THE RELATIVE IMPORTANCE OF CONTEXT AND INDIVIDUAL DIFFERENCES IN PREDICTING PROACTIVE SERVICE PERFORMANCE

The current study addresses the gap in how context and individual differences help in developing a proactive workforce in the hospitality industry. Based on the model of proactive motivation and selfdetermination theory, this study investigates the simultaneous impact of interpersonal leadership and learning goal orientation through the mediating role of employee engagement. The study sample (N = 438) is collected from frontline hospitality employees in Malaysia. Results of partial least squares structural equation modeling (PLS-SEM) show that both context (i.e., interpersonal leadership) and individual differences (i.e., Learning Goal Orientation) are significant predictors of employee engagement and, in turn, of proactive service performance. Moreover, engagement is found as a significant mediator between both the predictors and the criterion. Interestingly, however, in comparison with interpersonal leadership, learning goal orientation is noticed as a substantial predictor of employee engagement and proactive service performance. Implications of the findings for research and practice are discussed accordingly.


INTRODUCTION
In the times of hyper-competition in tourism services industry, with digital platforms (e.g., Airbnb) disrupting conventional hospitality practices (Guttentag & Smith, 2017), it is becoming more and more difficult for the hard-core hotel sector to earn its well-deserved share in the continuously growing global travel and tourism industry ("WTTC Global Travel & Tourism,"2016). However, as people remain at the very center of travel and tourism business, the hotel industry has a natural competitive advantage of having 'human touch' to satisfy the diverse needs and expectations of its customers through personal attention provided by its frontline employees ("The individual is at the,"2017). In doing so, along with strategies to compete on the digital front, hotel industry managers need to concentrate on the personalized service for customer satisfaction because, according to Accenture strategy, the unpleasant service experience is a prominent reason of customer withdrawal (Quiring, De Angelis, & Gasull, n.d.). Moreover, in today's always-on age of the internet, only premier customer experience can guarantee the necessary competitive advantage a hospitality organization needs ("Service is the new sales,"n.d.). Nevertheless, the question remains about how to achieve customer satisfaction for hospitality business success?
Proactively motivated frontline employees can ensure the above mentioned competitive advantage for the hospitality industry by taking the initiative in providing excellent guest experience (Vachon, 2013). It is so because, on the one hand, proactive people do not follow the status quo and take the initiative to achieve a self-determined future for not only themselves but for others too (Grant, n.d.), on the other hand, they can inspire bystanders as well for taking self-started action to thrive (AngelaN, 2019). In other words, proactivity is a commendable phenomenon that makes people forward-looking and active in foreseeing problems to deal with and opportunities to identify beforehand (Grant & Ashford, 2008). This assertion is backed by literature on the positive outcomes associated with proactivity (see for review, Fuller & Marler, 2009;Thomas, Whitman, & Viswesvaran, 2010). Moreover, empirical evidence suggests that proactive behavior can be developed (Campos et al., 2017). Hence, it is likely that the proactive behavior of frontline hospitality employees might have the potential to turn dreadful business situations around for the hospitality industry. Following this, hospitality researchers have recently started giving importance to proactive service performance by investigating well-thought research models to find best possible predictors of the phenomenon (e.g., Chen, Lyu, Li, Zhou, & Li, 2016;Lyu, Zhou, Li, Wan, Zhang, & Qiu, 2016;Raub & Liao, 2012;Wu, Chen, Lee, & Chen, 2016).
However, despite an increasing trend of empirical research on proactivity (Cangiano, Bindl, & Parker, 2017), the focus of organizational researchers linger, by and large, on reactive forms of employee behavior . More specifically, there is a shortage of empirical evidence in services literature on proactive employee behaviors (Raub & Liao, 2012). Furthermore, even in the available hospitality literature, there remains a scarcity of research models simultaneously considering both contextual and individual factors in predicting proactive service behaviors (Hong, Liao, Raub, & Han, 2016). Thus, it leads to a lack of understanding about the relative importance of situational and personal antecedents of proactive service performance in hospitality employees. This gap in the literature is devastating for research and practice and demands the attention of hospitality researchers to investigate the "what" and "how" of workplace proactivity in frontline employees. Therefore, addressing this paucity in hospitality literature might have significant implications for the industry. Consequently, based on theories discussed later, we developed a predictive model of proactive service performance (see figure 1) intending to simultaneously test the relative importance of interpersonal leadership and learning goal orientation through employee engagement as a mediating mechanism. Investigating the role of well-thought specific mechanisms in empirical research with mediation is vital for theory development Rungtusanatham, Miller, & Boyer, 2014).
Hence, based on the model of proactive motivation (Parker, Bindl, & Strauss, 2010) complemented by selfdetermination theory (SDT: Deci, Olafsen, & Ryan, 2017), the primary purpose of this research is to investigate a predictive model of proactive service performance to find the relative significance of interpersonal leadership and learning goal orientation through employee engagement. In doing so, we believe that the results of this study might contribute to theory and practice in unique ways. First, the current study is one of the few studies simultaneously investigating contextual and individual factors in a mediation model predicting workplace proactivity. We expect that the results of this study might add significant value in employee performance literature, especially in hospitality research. Second, to the best of our knowledge, this is the first study in hospitality literature integrating literature from the model of proactive motivation, self-determination theory (SDT), and employee engagement in developing a mediation model of proactive service performance.
Consequently, the results of this study might guide hospitality researchers on the fundamental role of well-established theory as an authentic body of knowledge in theorizing relationships in the quantitative way of doing research . Third, from a practice point of view, the results of this study might provide unique insight on the importance of growth mindset in the form of learning goal orientation as an individual difference for enhancing employee engagement and, in turn, proactive service performance. Finally, as motivational research has the direct practical value for real-world business issues (Vansteenkiste & Gagné, 2013), the results of the current study might be of practical importance for hospitality managers in developing a proactive frontline workforce.

