central idea underlying the trait activation theory (Tett & Burnett, 2003; Tett & Guterman, 2000) is that individuals’ traits are more likely to be activated in response to trait-relevant situational cues. According to the trait activation theory, situational cues are moderators and strengthen the relationship between individual factors and behaviors (Tett & Burnett, 2003). Thus, individuals’ traits such as LGO can be strongly expressed under the conditions that provide relevant cues that require competence for effective performance. In addition, it is argued that individuals thrive more when certain conditions exist in the workplace (Spreitzer et al., 2005). In other words, not all individuals with LGO experience thrive at the same level. Rather, it can differ depending on the contexts in which they work. In this regard, we expect that ICG moderates the relationship between LGO and thriving at work.

Even though the overall perception of OCG helps employees gain learning experience and feel alive, we argue that individuals with LGO increasingly feel a sense of thriving at work when they perceive high levels of ICG rather than ECG. Prior studies have shown that ICG and ECG function differently in predicting outcomes (McElroy & Weng, 2016; Son & Kim, 2021; Weng et al., 2010). For example, it was found that ICG, not ECG, was positively associated with work engagement (Son & Kim, 2021). As individuals with LGO are intrinsically motivated to learn and self-develop (Dweck, 1986; VandeWalle, 1997), the perception of ICG rather than the perception of ECG may make them feel more energized and alive. This is because ICG is closely associated with pursuing and attaining intrinsic goals, while tangible rewards such as compensation could undermine intrinsic motivation (Deci et al., 1999; Kochan, 2002). Therefore, employees’ perception of ICG could serve as a situational cue for individuals with LGO.

Given that ICG indicates the degree to which employees believe their organization supports and gratifies their intrinsic growth (e.g., career goal progress, professional ability development), individuals with LGO perceive such environments as learning opportunities that motivate them to learn more. That is, the perception of organizational support for career growth may trigger individuals with LGO to feel energetic and perceive a sense of learning. However, the relationship between LGO and thriving at work will be weakened when they perceive low ICG. Employees with LGO will be less likely to feel energetic and perceive learning under a situation that provides fewer opportunities for career goal attainment and skill development. Prior research has demonstrated that situational factors (e.g., workplace environment) strengthen or weaken the effect of individual factors such as LGO on employee behaviors (Zhu & Akhtar, 2019; Zia et al., 2020). For example, Zia et al. (2020) have indicated that an empowering environment strengthens the relationship between LGO and self-development. Therefore, we proposed the following hypothesis:

Hypothesis 2: ICG moderates the relationship between employees’ LGO and thriving at work such that this relationship is stronger when employees perceive high ICG (vs. low).

Based on H1 and H2, we proposed a moderated mediation model of the processes linking LGO and promotive voice behavior. The indirect relationship of employees’ LGO with promotive voice through thriving at work will vary based on the ICG level. Individuals with high LGO will be more likely to engage in speaking up at the workplace by experiencing improved energy and learning, particularly when they perceive high ICG in their organization. However, the effect of employees’ LGO on voice behavior through thriving will be weaker when they perceive low ICG. Therefore, we proposed the following hypothesis:

Hypothesis 3: ICG moderates the indirect relationship between employees’ LGO and promotive voice behavior through thriving at work, such that this indirect relationship is stronger among employees with higher ICG.

Method

Research Setting and Participants

Data were collected from Korean employees enrolled in an online panel managed by INVIGHT—a panel company located in Seoul, Korea. The panel company comprised over 180,000 registered members who had agreed to be participants in research surveys on a variety of topics. More than 5000 members of the online respondent pool were randomly selected and received an email invitation including information about the study and a web link to the survey, and were assured that their responses to each questionnaire would remain confidential. The survey was administered with a time lag of four months, in two waves, to minimize potential common method bias. A research design with time separation may aid the identification of causal effects since temporal order with a sufficient time period allows for the unraveling of antecedents and consequences from the theoretically developed links among LGO, ICG, thriving, and voice behavior. In the first wave (T1, May 16–31, 2020), respondents reported on LGO and ICG and provided their demographic information. Four months later (T2, September 16–30, 2020), they reported on thriving and promotive voice behavior. Although there are no specific rules of thumb to determine the appropriate length of the time lags between study waves (Taris & Kompier, 2014), we consider four months as an appropriate interval enough to capture the effect of trait-like LGO on behaviors. In addition, it could minimize the sample attrition caused by a longer time lag between waves. To encourage participation in both surveys, we compensated all respondents with 1600 INVIGHT points for each survey as an incentive for their participation.

