Conflict and Employee Engagement

The majority of respondents have a bachelor's degree (73.0 percent or 81 people). 27.0 percent or 30 people have a diploma. In addition, the majority of respondents were at the age of 30-39 years, i.e. 52 people or 46.8 percent of respondents. 20 people or 18.0 percent of respondents were at the age of 40-49 years. The remaining, 39 people or 35.1 respondents, were at the age of 21-29 years. 

More than half of the respondents, female employees at PT Freeport Indonesia, have been working for more than 7 years. The majority have been working at this company for 7–9 years(54 people or 48.6 percent), and 8.1 percent or 9 people have been working for more than 10 years. 38 people or 34.2 percent have been working for 4-6 years, and 10 people or 9.0 percent have been working for 0-3 years. 

Table 1 indicates that the scale, magnitude, and statistical appropriateness are acceptable. The average variance extracted (AVE) value for all latent variables in the work-family conflict model, job engagement, emotional exhaustion, and turnover intention was 0.690; 0.719; 0.544; and 0.860 (above 0.50) respectively, while the hAVE value of all constructs is 0.737 to 0.927, which is greater than the correlation between latent variables (-0.087 to 0.608). Thus, it can be said that all constructs meet the validity requirements based on discriminant validity criteria.

Table 1 AVE Score, Composite Reliability (CR), Mean, Standard Deviation and Correlation Coefficient

		AVE	√AVE	CR	Mean	SD	1	2	3	4
1	Work-family conflict	0.690	0.830	0.870	3.23	2.197	(0.779)
2	Job engagement	0.719	0.848	0.884	3.55	2.375	-0.087	(0.804)
3	Emotional exhaustion	0.544	0.737	0.824	2.95	2.410	0.449	0.149	(0.725)
4	Turnover intention	0.860	0.927	0.948	2.52	2.980	0.520	0.074	0.608	(0.919)

Note: *p, 0.05; **p, 0.01; (two-tailed significance); Cronbach's alphas for each scale is italicizedand addressed diagonally.

On average, Cronbach's alpha for reliability criteria is quite high; turnover intention had the highest value. Sequentially, Cronbach's alpha coefficient values for the four scales reported in this study ranged from 0.725 to 0.919; they are still acceptable. The value of composite reliability (CR) is 0.870; 0.884; 0.824; and 0.948 (above 0.80) in sequence. It can be concluded that the entire construct is reliable, both according to the criteria of composite reliability and Cronbach's alpha.

The mean, standard deviation and range of items in the measurements used in this study are as follows: (a) work-family conflict (9 items; M = 3.23, SD= 2.197); (b) job engagement (9 items; M= 3.55, SD = 2,375); (c) emotional exhaustion (8 items; M = 2.95, SD = 2,410); and (d) turnover intention (3 items; M = 2.52, SD = 2,980). 

Table 2 presents the R-squared values and communality among the variables of the research model observed. Goodness of Fit (GoF) in this study was assessed by the following equation: √AR2 * A.Com = √0.2338 x 0.7035 = 0.4056. 0.4056 indicates that the model has a large enough predictive value. 58.22 percent of the Q-Squared Predictive Relevance value for turnover intention is explained by work family conflict, job engagement, and emotional exhaustion.

Table 2 Goodness of Fit Evaluation

Construct	R-Square	Communality
Work-family conflict		0.6902
Job engagement	0.0523	0.7192
Emotional exhaustion	0.2015	0.5444
Turnover intention	0.4478	0.8605
Average	0.2338	0.7035

Furthermore, based on the results of the hypothesis test presented in Table 2, the analysis model is illustrated in Figure 1

The following is the illustration path analysis model:

Figure 1 Path Analysis Model

Hypothesis 1 is accepted. The effect of work family conflict on turnover intention was found to be significant, with β = 0.3160 and t-statistic = 3.7771 (above the critical value of 1.96). Based on these findings, then hypothesis 1 is accepted, which means that there is a significant and positive effect of work-family conflict on employee turnover intention. Therefore, high work-family conflict can increase the turnover intention of PT Freeport Indonesia's female employees. Work family conflicts create problems for workers by positively influencing female employees to leave the company. Lack of flexibility at work is believed to be one that most people think as pressure that triggers workfamily conflict. The findings of this study are consistent with the results of research which were conducted by Haar, Yavas, et al., Noor and Maad, and Ozbag and Ceyhun. Conversely, the results of this study cannot confirm the findings of research which was conducted by Hangyue, et al. as well as Panatik, et al. that high work-family conflict can reduce a person's tendency to move or leave the company.

Table 3 Path Analysis Results

Relationship Between Variables	t-statistic (cut-off = 1.96)	Coefficient of Direct Effect	Coefficient of Direct Effect (through job engagement)	Coefficient of Direct Effect (through Emotional exhaustion)
1. Work-family conflict ☐ Job engagement	1.3548	- 0.1935
2. Work-family conflict ☐  Emotional exhaustion	7.1812	0.4488**
3. Work-family conflict ☐ Turnover intention	3.7771	0.3160*	-0.0064	0.207**
4. Job engagement ☐ Turnover intention	0.4480	0.0326
5. Emotional exhaustion  ☐ Turnover intention	5.9531	0.4617**	0.020
6. Emotional exhaustion ☐ Job engagement	1.9862	0.2363*

Note: ** p, 0, 01; (2-tailed); * p, 0, 05; (2-tailed);

Hypothesis 2 is rejected. The effect of workfamily conflict on job engagement was found to be insignificant, with the beta coefficient amounted to -0.194 and a t-statistic value of 1.355 (below a critical value of 1.96); therefore, hypothesis 2 was rejected. It means that high work-family conflict cannot reduce or increase job engagement of PT Freeport Indonesia's female employees. The increasing work demands felt by employees that causing employees to get a lot of complaints from the family is considered a potential source of conflict. However, the conflict between work and family role does not necessarily directly reduce or increase the attachment of PT Freeport Indonesia's female employees to their job. One of the factors that caused the lack of direct influence could be because the company already has special programs that can intervene in the impact of dual role conflict experienced by an employee. The findings of this study cannot confirm the results of the studies which was conducted by Baka and Derbis; Kesumaningsari and Simarmata; and Karatepe and Karadas.

Hypothesis 3 is rejected. The effect of job engagement on turnover intention was found to be insignificant, with the beta coefficient amounted to 0.033, and a t-statistic value of 0.448 (smaller than the critical value of 1.96). Based on these findings, hypothesis 3 is rejected, which means that higher job engagement cannot reduce or increase the turnover intention of female employees of PT Freeport Indonesia. The attachment to work is seen as a mechanism for individual attitudes towards his work, not towards the organization. Job satisfaction predicts more work-related behavior, for example, employee performance or extra role performance. The findings of this research are in line with the results of research which was conducted by Mxenge, Dywili, Bazana. On the contrary, the findings of this study cannot confirm the studies which were conducted by Schaufeli and Bakker, Mangi and Jalbani, Agoi, Biswakarma, Rachman and Dewanto who found a significant negative effect of job engagement on turnover intention.

Hypothesis 4 is accepted. The effect of workfamily conflict on emotional exhaustion was found to be significant, with β = 0.449 and t-statistic amounted to 7.181 (greater than the critical value of 1.96). It can be said that hypothesis 4 is accepted, which means that higher work-family conflict can increase emotional exhaustion in female employees of PT Freeport Indonesia. The results of this study confirm the results of studies conducted by Hangyue, et al., Zhang, et al., Wang, et al.; Mete, et al., and Akgunduz, et al. Some of these studies confirm a positive linear relationship between work-family conflict with emotional exhaustion; higher the level of conflict between work and family that is experienced by an individual will increase his emotional fatigue at work. 