LITERATURE REVIEW
Proactive service performance is the phenomenon of interest in this study and is defined as "individual's self-started, long term oriented, and persistent service behavior that goes beyond explicitly prescribed performance requirements" (Rank, Carsten, Unger, & Spector, 2007, p. 366). This inclusive form of proactive service performance definition was developed exclusively for the frontline service sector employees. On the one hand, this definition is consistent with the generally accepted rule of thumb for any action to be called proactive, i.e., self-started, change-oriented, and future-focused (Parker et al., 2010). On the other hand, it is consistent with the literature advocating proactivity as a wide-ranging form of employee behavior, that is, beyond segregation into in-role and extra-role (cf. Carpini & Parker, 2017;Grant & Ashford, 2008). Besides, hospitality researchers have been regularly using this definition in the operationalization of proactive service behavior of frontline hospitality employees (e.g., Chen et al., 2016;Lyu et al., 2016;Raub & Liao, 2012).
Interpersonal leadership (IL) is conceptualized in this study as a distal contextual predictor of proactive service performance. This higher-order construct is an inclusive leadership style formed by three distinct but theoretically related constructs, i.e., transformational leadership, interpersonal justice, and informational justice (cf. Hansen, Byrne, & Kiersch, 2014). Thus, the discrete interpersonal leadership dimensions collectively represent support, empathy, care, and respect for subordinates. While dimensions form interpersonal leadership, it is appropriate to mention the evidence on the relationship between transformational leadership and proactive employee behaviors (e.g., Den Hartog & Belschak, 2012;Schmitt, Den Hartog, & Belschak, 2016;Wu & Parker, 2011). Similarly, recent literature acknowledges the intersection between leadership and organizational respect (e.g., Al-Atwi, 2018; Rogers & Ashforth, 2017).
Learning goal orientation (LGO) is theorized as an individual distal predictor of proactive service performance in the current study. This construct is based on the growth mindset, which is one of the dimensions of Carol S. Dweck's (1986) phenomenal concept, mindset. People with a growth mindset believe that intelligence is a malleable characteristic of human nature and new skills to achieve challenging individual and/or professional goals can be learned. Accordingly, learning goal orientation is defined as an employee's dispositional desire to learn new things to increase competence and achieve mastery in a given activity (Button, Mathieu, & Zajac, 1996;Dweck, 1986). There is minimal empirical evidence available in the management literature (e.g., Chughtai & Buckley, 2011;Yean, Johari, & Yahya, 2016) conceptualizing LGO in employee performance studies. However, none of these studies were conducted in the hospitality employees' context. Furthermore, both of these studies used LGO as a mediating mechanism between work engagement and employee performance rather than using it as a distal individual dispositional factor predicting the active form of employee behaviors.
Employee engagement is theorized as a motivational psychological state of an employee and is defined, based on previous literature (i.e., Kahn, 1990;Rich, Lepine, & Crawford, 2010;Shuck, Osam, Zigarmi, & Nimon, 2017), as an employee's positive, active, work-related psychological state of motivation represented by the simultaneous investment of his/her cognitive, emotional, and physical energies in performance outcomes. Only a handful of studies are available on the relationship between engagement and proactive service performance (e.g., Li, Chen, Lyu, & Qiu, 2016;Maria Stock, Jong, & Zacharias, 2017).

