The Impact of Social Media on Employee Work Performance with Trust as a Mediation Variable

Can the use of social media which is currently a trend in the community, have an influence on improving employee performance? This is what makes the research carried out. The analysis technique used is the Structural Equation Model (SEM) method with AMOS version 23. The sample consisted of 102 people who were employees in various types of private companies in DKI Jakarta. The results showed that the more social media is used to communicate with colleagues in completing work, the more trust among employee increases. Also the more social media is used to communicate with colleagues in completing work, the more improves employee work performance. But trust can not influence performance of employee, and trust is not a variable that mediates between social media and employee performance.


INTRODUCTION
The rapid development of technology and information has a big influence on the development of the current global situation. In fact, information technology always gives feedback, both positively and negatively to every aspect of life. The development of technology and information that is very popular in today's society is Social Media. Nowadays, social media has become a basic necessity for everyone. Social media networks that are often used include Facebook, Twitter, Telegram, Instagram, WhatsApp and others.
Outcome of this study are expected to add insight into how social media affects employee performance, and can also provide input to organizations that Social Media can affect employee performance.

THEORITICAL REVIEW
Social Media. (Parveen et. al., 2016) states, social media has changed the way business is done. The statement from (Pride et al., 2017) is, social media represents the online interactions that allow people and businesses to communicate and share ideas, personal information, and information about products and services. According to (DiMicco et al., 2008), Employees use the internet with the main motivation is to build stronger ties (Network Ties) between employees and also to reach employees they don't know well. (Jackson et al., 2007) also fond out how useful blogs at work is to create, maintain and strengthen relationships. At companies like IBM and Microsoft, young and new employees will be familiar with Facebook. This social media is used extensively to learn about their new colleagues and also create new social relationships (DiMicco et al., 2008). Social media is an effective platform that consists of a collection of information and tools of communication. Social media provides a variety of channels in communication, both in environment of work and the social (Cao et al., 2016). Employees use social media to search and share information, also to meet friends. They use social media to obtain and develop relationship among people (Cao et. al., 2016). (Smith, 2009) states that the independence of space on social media creates open communication that gives rise to a new form of collaboration, where users can share at any time without time and location restrictions. Users interact with various levels of intensity on an unprecedented scale. Using social media also influence of organizational performance, with the most dominant social media usage elements being on 'social media for information accessibility' (Nilasari et. al., 2019).
Trust. An individual's ability to authenticate knowledge that is accepted as relevant and factual, without risk, without error and inconsistency, is known as a trust (Lin, 2007). (Skinner et al., 2014) stated, trust is a "process" from individual beliefs which focus and lead to decisions to believe and show trust. In general, trust grows over time with interactions, relationships and experiences (Fang and Chiu, 2010). Deep understanding can minimize the behavior of uncertainty and intentions of others, as a prerequisite for trust (Valenzuela et al., 2009). Recognize other people more, will believe (trust) or the more will not believe (distrust). Individual trust (trust) can be obtained through communication in various kinds of media, one of them is social media (Cao et al., 2016). Many studies of trust between employees and management have been conducted before (Kampkötter et al., 2016). Studies (Weibel et al., 2016) has also examined about organizations which is try to organize behavior of employee as a way to improve employee confidence in organizations. The important thing in terms of knowledge transfer is trust between employees. The focus of interpersonal trust is on individuals who are willing to be vulnerable and believe in the positive behavior of their peers (De Jong et al., 2016). (Fang and Chiu, 2010) stated, interaction of face to face, mutual experience, and length of intimacy will lead to ready and willing in transfer knowledge. (Cao et al., 2016) argues, employee social capital can be driven by social media, which also facilitates the transfer of knowledge. Improving employee performance can be helped by social capital and knowledge transfer. (Fang and Chiu, 2010) discusses many studies and theories which state that trust has a positive effect on the intention to share individual knowledge and behavior in organizations.
Employee Work Performance. In the view of (Halabi and Lussier, 2005), performance is an expansion in which expectations or objectives have met. As according to (Hariandja, 2004), performance is defined as the results of work produced by employees or actual behavior that is displayed in accordance with its role in the organization. Whereas (Tika, 2010) states that company performance is the result of the work function or activities of a person or group of people in an organization which is influenced by various factors for the achievement of organizational goals at a certain period of time. If indicators used as performance measures increase, the strategy has been implemented well (Anthony and Govindarajan, 2001). Based on description above it can be stated, that good performance of employee is a step to achieve organizational goals.

