A Multi-Level Research on the Antecedents and Consequences of Group Task Satisfaction

In this study, in order to examine antecedents and consequences of group task satisfaction, We collected 204 questionnaires with 35 hotel groups from 4 hotels located in Guangzhou and Shenzhen and used HLM to test the relationship between group task satisfaction and other variables in group level and individual level. The research findings show that employee individual job satisfaction mediates the effect of group task satisfaction on employee turnover intention and job performance; group task satisfaction mediates the positive effect of group cohesion and group job characteristics on individual job performance; group cohesion moderates the effect of individual job satisfaction on turnover intention and job performance.


Introduction
In recent years, due to low switching cost and low salary of hotel employees, hotel employee turnover rate is high.The high employee turnover rate will cause certain pressure toward the human resources department, training department, finance department and management of hotels.Therefore, many hotel managers spent a lot of human, financial and material resources in seeking the reasons for high employee turnover rate and the corresponding solutions.Meanwhile, a heated discussion was also raised in the academic circle [1]- [3].Many scholars explored the reasons from job satisfaction in individual level [4]- [7], but in our depth interview, a number of staff responded that no sense of belonging in the work and intriguing against each other among col-leagues were one of the important reasons that leaded to their turnover intention.Based on what mentioned above, this article mainly discusses how to improve the hotel employee job satisfaction in the group level in order to improve the employee job performance and reduce employee turnover rate.

Group Task Satisfaction
At present, the academic research on employee job satisfaction was more and more mature.However, the academic research on employee job satisfaction in group level is very rare.At the beginning of this century, Mason and Griffin [8] first put forward the concept of "Group Task Satisfaction" (GTS).They pointed out that GTS was a group-level concept, i.e. group members' consistent attitude to work task and work environment instead of the aggregation or the average of employee individual job satisfaction.Mason and Griffin [8] argued that measuring GTS should ask group members to assess their perceptions of job satisfaction in group level, namely using transformation reference object consensus method to measure.Empirical study about antecedents of group task satisfaction was little.So, the present study mainly explores the influence of two group-level variables (i.e. group cohesion and group job characteristics) on group task satisfaction.

Group Cohesion and Group Task Satisfaction
Concept of cohesion was first put forward by Lewin [9] who argued that cohesion was the "glue" of keeping group members together and maintaining the relationship between them.According to the team-member exchange theory, each member in the group will produce different emotional reactions when contacting with other members in the group, and one may spread to other members and produce similar sentiment.In a high-quality team-member exchange relationship group, positive emotion diffuses among group members, harmonious group climate forms, emotional cohesion increases among members, and the group members are more likely to gain the satisfaction from group relationship.On this basis, the author puts forward the hypothesis: H1: Group cohesion has a significant positive influence on group task satisfaction.

Group Job Characteristics and Group Task Satisfaction
Group job characteristics, as the name implies, are inherent attributes of a group work or task.At present, a typical job characteristics theory proposed by Hackman and Oldham [10] showed that dimensions of job characteristics (i.e., task importance, autonomy, diversity, completeness, feedback) would affect employees' outcomes through "the key psychological state".Therefore, group job satisfaction, as a key psychological state in a group level, might be influenced by group job characteristics.Mason and Griffin [8] had also pointed out that group task satisfaction and job characteristics were correlated.For example, a challenging and significance tasks are more attractive than a regular task without any feedback.Job involvement, tasks diversity, task significance and other job characteristics would affect group members' job satisfaction.On this basis, the author concludes that: H2: Group job characteristics have a significant positive influence on group job satisfaction.

Group Task Satisfaction and Job Satisfaction
Mason and Griffin [8] argues that group job satisfaction has a direct impact on employee's job satisfaction.The consistent work attitude among group members would prompt them to obey conduct code in the group and keep consistent with work attitude in the group, thus in a high overall satisfaction group, the members are more likely to experience the satisfaction with work.On this basis, the author puts forward hypothesis: H3: group job satisfaction has significant positive influence on job satisfaction.From the above, the author argues that group job characteristics and group cohesion do not directly affect employee job satisfaction, but indirectly affect it through group task satisfaction.On this basis, the author puts forward hypotheses: H4a: Group task satisfaction mediates the positive influence group job characteristics on job satisfaction; H4b: Group task satisfaction mediates the positive influence group cohesion on job satisfaction.