THEORY AND THE HYPOTHESES DEVELOPMENT
As mentioned above, a mediation model like the one developed in the current study (see Figure 1) is vital for theory and knowledge development. The foundation for such a framework, however, comes primarily from well-established theories because "strong theoretical argument combined with good measurement" is the key for mediation models (Hayes & Scharkow, 2013, p. 1918. Moreover, mediation models explain that "how" distal predictors impact the criterion through a mechanism (MacKinnon, 2011). Following this line of action, we developed our research framework based on the descriptive model of proactive motivation (Parker et al., 2010), explaining the hypothesized relationships primarily by the tenets of selfdetermination theory (SDT: Deci et al., 2017). According to the basic theoretical framework of the model of proactive motivation, context and individual differences are the distal predicting factors of proactive employee behavior through proximal motivation mechanisms. Self-determination theory (SDT), being a well-established individual-level motivation theory, logically explains this process in detail by putting the condition of satisfaction of basic psychological needs for autonomy, relatedness, and competitiveness for employees' motivation to thrive in individual and professional life Vansteenkiste, Aelterman, De Muynck, Haerens, Patall, & Reeve, 2018). In other words, it is the contextual and individual factors that satisfy or thwart the basic psychological needs of employees and, in turn, motivate them for premier performance.
Accordingly, our research theorizes interpersonal leadership (IL) as a contextual factor satisfying basic psychological needs to make frontline hospitality employees autonomously motivated for proactive service performance. It is so because interpersonal leadership signals support, care, empathy, and respect in satisfying basic psychological needs. Besides, we theorize learning goal orientation as a theoretically convincing individual factor. According to the previous literature in education psychology (e.g., Hochanadel & Finamore, 2015;Yeager et al., 2016), a growth mindset (i.e., learning goal orientation) is a highly effective predictor of improved academic performance. So, learning goal orientation is theorized in the current study as a personal resource, having huge potential to autonomously motivate frontline hospitality employees for proactive service behavior through fulfilling their basic psychological needs. It posits the innate desire of frontline hospitality employees to thrive and, in turn, satisfying their basic psychological needs to make them psychologically engaged to behave proactively. At this point, however, we do not prioritize context or individual differences, because only the results of the study would reveal the relative impact of both the distal factors, i.e., interpersonal leadership and learning goal orientation. Furthermore, as employee engagement is like the proxy from self-determination theory's autonomous motivation perspective (cf. Meyer & Gagne, 2008), it is theorized as the mediating mechanism connecting both the distal predictors and the criterion in the current study.
Hence, drawing on the above, we hypothesize: H1: Interpersonal leadership positively predicts employee engagement.
H2: Learning goal orientation positively predicts employee engagement.
H3: Employee engagement positively predicts proactive service performance.
H4: Employee engagement mediates the relationship between interpersonal leadership and proactive service performance.
H5: Employee engagement mediates the relationship between learning goal orientation and proactive service performance.