A total of 350 employees participated in the survey at T1, but 279 employees completed both surveys. For the possible attrition biases, we examined whether there were significant differences between the T1 and T2 samples, and we found no significant demographic differences (i.e., age, gender, tenure, and organizational status) between samples. Additionally, there was no significant difference in the mean values of the primary T1 variables such as LGO, ICG, PPGO, and PAGO. We also examined mean value differences in the primary T1 variables between those who only participated in T1 and those who participated both times, but there was no significant difference. In the final sample, most respondents were employed on a full-time basis (93.5%) in privately owned firms (78.1%). More than half of the respondents were female (55.2%), married (58.4%), aged 33–50 years (52.3%), with years of tenure between two to eleven (54.4%), and were well-educated (71.7% had at least a bachelor’s degree).

Measures

All items were measured using a five-point Likert scale ranging from 1 (strongly disagree) to 5 (strongly agree).

LGO. LGO was measured with five items developed by VandeWalle (1997). The scale was originally developed in English but translated to Korean and validated by Rhee and Choi (2017) with a Korean sample. A sample item for LGO is “I am willing to select a challenging work assignment that I can learn a lot from.” Each item was evaluated using a five-point Likert scale ranging from 1 (strongly disagree) to 5 (strongly agree). The Cronbach’s α for LGO was .87.

Thriving at Work

Thriving at work was also measured by the Korean version of Porath et al.’s (2012) 10-item thriving at work scale, which was recently validated by Lee and Lee (2020) with a sample of Korean employees. This scale originally consisted of five items for vitality and five items for learning. The sample items are “At work, I find myself learning often” and “At work, I feel alive and vital.” Based on previous studies (e.g., Cullen et al., 2018; Walumbwa et al., 2018), we measured thriving as a whole by combining learning and vitality together. The Cronbach’s α for thriving was .92.

Voice Behavior

Promotive voice behavior was measured using a five-item scale developed by Liang et al. (2012). The Korean version of the scale items was also validated previously with a sample of Korean employees (Dedahanov et al., 2019). A sample item for promotive voice is “Proactively develop and make suggestions for issues that may influence the unit.” The Cronbach’s α for voice behavior was .92.

ICG

ICG including two dimensions—career goal progress and professional ability development—was measured using eight items originally developed by Weng and Hu (2009). First developed in English, its Korean version has been validated by Kim et al. (2016) with a sample of Korean employees. Each dimension has four items. A sample item for career goal progress was “My present job moves me closer to my career goals.” A sample item for professional ability development was “My present job encourages me to continuously gain new job-related skills.” The Cronbach’s α for ICG was .94.

Control Variables

Based on studies about employee voice behavior (Detert & Burris, 2007; Takeuchi et al., 2012; Van Dyne & LePine, 1998), we included four control variables—age, job tenure, gender, and organizational status—in our analyses to minimize their confounding effects on employee voice behavior. As an example, the control for job tenure was used because employees with senior tenure in their job tend to feel more confident speaking up, compared with newcomers (Stamper & Van Dyne, 2001). Age and tenure were measured in years on a continuous scale, whereas a dichotomous scale was used for gender (1 female employee, 2 male employees) and organizational status (1 permanent, 2 contingent). Furthermore, to address the unique effect of LGO on promotive voice, both PPGO and PAGO were included as control variables.

Results

Test of Measurement Model

We employed item parcels rather than individual items to minimize the complexity of the model by creating a more favorable ratio of indicators to sample size. Item parcels for two latent constructs—ICG and thriving—were generated using the item-to-balance method (Little et al., 2002). Specifically, ICG was represented by four-item parcels, with each parcel including two or three items that belong to each sub-dimension of ICG—in this case, career goal progress and professional ability development. After item parceling, the mean score of items belonging to the same parcel was adopted as the indicator for the corresponding latent construct. The same pattern also applied to thriving (i.e., four-item parcels made up of two or three items of each sub-dimension—in this case, vitality and learning). The average variance extracted for each construct was .58 for LGO, .78 for ICG, .65 for thriving, and .70 for promotive voice, providing support for the convergent validity of the measures used in this study.