Figure 1. Conceptual Framework
Hyphotesis. The use of Enterprise Social Media (ESM) can increase interactivity and create social networks that can create trust among colleagues (Cao et al., 2016). The study of virtual teams and the distribution of work found that trust is positively related to knowledge transfer. This shows that understanding through the use of ESM and reciprocal communication will increase interpersonal trust in virtual team members (Chen and Hung, 2010). Hopefully the successful adoption of ESM by employees will enable organizations to increase trust among them. Based on previous research, the first hypothesis can be arranged as follows:

H1: Social Media influences Trust among employees.
Knowledge is the main competitive advantage for every individual and even organization. (Stasser and Titus, 1985) said that the transfer of knowledge that is often done will be able to provide a way to solve problems, which helps in making superior decisions. (Andres and Zmud, 2002) also stated that increased knowledge exchange would be very important to overcome interdependence of task and uncertainty. (Cao et al., 2016) argue that trust based on the transfer of knowledge on social media has increased the ability of employees to solve problems effectively, and eventually will improve work performance. Based on previous research, the second hypothesis can be arranged as follows: H2: There is an influence of Trust on Employee Work Performance.
Social media influences employee performance. Previous research from (North, 2010) supports this statement that the use of social media in the workplace is beneficial, enabling the realization of new relations also maintaining existing one. Research (Leidner et al., 2010) showed that by using ESM, new employee performance will be faster and better, by increasing new relationships and transferring knowledge tacit and explicitly. The use of ESM will increase collective knowledge so that it contributes to employee job performance (Bennett et al., 2014). Based on previous research, the third hypothesis can be arranged as follows: H3: Social Media influences Employee Work Performance. (Fang and Chiu's, 2010) research shows that the use of Enterprise Social Media (ESM) can create and strengthen social networks and interactions that can create trust among colleagues. (Cao et al., 2016) argue that trust based on knowledge transfer on social media has increased the ability of employees to solve problems effectively, and eventually will improve work performance. Based on previous research, the fourth hypothesis can be arranged as follows:

H4: Trust as a mediation between the influence of Social Media on Employee Work
Performance.

METHODOLOGY
Design of this research was a hypothesis test (hypotesis testing). Hypothesis testing is research that aims to test hypotheses and generally is research that explains phenomena in the form of relationships between variables. The time horizon of this study is cross-sectional meaning that information is collected only once even though it may be done for a certain period so that it is often known as one-shot (Sekaran and Bougie, 2010). One shot of this research at a certain time in January -March 2020.
Variables and Measurements. The research variable is an exogenous variable (independent variable) that is Social Media (Social Media Use at Work). Then the endogenous variable (dependent variable), namely Trust and Employee Work Performance, and Trust are also mediating variables. (1) I often use social media to obtain work-related information and knowledge (2) I regularly use social media to maintain and strengthen communication with colleges in my work (3) What is your frequency of usage of social media in the work?
Based on (Kankanha lli et al., 2005) Trust (1) I assumed that members in the virtual community created by social media would always look out for my interests (2) I assumed that members in the virtual community created by social media would go out of their ways to make sure I was not damaged or harmed (3) I felt like members in the virtual community created by social media cared what happened to me (4) I believed that members in the virtual community created by social media approached their jobs with professionalism and dedication (5) Given members in the virtual community created by social media track record, I saw no reason to doubt their competence and preparation Based on (Levin and Cross, 2004) Employee Work Performance (1) I almost always perform better than an acceptable level (2) I often perform better than can be expected from me (3) I often put in extra effort in my work (4) I intentionally expend a great deal of effort in carrying out my job (5) I try to work as hard as possible (6) The quality of my work is top-notch Based on (Kuvaas, 2006) Samples. The population of this research is employees who work in various types of private companies in DKI Jakarta. The number of samples is 102 people according to (Hair et al., 2014) that the recommended number of samples is in the range of 100 -400. (Hair et al., 2014) also states that the minimum sample size is the number of indicators multiplied by 5. The number of indicators of this study is 14, multiplied by 5 = 70 samples, therefore a total sample of 102 means that they meet the requirements. Data Collection. Data were collected using a questionnaire technique, using an online questionnaire that can be accessed at: https://docs.google.com/forms/d/e/1FAIpQLSf1J-WJgGyuMduTSV8qvZNhuk_oYOrjfhiZBnnkWGLbGZl3LA/viewform?vc=0&c=0&w=1 This questionnaire is closed where the answer is already available. Questionnaire as a data collection tool using 5 Likert scale. Scale 1 = Strongly Disagree until scale 5 = Strongly Agree. Test Reliability and Validity. Before a questionnaire which is an instrument in research is widely used first, it is tested to measure the validity and reliability of the measuring instrument. Validity and reliability tests in this study, using SPSS version 23 computer aids. Validity tests are used to measure the validity of a questionnaire. A questionnaire is said to be valid if the questions on the questionnaire are able to reveal something that will be measured by the questionnaire. Test Reliability with the "Internal Consistency Reliability Method", using the Cronbach's Alpha coefficient. Cronbach's coefficient alpha as a coefficient of reliability is used to test each construct. According to (Sekaran and Bougie, 2010), the closer the reliability coefficient is to 1 (one), the better it is considered. In general, reliability less than 0.6 is considered weak, around 0.7 is considered acceptable, while above 0.8 is considered good. Based on the statement above, the basis for making a reliability test can be concluded as follows: (i) If the cronbach's coefficient alpha ≥ 0.6, the construct is considered reliable.
(ii) If the cronbach's coefficient alpha <0.6, the construct is considered unreliable. Test of Goodness of fit. Goodness of fit evaluation is done to see to what extent the data and models meet SEM assumptions. Evaluation is carried out on the overall model then proceed with evaluation of the measurement and structural models separately (Hair et al., 2014). (Hair et al., 2014) states that from several Absolute Fit Measure and Incremental Fit Measure tests, if the results of one test are 'fit' then it can be concluded that the model used is fit.