Participants and Procedures
Frontline employees of the hospitality industry in Malaysia are the target population of this study, and the sample (N = 438) is collected by the respondents recruited from LinkedIn and Facebook through a cross-sectional self-report online survey. The term frontline is used as a generic reference for all those hotel employees who come into regular contact with the hotel guests for customer/guest service (cf. Karatepe & Karadas, 2016;Seo, Nahrgang, Carter, & Hom, 2018). Non-probability sampling method (purposive sampling) of this study is suitable in the current context in several ways. First, the researcher could not get the full list of frontline hospitality employees in Malaysia, which is the necessary condition for sampling frame in probability sampling (Hulland, Baumgartner, & Smith, 2018;Memon, Ting, Ramayah, Chuah, & Cheah, 2017). However, as the unavailability of a sampling frame is a usual problem in social science research, this issue can be solved with an adequate sample size in non-probability sampling methods (Rowley, 2014). Second, the purposive sampling method is appropriate for the studies targeting loosely defined big populations having specific characteristics (Etikan, Musa, & Alkassim, 2016), as is the case in this study targeting the frontline employees of the hospitality industry in Malaysia. Third, similar to the utility and increasing trend to rely on different online crowdsourcing platforms (e.g., Amazon's MTurk) for the recruitment of research respondents (Chambers, Nimon, & Anthony-McMann, 2016), LinkedIn and Facebook are two relevant social media platforms with billions of users connected as per their professional and individual interests. As a result, seeing the vast potential for knowledge development, academic researchers have recently been started acknowledging both LinkedIn and Facebook as a good source for data collection in fundamental academic research (e.g., Kosinski, Matz, Gosling, Popov, & Stillwell, 2015;Roulin & Levashina, 2019;Van Wingerden, Berger, & Poell, 2018). Finally, we used a screening question "Is customer or guest service your core job responsibility?" as well to make sure the aptness of responses collected.

Measures
All the measures used in this study are adapted from reliable sources of literature and used a 7-point Likert scale with anchors, strongly agree (7) to strongly disagree (1) unless mentioned otherwise.
Proactive service performance is measured with a 7-items scale (Rank et al., 2007). Sample items include, "I proactively share information with customers to meet their service needs" and "I actively create partnerships with other service employees to better serve customers." Employee engagement is a higher-order reflective construct and is measured with 12-items job engagement scale (JES: Rich et al., 2010), measuring cognitive, emotional, and physical engagement with 4-items each. Sample items include, "At work, my mind is focused on my job" for cognitive engagement, "I am enthusiastic in my job" for emotional engagement, and "I devote a lot of energy to my job" for physical engagement.
Interpersonal leadership is a reflective-formative higher-order construct (cf. Becker, Klein, & Wetzels, 2012) measured with 7-items global transformational leadership (GTL) scale for transformational leadership (Carless, Wearing, & Mann, 2000), 3-items for interpersonal justice and 3-items for informational justice from the abridged measure of organizational justice (Hansen, Byrne, & Kiersch, 2013). Sample items include, "My supervisor communicates a clear and positive vision of the future" for transformational leadership, "My supervisor treats staff in a polite manner" for interpersonal justice, and "My supervisor explains the procedures thoroughly" for informational justice. These items are measured on a 7-point Likert scale with anchors from Always (7) to Never (1). Besides, a global item on interpersonal leadership (i.e., Overall my supervisor holds and practice good interpersonal leadership skills) is also collected on the same scale and anchors.
Learning Goal Orientation is measured with an 8-items scale (Button et al., 1996). Sample items include, "The opportunity to do challenging work is important to me" and "I do my best when I am working on a difficult task." PLS Marker Variable is measured with three items (Lin, Huang, & Hsu, 2015) including, "Once I have come to a conclusion, I am not likely to change my mind," "I do not change my mind easily," and "My views are very consistent over time."