We then conducted confirmatory factor analyses (CFAs) to verify the distinctiveness of the variables—LGO, thriving, ICG, and promotive voice behavior—using structural equation modeling with AMOS 21. We compared the hypothesized model (the four-factor model) with several alternative measurement models using the chi-square test (χ), comparative fit index (CFI), Tucker-Lewis Index (TLI), standardized root mean squared residual (SRMR) and root mean square error of approximation (RMSEA) to assess the model fit (Cheung & Rensvold, 2002). The minimal thresholds for acceptable SEM model fit are an RMSEA smaller than .08 (.05–.08 reasonable fit, ≤ .05 good fit; Browne & Cudeck, 1993), a SRMR smaller than .10 (≤ .10 reasonable fit, ≤ .05 good fit; Hu & Bentler, 1999) a CFI and TLI are larger than .90 (≥ .90 reasonable fit, ≥ .95 good fit; Hu & Bentler, 1999). As presented in Table 1, the four-factor model demonstrated a better fit (χ (129) = 319.84; CFI = .95; TLI = .94; SRMR = .05, and RMSEA = .07), compared with alternative models. Therefore, the results support the distinctiveness of the four constructs in this study.

Table 1

Results of CFAs

	χ 2	df	CFI	TLI	SRMR	RMSEA	△χ 2(△df)
Hypothesized model	319.84	129	.95	.94	.05	.07
Model 1	791.10	132	.83	.80	.11	.13	471.26***(3)
Model 2	793.93	132	.82	.80	.09	.13	474.09***(3)
Model 3	813.01	132	.82	.79	.12	.14	493.2***(3)
Model 4	860.13	132	.81	. 78	.11	.14	540.29***(3)
Model 5	867.07	132	.81	. 77	.09	.14	547.23***(3)
Model 6	1160.03	132	. 73	. 68	.14	.17	840.19***(3)
Model 7	1258.09	134	.70	.66	.12	.17	938.25***(5)
Model 8	1293.74	134	.69	.65	.13	.18	973.9***(5)
Model 9	1546.29	134	.63	.57	.13	.20	1226.45***(5)
Model 10	1555.20	134	.62	.57	.12	.20	1235.36***(5)
Model 11	1916.91	135	.53	.46	.13	.22	1597.07***(6)

CFI means comparative fit index; TLI Tucker Lewis Index, SRMR standardized root mean squared residual, RMSEA root-mean-square error of approximation. ** denotes significance at the 0.1% level. The values of △χ and △df are differences between the hypothesized model and the other models. The models listed above are described as follows:

Hypothesized model: four factor model (LGO, thriving, ICG, promotive voice)

Model 1: Three factor model combining LGO and promotive voice as one single factor, thriving, ICG

Model 2: Three factor model combining ICG and thriving as one single factor, LGO, promotive voice

Model 3: Three factor model combining LGO and ICG as one single factor, thriving, promotive voice

Model 4: Three factor model combining thriving and promotive voice as one single factor, LGO, ICG

Model 5: Three factor model combining LGO and thriving as one single factor, ICG, promotive voice

Model 6: Three factor model combining ICG and promotive voice as one single factor, thriving, LGO

Model 7: Two factor model combining as LGO, ICG and thriving as one factor, and promotive voice

Model 8: Two factor model combining as LGO, thriving and promotive voice as one factor, and ICG

Model 9: Two factor model combining as LGO, ICG and promotive voice as one factor, and thriving

Model 10: Two factor model combining as ICG, thriving and promotive voice as one factor, and LGO

Model 11: One factor model combining all variables

Although we used four-month time-lagged data to avoid the common method variance problem, data collection relied on employees’ self-reports. Also, some variables were answered at the same point in time. Thus, our data may not be entirely free from the same source bias. Accordingly, we performed an additional analysis using the CFA marker technique (Williams et al., 2010) to examine a possible presence of the common method variance (CMV) in our data. Following the recommendation from Williams et al. (2010), we chose financial strain (Creed & Macintyre, 2001) as a marker variable because of its lowest correlations with the other variables. Financial strain was measured with four items, each rated on a 5-point scale. Sample items were “Do you have serious financial worries?” and “Are you often not able to do the things you like to do because of shortages of money.” Chi-square difference tests comparing the baseline model (χ2 = 433.78, df = 211, p < .001) with the Method-C model (χ2 = 433.30, df = 210, p < .001) was not significant, Δχ2 = .48, Δdf = 1, p = .49, indicating that CMV was not the serious concern in our results.

Descriptive Statistics

Table 1 presents descriptive statistics, correlations, and reliabilities of the study variables.