The Mediating Effect of Individual Job Satisfaction
The more satisfied employees are more willing to stay in the enterprise instead of quitting [11], i.e. employee job satisfaction negatively correlates with turnover intention.In addition, employees with high job satisfaction are more likely to love the job, to be highly engaged in job and to improve job performance.Therefore satisfied employees would improve their job performance.Combining with hypothesis 3, the author proposes hypotheses: H5a: Employee job satisfaction mediates the negative influence of group task satisfaction on turnover intention; H5b: Employee job satisfaction mediates the positive influence of group task satisfaction on job performance.

The Moderating Effect of Group Cohesion
Little research regarded group cohesion as a moderator variable, and cross-level moderating effect was less.Group cohesion could be divided into two dimensions.One is emotional cohesion, also called social cohesion, which refers to close relationship among group members, and interdependence tendency formed in daily life in which they get along with each other.Another is task cohesion which puts emphasis on coordination and distribution of responsibilities and complementary skills of group members when they are completing group task together [12].In a group of strong emotional cohesion, satisfied employees would emotionally attach to their group, reducing their turnover intention; in a group of strong task cohesion, group members would closely cooperate, improving job performance of their group.Therefore, the author concludes that: H6a: Group cohesion moderates the positive effect of employee job satisfaction on turnover intention; H6b: Group cohesion moderates the negative effect of employee job satisfaction on job performance.From what has been discussed above, the author puts forward hypothesis model shown in Figure 1.

Sample
The author issued formal questionnaires to the employees in the front-line service group and back office group of three five-star hotels and a four-star hotel in Guangzhou and Shenzhen.Before issuing, the author packed corresponding questionnaires into the envelope according to the number of each group and marked the group name on it to avoid packing by mistake.After employees completed, the author sealed the envelope and took them back.A total of 252 questionnaires from 40 groups were collected.Eliminating invalid questionnaires such as incomplete questionnaires, 204 questionnaires from 35 groups were valid, with an 81% effective rate.In the valid study samples, women accounted for 56.1%; employee between 18 and 28 years old accounting for 74.6%;between 29 and 45 years old accounting for 23.9%; 46 years old or above accounting for 1.5%; below high school degree, 43.4%, college degree or above, 56.6%; monthly income below 3000 RMB, 68.2%, between 3000 and 8000, 30.3%, above 8000, 1.5%; respondents who work in current hotel are less than 1 year, 55.6%, between 1 year and 5 years, 39.5%, above 5 years, 4.9%.

Measures
The measurements of key variables in present study adopt Likert 5-points scale, 1 for complete disagreement and 5 for complete agreement.Group task satisfaction (GTS) was measured with group task satisfaction scale developed by Mason and Griffin [13].The measurement of group cohesion (GC) referred to group identity scale by Henry et al. [14], Henry et al. argued "group cohesion from the group identity".Group job characteristics (GJC) was measured with job diversity, autonomy and feedback three dimension, and items design referred to the study of Teas [15].Job satisfaction (JS) was measured with job satisfaction 5-scale developed by Smith et al. [16].The measurement of turnover intention (TI) referred to the scale developed by Wayne et al. [17].The measurement of job performance (JP) referred to "task-contextual performance", two-dimension model proposed by Borman and Motowidlo [18].

Reliability and Validity Test
Firstly, this research has adopted the SPSS17.0 to conduct exploratory factor analysis (EFA) of the three multidimensional variables, which are group task satisfaction, group job characteristic and group cohesion.The subscales of group task satisfaction and group job characteristics hold very high loads among their common factors; meanwhile, they have achieved convergent validity and discrimination validity.Among the sub-scales of group cohesion, the item coefficient of GC11 was less than 0.45, it had to be deleted; while GC3, GC5 and GC7 fell over the dimension of planned construct and they couldn't be explained easily, therefore, they also had to be deleted.
After revising the scale, this research will adopt LISREL8.80 to conducted the confirmatory factor analysis (CFA ) of group task satisfaction, group cohesion and group job characteristics, which furthermore will use indexes of ×2 testing, RMSEA, NNFI, CFI, SRMR to judge the fitting degree of models.As shown in Table 1, the measurement models of group task satisfaction, group cohesion and group job characteristics basically fit the data, demonstrating that the scales have good construct validity.After reversion, all variables "Cronbach" α reached between 0.609 and 0.832, which showed that the reliability of the scales reached acceptable scope.