DATA ANALYSIS
To avoid "common errors in data analysis" (Green, Tonidandel, & Cortina, 2016, p. 14), such as stepwise mediation analysis of Baron and Kenny (1986), we used structural equation modeling (SEM) for a robust simultaneous statistical analysis of the research model of this study. More specifically, keeping in mind the predictive nature of the research model, partial least squares structural equation modeling (PLS-SEM) is used. Because, PLS-SEM, while accounting for measurement error, is a composite based exploratory structural modeling method, more suitable for prediction and theory development Richter, Sinkovics, Ringle, & Schlägel, 2016). Furthermore, PLS-SEM is selected because it is a constructive multivariate data analysis technique for better results (Hair, Ringle, & Sarstedt, 2013) in various fields of social sciences and business research (Hair, Sarstedt, Hopkins, & Kuppelwieser, 2014;Henseler, 2016) such as information systems (Hair, Hollingsworth, Randolph, & Chong, 2017), knowledge management (Cepeda-Carrion, Cegarra-Navarro, & Cillo, 2019), and tourism (do Valle & Assaker, 2016). Furthermore, recent literature endorsed PLS-SEM as the preferred method of analysis for complex and prediction oriented statistical models in hospitality and human resource management (HRM) research (Ali, Rasoolimanesh, Sarstedt, Ringle, & Ryu, 2017;Ringle, Sarstedt, Mitchell, & Gudergan, 2018).

RESULTS
SmartPLS 3.2.8 (Ringle, Wende, & Becker, 2015) is used as a comprehensive PLS-SEM software program  in testing hypothesized relationships of the research model of this study. Its outer model analysis is to test the factor structure of the study constructs, and the inner model analysis is to find out the status of hypothesized relationships between the constructs.

DESCRIPTIVE STATISTICS
Data was recorded in IBM's statistical package for the social sciences version 22 (SPSS) for initial analysis and descriptive statistics. First of all, as per the demographic information (see Table 1) of the collected study sample, out of 438 respondents, 230 were male, and most of the respondents fall into the age category of 21 to 30 years. Moreover, in the current experience section, most respondents (129) fall in the category of 1 to 3 years of experience. Besides sample characteristics, reporting descriptive statistics of the study is encouraged in literature as a best practice (Hair, Hollingsworth, et al., 2017). So, Table 2 represents mean, standard deviation (SD), and correlation matrix as descriptive statistics of the study variables.

Reliability and Convergent Validity
Proactive service performance and learning goal orientation are unidimensional reflective constructs, while employee engagement is a multi-dimensional Type-I reflective-reflective construct (Becker et al., 2012). Moreover, transformational leadership, interpersonal justice, and informational justice are the firstorder reflective dimensions of the higher-order Type-II reflective-formative construct, interpersonal leadership. In stage one, while assessing measurement models of the study, we used Mode-A repeated indicators method for employee engagement (Type-I, reflective-reflective) and Mode-B repeated indicators method for interpersonal leadership (Type-II, reflective-formative) in the higher-order constructs of our research model (Sarstedt, Hair, Cheah, Becker, & Ringle, 2019).
Outer model analysis for all the reflective constructs, as is shown in Table 3, depicts satisfactory results with factors complying cut-off values for internal consistency (i.e., Cronbach alpha ≥ 0.70 and compositereliability-CR ≥ 0.70), and convergent validity (i.e., item-loadings ≥ 0.708, average variance extracted-AVE ≥ 0.50) (Hair, Hult, Ringle, & Sarstedt, 2017). However, even though loadings for two items of proactive service performance (PB1=0.667, PB3=0.668) are less than 0.708, it is acceptable because both item loadings are above 0.60 and average variance extracted from all seven items (AVE=0.562) is more than 0.50 . Moreover, as it is suggested to establish higher-order construct level reliability-validity as well , we found adequate measures for reliability (CR = 0.93) and convergent validity (AVE = 0.82) for our reflective-reflective higher-order construct, employee engagement.
Similarly, the outer model analysis of our only higher-order formative construct, interpersonal leadership, fulfills all three construct level conditions of validity. First, as shown in Table 4 and graphically represented in figure 2, redundancy analysis (β = 0.871) of the construct confirms its convergent validity by having a path coefficient of more than 0.70 (Hair, Hult, et al., 2017). Second, as is evident from table 4, weights of the first order dimensions forming interpersonal leadership are significant, i.e., transformational leadership (β = 0.63, p ˂ .001), interpersonal justice (β = 0.25, p ˂ .001), and informational justice (β = 0.24, p ˂ .001). Finally, the variance inflation factor (VIF) measures for all three dimensions of the higher-order formative construct are considerably less than 5, confirming the validity of the dimensions forming this construct without any multicollinearity issue. Hence, the satisfaction with the conditions of a formative higher-order construct confirms that interpersonal leadership is a valid reflective-formative hierarchical construct (Hair, Hollingsworth, et al., 2017).