Data Analysis and the Structural Model.
Based on the results of the suitability of the model above, then the next step in the form of hypothesis testing using Structural Equation Modeling (SEM) can be done. Structural Equation Model (SEM) using the AMOS version 23 program. SEM is basically a collection of statistical techniques that allow the testing of a series of relatively complex relationships simultaneously (Ferdinand, 2002). Structural Equation Modeling This research can be seen in the following figure:

THE RESULTS OF STATISTICAL TESTS AND DISCUSSION
Results. Hypothesis testing is done by testing the C.R (t-value) whether greater than 1.96 or p-value less than 0.05. To compare the p-value with the level of significance, based on the following conditions (Hair et al., 2013): 1. If the p-value ≤ 0.05 then Ho is rejected and the research hypothesis is supported. 2. If the p-value> 0.05 then Ho fails to be rejected and the research hypothesis is not supported. SEM test results can be seen in the following table: Discussion. In testing hypothesis 1, the results of data processing show that Social Media influences Trust, having a coefficient value of 0.582 which means that the higher the perception of Social Media, the higher the perception of Trust. Test results on the value of C.R (t-value) of 3.483 is greater than 1.96 or p-value of 0.000 <0.05 (alpha 5%), it can be concluded statistically at a 95% confidence level there is a positive influence of Social Media on Trusts. Hypothesis 1 test results show that Social Media significantly influences Trust. Social media at work such as using social media to get work-related information, and using social media to maintain and strengthen communication with colleagues at work.
The results of this study support the results of previous studies conducted by (Cao et al., 2016) that the use of Enterprise Social Media (ESM) can create and strengthen social networks and interactions that can create trust among colleagues. The results of the study are also supported by studies of virtual teams and distributed work that find that trust is positively related to knowledge transfer (Chen and Hung, 2010). Thus, it can be concluded that if Social Media in the workplace is getting better, the Trust will be more positive.
The most dominant element of social media is the element 'I regularly use social media to maintain and strengthen communication with colleagues in my work' with a value of 1.59. This shows that the more social media used to communicate with colleagues in completing work, it will lead to increasing trust among employees.
In testing hypothesis 2, the results of data processing show that Trust does not affect Employee Work Performance, has a coefficient of -0.014. The test results on the CR value (t-value) of -0.150 smaller than 1.96 or p-value of 0.880> 0.05 (alpha 5%), it can be concluded statistically at a 95% confidence level there is no positive influence of Trust on Employee Work Performance. Hypothesis 2 test results, showed that Trust does not significantly influence Employee Work Performance. The results of this study do not support the results of previous studies conducted by (Cao et al., 2016) that trust based on knowledge transfer on social media has increased the ability of employees to solve problems in an effective way, which in turn will improve employee work performance.
In testing hypothesis 3, the results of data processing show that Social Media influences Employee Work Performance, has a coefficient value of 0.432 which means that the higher the perception of Social Media, the higher the perception of Employee Work Performance. Test results on the CR value (t-value) of 2.987 is greater than 1.96 or p-value of 0.003 <0.05 (alpha 5%), it can be concluded statistically at a 95% confidence level there is a positive influence of Social Media on Employee Work Performance. Hypothesis 3 test results, showed that Social Media significantly influence Employee Work Performance. Social media at work such as using social media to get work-related information, and using social media to maintain and strengthen communication with colleagues at work.
The results of this study support the results of previous studies conducted by (Leidner et al., 2010) that new employees perform better and faster after using ESM when building new relationships and facilitating quick tacit and transfer of explicit knowledge. Supported also by (Bennett et al., 2010) who found that, collective knowledge increases when using ESM thereby contributing to employee performance. Thus, it can be concluded that if Social Media in the workplace is getting better, then Employee Work Performance will be more positive.