As shown in Table 2, promotive voice significantly correlates with all three predictors in the expected direction. The same pattern is also applied to the relationships of thriving as a mediator with the other two predictors, r = .41 (p < .01) for LGO, and r = .57 (p = .01) for ICG.

Table 2

Means, standard deviations, and correlations for all variables

Variables	Mean	SD	Min.	Max.	1	2	3	4	5	6	7	8	9	10	11
1. Age	41.41	10.52	24	65
2. Gender	1.45	.50	0	1	−.08
3. Tenure	8.07	7.96	.17	40.25	.45**	−.06
4. Status	1.07	.25	0	1	−.07	.00	.12*
5. PPGO	3.39	.71	1.50	5.00	.00	−.20**	−.13**	−.10	(.83)
6. PAGO	2.86	.73	1.00	4.75	−.34**	.00	−.13*	.05	−.01	(.80)
7. LGO	3.24	.73	1.00	5.00	.20**	−.18**	.02	−.07	.59**	−.38**	(.87)
8. ICG	3.19	.78	1.13	4.88	.14*	−.18**	.22**	.19**	.25*	−.21**	.47**	(.94)
9. Thriving	3.10	.71	1.00	4.90	.30**	−.10	.21**	−.03	.22**	−.20**	.41**	.57**	(.92)
10. PMV	3.33	.79	1.00	5.00	.23**	−.17**	.15*	−.04	.35**	−.31**	.47**	.41**	.50**	(.92)
11. FS	3.31	.94	1.00	5.00	−.02	−.11	−.12*	−.08	.18**	.03	.16**	−.09	−.10	−.04	(.87)

N = 279, Cronbach’s alpha is shown in the parentheses. Gender was dummy coded as male = 0 and female = 1; Status was dummy coded as part-time = 0 and full-time = 1; PPGO performance prove goal orientation, PAGO performance avoidance goal orientation, LGO learning goal orientation, ICG intrinsic career growth, PMV promotive voice behavior, FS financial strain. * p < .05, ** p < .01

Hypothesis Testing

We performed the Hayes’ PROCESS syntax in SPSS with 10,000 bootstraps to test the hypotheses because this method simultaneously tests significance of all the possible paths simultaneously, unlike the traditional piece-meal approaches with a hierarchical multiple regression. To examine the mediating role of thriving at work in the relationship between employees’ LGO and promotive voice behavior, the Hayes’ PROCESS syntax in SPSS (Model 4) was conducted with LGO as an independent variable and promotive voice behavior as a dependent variable while controlling for age, gender, organizational status, tenure, and two goal orientation variables (i.e., PPGO, PAGO). According to Cheung and Lau (2008), the mediations are statistically significant if the bias-corrected (BC) confidence intervals for these indirect effects do not overlap with zero. The result showed that 95% CI (.36, .59) in total effect did not include zero, which indicated that employee LGO is positively related to promotive voice behavior. Moreover, both 95% CI (.08, .22) in indirect effect and 95% CI (.22, .45) in direct effect did not include zero, thereby indicating that thriving at work partially mediated the relationship between LGO and promotive behavior. Therefore, H1 was supported.

To test H2 (moderation) and H3 (moderated mediation) simultaneously, we utilized the PROCESS syntax in SPSS (Model 7). The bias-corrected 95% confidential intervals were constructed around the effect size coefficients to determine their significance. Prior to the analysis, mean-centering was conducted for the independent (LGO) and moderation (ICG) variables. The results of moderated mediation analysis are presented in Table 3.

Table 3

Results of moderated mediation analysis

	Indirect effect (b, boot SE)	95% CI (Lower, Upper)
LGO ➔ Thriving ➔ PMV| ICG = Low	(.02, .03)	(−.0426, .0790)
LGO ➔ Thriving ➔ PMV| ICG = Med	(.05, .03)	(−.0013, .1126)
LGO ➔ Thriving ➔ PMV| ICG = High	(.09, .04)	(.0183, .1665)
Index of moderated mediation	(.04, .02)	(.0010, .0936)

N = 279. LGO learning goal orientation, PMV promotive voice behavior, SE standardized error, CI confidence interval

The results revealed that the LGO x ICG interaction is positively related to thriving at work (b = .11, 95% CI = [.0064, .2115]). To fully understand the specific pattern of relationships at different levels of the ICG, we illustrated the relationship between employee LGO and thriving at work for the values of ICG at one standard deviation above and below the mean (see Fig. 2). Consistent with H2, analyses of simple slopes revealed that for employees with high ICG, the relationship between LGO and thriving at work was significantly positive (b = .23, 95% CI = [06, .39]). The same pattern also applied to those with a mean level of ICG (b = .15, 95% CI = [.01, .28]). However, this relationship was not significant for employees with low ICG (b = .06, 95% CI = [−.09, .21]).