Discriminant Validity
Discriminant validity is an important measure of the degree of accuracy to which items measure their respective constructs. In other words, it "ensures that each construct is empirically unique and captures a phenomenon not represented by other constructs in a statistical model" (Franke & Sarstedt, 2019, p. 430).
There are different measures of discriminant validity available, but Fornell-Larcker (Fornell & Larcker, 1981) and Hetrotrail-Monotrait ratio of correlations (HTMT: Henseler, Ringle, & Sarstedt, 2015) are of significant importance. However, the latter is a comparatively new and preferred way of testing discriminant validity in PLS-SEM, while the former has historical value as a traditional measure of discriminant validity (Hair, Risher, Sarstedt, & Ringle, 2019 Hence, according to the Fornell-Larcker criterion, as shown in Table 5, the discriminant validity of the constructs is established because the square root of the AVE (given on the diagonal line) is larger than the shared variance between variables . Likewise, as shown in Table 6, discriminant validity is established from the HTMT0.85 criterion as well because all values are less than HTMT0.85. Additionally, bias-corrected confidence intervals do not contain the value of 1 in any of the constructs (Henseler et al., 2015;Ramayah et al., 2018). So, the discriminant validity of the constructs of this study is established.

Common Method Variance
As the single-source cross-sectional design of this study is prone to common method bias (Podsakoff, MacKenzie, & Podsakoff, 2012), we employed a marker variable approach to deal with this issue. There is a variety of marker variable approaches available in the literature (see for review, Tehseen, Ramayah, & Sajilan, 2017;Williams, Hartman, & Cavazotte, 2010). We, however, used the PLS marker variable approach of Rönkkö and Ylitalo (2011). This marker variable approach was used in Lin et al. (2015) with three marker items to create a method factor. We adopted these three marker items in our study survey (see Measures for items). Further, a method factor was created and added as an independent variable in the baseline study model predicting both endogenous variables of the study, i.e., proactive service performance and employee engagement. However, results revealed an insignificant relationship between method factor and both endogenous factors. Moreover, no significant change in R2 values of the endogenous variables was observed. As a result, analysis of the PLS marker variable found no common method variance issue in the current study.

Control Variables
Following previous PLS literature on testing control variables (Felipe, Roldán, & Leal-Rodríguez, 2017), we examined gender, age, and experience with the current organization as potential covariates of proactive service performance, by adding all three of these categorical variables into the baseline path model. By doing so, we tested their impact on the dependent variable. The results of the model with control variables, however, show that all three variables are neither statistically significant nor cause any significant R2 change in proactive service performance (see Appendix). Therefore, the insignificant potential covariates were dropped from the baseline research model for further analysis (Bernerth & Aguinis, 2016).