The most dominant element of social media is the element 'I regularly use social media to maintain and strengthen communication with colleagues in my work' with a value of 1.59. This shows that the more social media is used to communicate with colleagues in completing work, it will lead to increasing employee performance. Employee performance will be increased by the use of social media. Employee performance is referred to as: better performance than accepted level, perform better than expected, try extra at work, make a lot of effort in carrying out work, trying to work as hard as possible, and the best quality of work. The most dominant element of Employee performance is the element 'I often put in extra effort in my work' and 'I intentionally expend a great deal of effort in carrying out my job' with a same value of 1.21.
In testing hypothesis 4, the results of data processing show that the Trust does not mediate the influence of Social Media on Employee Work Performance, has a coefficient of -0.008. Test results on the CR value (t-value) of -0.145 smaller than 1.96 or p-value of 0.884> 0.05 (alpha 5%), it can be concluded statistically at a 95% confidence level Trust does not mediate the positive influence of Social Media on Employee Work Performance. Hypothesis 4 test results, show that the Trust does not significantly mediate the influence of Social Media on Employee Work Performance. Although previous research has found that the use of Enterprise Social Media (ESM) can create and strengthen social networks and interactions that can create trust among colleagues (Fang and Chiu, 2010), and (Cao et al., 2016) found that trust based on knowledge transfer on social media has increased the ability of employees to solve problems in an effective way that will ultimately improve work performance, but the results of this study do not support. Means that the existence of Trust among employees does not strengthen the influence of Social Media on Employee Work Performance.
Then also testing whether the Trust variable can mediate the effect of Social Media variables on Employee Work Performance. The results obtained using the SOBEL test can be seen as below:

CONCLUSION
Based on the results of the research, conclusions can be drawn as follows: (1). There is an influence of Social Media on Trust among employees. This shows that the more Social Media is used to communicate with colleagues in completing work, the more trust between employee increases. (2). There is no influence of Trust on Employee Work Performance. This shows that Trust among employees does not affect Employee Work Performance. This shows that employees' perceptions of virtual friends in the social media community always pay attention to their interests; pay attention to the situation; and care about what happens to him, it will not affect Employee Performance. So is the perception of employees who believe that virtual friends will do work with professionals; and trust in their competence, it will also not affect EmployeeWork Performance. (3). There is an influence of Social Media on Employee Work Performance. This shows that the more social media is used to communicate with colleagues in completing work, the more it improves employee Work performance. (4). Trusts do not mediate the influence of Social Media on Employee Work Performance. This shows that Trust among employees does not strengthen the influence of Social Media on Employee Work Performance.
Managerial Implications. Companies should continue to build networks in the workplace that connect experts in various social media so that it will improve employee performance effectively. Managers should encourage employees to use social media to gain knowledge and information. Managers continue to retain employees to communicate with colleagues using social media on a regular basis. Managers need to promote how communication on social media will build trust between employees in the workplace, so employees will continue to maintain and further strengthen that communication.
Suggestions for Further Research. This research was conducted with a limited number of samples and samples are employees of various industries. For further research, it can add the number of samples and with certain criteria, namely employees who have used social media in completing their work. Every job in every industry will be different in the use of social media. Therefore, in subsequent studies, samples were taken from employees in the same industry so that results can be more accurate. Addition of variables that mediate the influence of Social Media on Employee Performance, can be done. Such as knowledge transfer variables (Offong and Costello, 2017), network ties (Cao et al., 2016), or communication (Yu et al., 2018).