Fig. 2

Moderating effect of ICG on the relationship between LGO and thriving at work

H3 is related to a first-stage moderated mediation—the indirect effect of employee LGO on promotive voice behavior through thriving at work would be conditional, depending on the level of ICG. According to Hayes (2015), the index of moderated mediation should be significant to substantiate the claim for the linear moderated mediation path. In our case, the index of moderated mediation was significant and positive (index = .04, se = .02, 95% bias-corrected CI = [.0010, .0936], see Table 3), thereby suggesting that the indirect effect of LGO on promotive voice behavior through thriving is moderated by ICG. More specifically, employees who experienced high ICG exhibited the positive indirect effect of LGO on promotive voice behavior through thriving at work (indirect effect = .09, 95% CI = [.0183, .1665]), but this pattern was not observed among employees experiencing low ICG (indirect effect = .02, 95% CI = [−.0426, .0790]). This suggests that the mediating effect of thriving at work in the LGO–promotive voice relationship primarily depends on whether employees experience ICG within their current organization. Consequently, H3 was supported.

Discussion

Drawing on the broaden-and-build theory (Fredrickson, 2001) and trait-activation theory (Tett & Burnett, 2003), this study aimed to investigate the mediating effect of thriving at work in the relationship between employees’ LGO and promotive voice behavior and the moderating effect of ICG in the relationship between LGO and thriving at work. Using time-lagged data from South Korea, the proposed relationships among employees’ LGO, thriving at work, promotive voice, and ICG were supported. The results of the study revealed that employees’ LGO was positively associated with promotive voice behavior. In addition, thriving at work mediates the relationship between LGO and promotive voice behavior and ICG moderates the relationship between LGO and thriving at work. Furthermore, ICG also moderates the mediating effect of thriving at work on the LGO–promotive voice behavior relationship such that the mediating effect was only significant when employees perceived high ICG.

Theoretical Implications

This study has some theoretical implications for the voice literature. First, the current study contributes to the voice literature by confirming the effect of LGO on promotive voice behavior. Despite the extensive literature on voice behavior, relatively little attention has been paid to the effect of LGO in predicting employee voice behavior. Consistent with prior research emphasizing the role of LGO (Kensbock & Stöckmann, 2021; Ng & Lucianetti, 2018), the current study highlighted employees’ LGO as an important individual difference factor that affects voice behavior. In particular, as goal orientation includes a motivational component different from other trait-like individual differences, it can better explain individuals’ behavior in specific situations (Vandewalle et al., 2019) and ultimately provide a comprehensive understanding of individual differences in relation to employees’ adaptive behaviors such as voice behavior. By investigating the influence of LGO on promotive voice, particularly using a two-wave design, the current study enriches voice literature.

Second, by examining the underlying mechanism of thriving at work on the relationship between employees’ LGO and voice behavior, this study provides a fuller explanation regarding why and how learning goal-oriented employees are able to speak up at work. Based on the broaden-and-build theory (Fredrickson, 2001), this study found that LGO, both directly and indirectly, enhanced voice behavior through thriving at work. In other words, thriving at work mediates the relationship between LGO and promotive voice. As learning goal-oriented individuals focus more on the task itself and have a growth mindset, they are more likely to feel positive emotions such as joy and interest from their work experiences. Such positive emotions may broaden their states with more knowledge and energy, which enables employees to engage in more promotive voice behavior. Consistent with the broaden-and-build theory, this finding verifies the notion that thriving at work can mediate the relationship between LGO and employees’ promotive voice behavior. However, the direct impact still remains significant and the effect size is large compared to the indirect effect. This suggests that there might be alternative mediators not included in the model, such as self-regulatory mechanisms (Johnson et al., 2011), psychological capital (Macey & Schneider, 2008), and job crafting (Matsuo, 2019).

Third, this study advances thriving literature by examining the conditions wherein employees are more likely to experience thriving at work. By drawing from the trait-activation theory (Tett & Burnett, 2003), the results of this study have revealed that learning goal-oriented employees are more likely to feel energized and improve their skills at work, particularly when they perceive ICG. This result is important because it suggests that organizational contexts play a critical role in activating employees to thrive at work. This finding is consistent with previous research suggesting that situational factors such as workplace environment strengthen or weaken the effect of individual factors such as LGO on employee behaviors (Zia et al., 2020). In particular, this study highlights the role of ICG for enhancing employees’ thriving at work and furthering their voice behavior, thereby answering the call by Weng and Zhu (2020) for more research on the different roles of ICG and ECG.