STRUCTURAL (INNER) MODEL ANALYSIS
We used latent variable scores from stage one for the inner model analysis in stage two for testing model robustness as well as the status of hypothesized relationships of the study. In doing so, we used bootstrapping with 1000 subsamples to test the structural model. First of all, as shown in SmartPLS structural model analysis output (see Figure 3), results in Table 7 indicate that theorized model of the study is statistically meaningful because both the endogenous variables accounted for large total variance (i.e., coefficients of the determination); R2 = 0.628 for employee engagement and R2 = 0.349 for proactive service performance (Cohen, 1988(Cohen, , 1992. Similarly, positive values of the Stone-Geisser's measure (Q2) for both endogenous constructs [employee engagement (Q2 = 0.485), proactive service performance (Q2 = 0.337)] established the predictive relevance of the research model of this study (Hair, Hult, et al., 2017).
Secondly, results (see Table 7) of the hypothesized direct relationships were supported such that interpersonal leadership significantly predicted employee engagement (IL→EE: β = 0.141, p = 0.001, tvalue = 3.489, 95% CI [0.059, 0.210]), learning goal orientation also significantly predicted employee engagement (LGO→EE: β = 0.70, p ˂ 0.001, t-value = 18.643, 95% CI [0.627, 0.766]), and employee engagement significantly predicted proactive service performance (EE→PSP: β = 0.591,p ˂ 0.001,tvalue = 14.476,95% CI [0.497,0.660]). Thus hypotheses 1, 2, & 3 are supported. While simultaneously considering context and individual differences, the results are interesting such that, in comparison to interpersonal leadership, learning goal orientation shows a large direct impact on employee engagement. More specifically, one standard deviation change in learning goal orientation predicts 0.7 standard deviation increase in employee engagement (i.e., 70% of the total variance accounted for EE), whereas one standard deviation change in interpersonal leadership predicts 0.141 standard deviation increase in employee engagement (i.e., only 14% of the total variance accounted for EE). It is significant because employee engagement, in turn, directly predicts proactive service performance with a large impact, such that one standard deviation change in EE predicts a 0.591 standard deviation increase in proactive service performance (i.e., 59% of the total variance accounted for PSP).
Finally, indirect relationships of the study were also found significant, such that employee engagement significantly mediates the relationship between interpersonal leadership and proactive service performance (IL→EE→PSP: β = 0.083, p = 0.001, t-value = 3.287, 95% CI [0.035, 0.131]), as well as the relationship between learning goal orientation and proactive service performance (LGO→EE→PSP: β = 0.413, p ˂ 0.001, t-value = 10.547, 95% CI [0.335, 0.490]). Consequently, confirming the acceptance of hypotheses 4 & 5. Results of the indirect impact of both the exogeneous predictors on proactive service performance through employee engagement are of significant value because learning goal orientation is found to be the large indirect predictor of proactive service performance with more than 40% expected impact on the criterion, whereas interpersonal leadership indirectly predicts the criterion with only around 8% expected impact.