The Multilevel Linear Model Analysis
Among the hierarchical linear models, level 1 (individual level) includes the employee's job satisfaction, turnover intention, job performance and control variables.Level 2 (group level)includes group job characteristics, group cohesion and group task satisfaction.RWG value, coefficient of ICC (1) and ICC (2) of employees' score of group job characteristics, group cohesion and group task satisfaction basically reached acceptable aggregation conditions recommended by Klein et al. [19].Therefore, we aggregated the individual scores of the three variables to level 2 variable values.

Analysis of Multilevel Main Effect and Moderating Effect
Using HLM 6.08 software, the author carried out data analysis according to Hofman's multilevel linear model analysis method [20], and the analytic results demonstrates that (see Table 2): 1) After controlling the control variables, the independent variables of level 1 and independent variables of level 2 (i.e., group cohesion and group job characteristic), group task satisfaction had significant positive influence on employee job satisfaction (γ 03 = 0.932, p < 0.01), supporting H3; meanwhile group task satisfaction had no significant effect on employee turnover intention (γ 03 = −0.283,p > 0.1) and job performance (γ 03 = −0.411,p > 0.1).

The Multilevel Mediating Effect Analysis
In this study, with the mediation effect analysis method of Barron and Kenny [22], the author tested the mediation effect.The analysis results of multilevel mediation effect (see Table 3) indicate that 1) group characteristics (β = 0.330, p < 0.05) and group cohesion (β = 0.394, p < 0.01) has significantly positive influence on group task satisfaction, supporting H1 and H2; 2) group task satisfaction fully mediates positive impact of group job characteristics (β = 0.929, p < 0.01)and group cohesion (β = 0.932, p < 0.01)on employee job satisfaction, supporting H4a and H4b; 3) employee job satisfaction fully mediates positive influence of group task satisfaction on employee turnover intention (β = −0.654,p < 0.01) and job performance (β = 0.445, p < 0.01), supporting H5a and H5b.

Discussion and Conclusions
This research has improved and deepened Mason and Griffin's group task satisfaction theory [8].And we have verified that: 1) Group cohesion and group job characteristics have a significant positive impact on group task satisfaction, which are similar to empirical research findings of Seers [23], Dobbins et al. [24] and Kidwell et al. [25]; 2) Employees' perception of group task satisfaction has a significant positive impact on the job satisfaction in individual level, which is consistent with Mason and Griffin's finding; 3) Group job characteristics and group cohesion do not directly affect employees' individual job satisfaction, but indirectly affect it through group task satisfaction; group task satisfaction does not directly affect turnover intention and job performance, but indirectly affects them through individual job satisfaction, which further improves and extends Mason and Griffin's group task satisfaction theory; 4) Group cohesion across-level moderates the effect of individual job satisfaction on turnover intention and job performance.There is little research about the moderating effect of group cohesion, not mention the research about cross-level moderating effect, therefore, this finding provides a new perspective on the effect of group cohesion.These research conclusions provide some revelations for hotel managers on how to reduce employee turnover intention and improve job performance from a group point of view.It is necessary for the hotel managers to create a high cohesive group to reduce the employee turnover rate and improve the employee job performance.In addition, hotel managers should also pay attention to the improvement of work diversity, autonomy and feedback.Especially in terms of autonomy, due to the limited mandate, the hotel junior employees only report to their superiors based on the principles of basic services when facing unexpected situations, which may result in service delays and other issues.At last, it may affect the customer experience and employee motivation initiative.Based on this, we think the group managers should give a reasonable authorization based on the group member personality, experience, skills, etc., in order to save time, avoid unnecessary delays in service and complaints, mobilize the member's initiation, and increase their sense of accomplishment.There are still some limitations in this research.First of all, we only investigated the hotel industry, so the general applicability of the findings needs testing.In addition, in this research, we failed to make paired questionnaires to collect data for the reason of operational difficulties, so there may be homologous error problems.

Figure 2 .
Figure 2. (a) Interaction effect of GC and JS on TI; (b) Interaction effect of GC and JS on JP.

Table 2 .
Analysis results of multilevel main effect and moderating effect.

Table 3 .
Analysis results of multilevel mediation effect.IV represents Independent variable; DV represents dependent variable; MV represents mediation variable.a the first value is the regression coefficient, and the value in parentheses is the standard error; * p < 0.05; ** p < 0.01.