Practical Implications

This study has managerial implications for organizations. The results revealed that LGO, directly and indirectly, promoted voice behavior. Therefore, HR managers need to consider employees’ LGO as a key individual characteristic when they hire new employees. Furthermore, organizations may need to put more effort to develop HR practices and programs, such as mentoring, coaching, and professional development, to enhance employees’ learning goals (Seijts & Latham, 2005). As some researchers have emphasized, learning goals may be trained and enhanced through organizational practices and programs (Seijts & Latham, 2005). Furthermore, our findings highlight thriving at work as a key mechanism for promotive voice. Therefore, organizations should create an environment that helps employees to thrive at work. In particular, organizations need to be aware of the fact that the effect of employees’ LGO on thriving is strengthened when employees perceive organizational support for career growth. Therefore, HR managers should provide diverse career opportunities that support an employee’s professional development and career goal progress. Designing individualized career paths and providing interventions for employees would help them to achieve their career goals, thereby increasing their thriving experiences at work.

Limitations and Future Research

Our study has several limitations that must be acknowledged. First, this study applied a two-wave design with a four-month time lag to examine the moderated mediation model. Even though the current design is better than a cross-sectional design, the optimal design to test the (moderated) mediation model involves collecting data across three or more time points (Cole & Maxwell, 2003). In our study, both the mediator (i.e thriving) and the outcome variable (i.e., PMV) were measured at the same time (T2), so that the causal relationship between thriving and PMV could be temporally unclear. Furthermore, even though the current study considers voice behavior as a consequence of a psychological state, there are other studies that suggest thriving or other closely related variables (e.g., engagement) as an outcome of voice (Chen et al., 2020a; Cheng et al., 2013; Rees et al., 2013). Therefore, future researchers should consider adopting a cross-lagged design with multiple time points to examine the reverse-causal or reciprocal relations between them.

Second, our data on employee voice behavior are self-reported. Employees might be the best source for us to study their behaviors because no one knows better than the individuals who actually perform them (Ng & Lucianetti, 2018). Nonetheless, the employees’ self-rating of promotive voice behavior may include even the ‘relatively trivial’ suggestions that might not be recognized by supervisors or others. In addition, some employees might be more self-critical, and may thus underrate their performance when compared with the ratings of their supervisors. To fully address this limitation, future research needs to consider including multiple raters (e.g., direct supervisors) to extend the generality of our findings by comparing self-rated and supervisor-rated promotive voice while also avoiding common source bias.

Third, there is a potential cultural impact on voice behavior as national culture is considered critical in shaping employees’ attitudes and behaviors at the workplace (Parboteeah et al., 2013). South Korea is characterized by a higher power distance culture (Hofstede et al., 2010) and Korean organizational cultures have also been shown to be hierarchical (Dastmalchian et al., 2000; Hong et al., 2016). In such a hierarchical organizational culture, where employees are expected to comply with and accept their supervisors’ directives, exhibition of voice behavior is likely to be perceived as highly risky behavior or less welcomed by authority figures. Therefore, it is worthy to explore what leadership or managerial styles can moderate and even reverse the adverse effects driven by hierarchical organizational culture on employee promotive voice.

Finally, we did not include the other form of voice, such as prohibitive voice, in this study. Prohibitive voice is the expression of concern about work-related issues that are possibly harmful to the organization (Liang et al., 2012). Prohibitive voice serves an important function as promotive voice, benefitting the organization because it creates a healthy environment by preventing potentially harmful situations from occurring. Therefore, future researchers need to consider including different types of voice behavior in relation to LGO and thriving. Similarly, this study only focused on the mechanisms explaining how LGO leads to promotive voice, without taking other types of goal orientation (e.g., performance-avoidance goal orientation) into account. Considering that performance-avoidance goal orientation (VandeWalle, 1997) is more strongly related with prohibitive voice (i.e., the expression of concern about work-related issues that are harmful to the organization; Liang et al., 2012), rather than focusing solely on promotive voice (Chamberlin et al., 2017), future research can benefit from further investigation of different potential moderators and mediators in explaining the various relationships between goal orientation and employee voice.