DISCUSSION
Based primarily on the descriptive model of proactive motivation (Parker et al., 2010) and explained through the tenets of self-determination theory (SDT: Deci et al., 2017), our study investigated the relative importance of contextual and individual difference factors in predicting proactive service performance through employee engagement, in the context of frontline hospitality employees. More specifically, this study is one of the first attempts simultaneously testing theory-based situational and personal variables in predicting proactive service performance. Moreover, using employee engagement as an intervening mechanism in the research model of this study from the autonomous motivation of the self-determination theory perspective is also a rare combination found in hospitality employee behavior research.
The results indicate that both interpersonal leadership and learning goal orientation significantly predict employee engagement and, in turn, proactive service performance. Moreover, employee engagement is found as a significant mediator between the predictors and the criterion. Though only a handful of studies have simultaneously tested contextual and individual factors in predicting proactive employee behaviors (e.g., Bettencourt, 2004;Parker, Williams, & Turner, 2006), the findings of the current study are congruent with the limited available evidence on the positive relationship between leadership and proactive behavior (Schmitt et al., 2016), as well as the positive relationship between dispositional factors and change-oriented citizenship behaviors (Crant, Hu, & Jiang, 2016;Crant, Kim, & Wang, 2011). Similarly, the significance of employee engagement as a mediator in the current study confirms the findings of a recent meta-analysis (Ng, 2017), suggesting engagement as one of the significant psychological mediators between contextual factors (transformational leadership in this case) and the multiple forms of individual performance. Besides, the findings also confirm another more relevant metaanalysis (Marinova, Peng, Lorinkova, Van Dyne, & Chiaburu, 2015), indicating employee engagement as a significant mediating process between both contextual and individual factors and the change-oriented employee behaviors.
An interesting finding, however, is the substantial direct and indirect impact of learning goal orientation on employee engagement and proactive service performance. It shows the relative importance of an individual's growth mindset (learning goal orientation) while simultaneously considering a contextual factor, interpersonal leadership, in predicting proactive service behavior of frontline hospitality employees. Despite scant evidence on the simultaneous assessment of contextual and individual factors in one research model, the massive impact of learning goal orientation on employee engagement and proactive service performance found in the current study is consistent with available studies where personal factors have been found relatively more significant than the contextual factors (cf. Parker et al., 2006). Moreover, because the findings of the current study provide the initial evidence on learning goal orientation as a highly significant dispositional predictor of employee engagement and workplace proactivity, it is hard to compare its results with previous literature. As per the authors' best of knowledge, Bettencourt (2004) is the only study in organizational literature used learning goal orientation in their research model predicting change-oriented employee behaviors. Nevertheless, even in this study, based on achievement motivation theory (Dweck, 1986), learning goal orientation was theorized as a direct predictor of change-oriented citizenship behaviors. Therefore, the results of the current study on the simultaneous effect of interpersonal leadership and learning goal orientation on employee engagement and, in turn, on proactive service performance provide evidence more relevant to the model of proactive motivation, which is the descriptive foundation of the current study. These findings might have significant implications for research and practice.

IMPLICATIONS FOR RESEARCH AND PRACTICE
The findings of our study are of significant importance for literature in numerous ways. First, this study provided initial evidence about the relative significance of the situational and personal factors for workplace proactivity. In doing so, we highlighted the importance of the growth mindset (operationalized as a learning goal orientation in this study) as a highly instrumental individual disposition in engaging frontline hospitality employees in proactive service behaviors. It suggests that, along with contextual motivators such as interpersonal leadership, individual disposition to strive for learning is a significant factor of motivation for proactive behavior in frontline hospitality employees. Second, by finding employee engagement as a significant mechanism capable of transmitting the impact of contextual and/or individual factors on the proactive employee behavior in hospitality sector, results of this study join hands with the rare previous research using engagement as a link to proactivity in hospitality literature (e.g., Li et al., 2016). It implies the importance of the integration of well-established motivation theory (SDT) and a well-accepted motivation construct (employee engagement) in finding ways to make frontline hospitality employees behave proactively.
The findings of the study are equally significant for practitioners of the hospitality industry as well. First, if hospitality managers want to secure a unique competitive advantage through the proactive service behavior of their frontline employees, they should focus on cultivating and enhancing growth mindset by promoting learning goal orientation as an acceptable norm in the organization. Second, though having a small impact, in this case, interpersonal leadership is also a significant booster of motivation and proactive service behavior. So, hospitality managers should develop leaders with support, care, empathy, and fairness characteristics. Finally, as engagement is a highly significant predictor of proactivity, hospitality managers should focus on engagement training as well, so that frontline hospitality employees feel innately motivated with their whole-self available for their work-roles.

LIMITATIONS AND RECOMMENDATIONS FOR FUTURE STUDIES
Along with strengths, it is essential to mention the limitations of the study as well, so that future researchers try to overcome the same. First, as this study used a self-response cross-sectional research design, there remains a probability of common method bias. Though we provided statistical evidence to rule-out the prevalence of CMV in the current study by using the Marker variable approach, researchers in the future can apply different research designs such as longitudinal and diary studies in replicating findings of this study. Second, as the research model of this study is statistically meaningful in predicting and explaining the total variance of proactive service performance, future hospitality researchers can use different combinations of factors from context and individual differences in developing models for proactive service motivation. Finally, while modeling an active form of employee behaviors, we recommend hospitality researchers to use self-determination theory (SDT) and employee engagement as two significant sources of workplace motivation to strive for excellence.