Hypothesis 5 is accepted. With a beta coefficient of 0.462 and a t-statistic value of 5.953 (greater than the critical value of 1.96), it can be said that the higher emotional exhaustion can increase the turnover intention of female employees of PT Freeport Indonesia. Emotional fatigue of female employees plays a key role in the turnover process; the lack of satisfaction or low commitment plays a role in cognitive processes, which eventually pushes employees to quit their jobs. Although it is dealt with neutrally, the emotional fatigue of female employees tends to be caused more by things in emotionally draining work. One's emotion drained by work is the potential for thought to withdraw from work and even from the organization (turnover intention). The findings of this study confirm the results of studies which was conducted by Wright and Cropanzano, Schaufeli and Bakker, and Hangyue, et al. However, the findings of this study cannot confirm the results of the study which was conducted by Geurts, et al. that there is no influence of emotional exhaustion on turnover intention. 

Hypothesis 6 is rejected. The analysis results showed a beta coefficient of 0.236 and a t-statistic of 1.986 (greater than the critical value of 1.96). Although the effect was found to be significant, the influence is positive, so this indicates that the higher emotional exhaustion can actually increase the job engagement of female employees of PT Freeport Indonesia. The higher the fatigue level, the greater the intensity and quality of interaction in the company. In short, the employees get tired because they try hard to perform their job well so that it will lead to a sense of attachment to the job. An engaged female employee is willing to make extra effort to achieve organizational goals. The findings of this study confirm the results of studies conducted by Bezuidenhout and Cilliers and Mangi and Jalbani. However, these findings cannot confirm the results of studies conducted by Bakker, et al., Sonn, and Praditia that emotional exhaustion hasa significant effect on job engagement of individuals. 

Hypothesis 7 is rejected. Based on the results of the mediation test (using Sobel test), the value of the direct influence of work-family conflict on turnover intention was found to be greater than the indirect effect of work-family conflict on turnover intention through job engagement. In addition, this mediation was not significant, so hypothesis 7 was rejected. Female employee engagement cannot encourage or reduce the influence of work-family conflict faced by the female employees of PT Freeport Indonesia on the turnover intention. The findings of this study cannot confirm the results of studies conducted by Schaufeli and Bakker and Mangi and Jalbani that job engagement can have a mediating effect on the antecedent effect of engagement on turnover intention. 

Hypothesis 8 is accepted. Based on the results of the mediation test (using the Sobel test), emotional exhaustion can play an important role in mediating the effect of work-family conflict on turnover intention; therefore, hypothesis 8 is accepted. Based on the results of mediation tests that have been conducted, emotional exhaustion is categorized as partial mediation, given the value of the indirect effect of work-family conflict on turnover intention through emotional exhaustion found to be greater than the direct effect of work-family conflict on turnover intention. It can be concluded that the emotional exhaustion felt by the female employees of PT Freeport Indonesia will further encourage the influence of work-family conflicts on the turnover intention. The findings of this study confirm the results of studies which was conducted by Hangyue, et al. and Yavas, et al., and Knight and Leimer who revealed the important role of emotional exhaustion in strengthening the impact of work-family conflict on turnover intention. 

Hypothesis 9 is rejected because the coefficient value of the indirect effect of emotional exhaustion on turnover intention through job engagement was found to be smaller than the direct effect of emotional exhaustion on turnover intention. In addition, the results of the mediation test (using Sobet test) also show that job engagement did not mediate the effect of emotional exhaustion on turnover intention. Thus, employee attachment to their job cannot predict the increase in emotional exhaustion on turnover intention. According to Maslach, et al., job characteristics, especially feedback and autonomy, are consistently related to burnout which is a positive antithesis of employee engagement. These results partially support the statement that emotional exhaustion has a negative effect on job engagement. The findings of this study do not confirm the results of studies conducted by by Schaufeli and Bakker and Mangi and Jalbani which explain the important role of job engagement mediation on the effect of emotional exhaustion on turnover intention.