Meta Analysis of the Influence of Training on Work Productivity

Increasing employee competency will affect employee performance so that the employee's work becomes more effective and efficient. So that it is efficient and effective. The purpose of this research is to do a meta-analysis of the effect of training on work productivity. This research method uses meta-analysis, a type of multiple regression research, in which the researcher conducts a meta-analysis of 10 published journals with similar research. The results of this meta-analysis study show that there is a significant effect between training on employee work productivity. Based on the results of the Meta test Analysis of the correlation coefficient, the relationship between variables is strong because it is close to number 1. Based on the meta test Correlation coefficient analysis and effect size test on multiple regression tests, Count T tests, Count F tests, R Square tests and correlation tests. The data is positively heterogeneous and there is a significant influence or relationship between variables.

Increasing employee competency will affect employee performance so that the employee's work becomes more effective and efficient. So that it is efficient and effective. The purpose of this research is to do a meta-analysis of the effect of training on work productivity. This research method uses meta-analysis, a type of multiple regression research, in which the researcher conducts a meta-analysis of 10 published journals with similar research. The results of this metaanalysis study show that there is a significant effect between training on employee work productivity. Based on the results of the Meta test Analysis of the correlation coefficient, the relationship between variables is strong because it is close to number 1. Based on the meta test Correlation coefficient analysis and effect size test on multiple regression tests, Count T tests, Count F tests, R Square tests and correlation tests. The data is positively heterogeneous and there is a significant influence or relationship between variables.

INTRODUCTION
research, we will find out through the journal meta-analysis, so it is hoped that we can find out how much influence the training provided has on employee work productivity through meta-analysis. Meta-analysis is a statistical technique to combine the results of two or more similar studies in order to obtain a combination of quantitative data. In this study, the effect of training variables on productivity variables was sought. So it is hoped that it can be known whether based on the meta-analysis test, the data is homogeneous or heterogeneous and how the results of the meta-analysis are the correlation coefficients. Does the meta analysis result have a strong relationship between the independent variables and the dependent variable or vice versa the relationship between the two variables is weak.

LITERATURE REVIEW
The lack of employee productivity can be seen from the ability of employees to speak English. Furthermore, the low work productivity of employees can be seen from the slow check-in and check-out process. The instructor only conveys material related to the field of training provided. Furthermore, not all departments carry out training. The research objective was to analyze the effect of training on employee work productivity at the Whiz Prime Hotel Padang. This type of research is classified as a quantitative research with a causal associative approach. The type of data used in this research is primary data. The research population is all employees who work at Hotel Whiz Prime Padang. The research sample consisted of 68 people, the data was collected by distributing questionnaires which had been tested for validity and reliability. the sampling technique is a saturated sample.
The data analysis technique used a simple linear regression test with the help of the SPSS Computer Program version 16.00. Based on the research that has been done, the following results are obtained: (1) Training is in the good category (75.00%), (2) Employee work productivity is in the bad category (64.71%) (3) The R Square value is 0.497 . This means that training affects employee work productivity by 49.7% and 50.3% is influenced by other factors. The results of a simple linear regression test obtained an calculated f value of 65,293 with a significance level of 0.000 <0.05, meaning that the training variable contributes significantly to work productivity. Furthermore, a t value of 8,080 was obtained with a significance level of 0,000 <0.05, so there was a significant relationship between the two variables. Then the regression coefficient value is 1,761 with a value at sig. 0.000 <0.05. This means that every increase of 1 training unit will increase 1,761 units of employee work productivity.
The effect of training on employee work productivity at PT PP. London Sumatra Indonesia, Tbk Medan. This study used a sample of 94 respondents and data collection was obtained from the questionnaire method and literature study and the data analysis method was computerized. The results of this study are evident from: test the validity and reliability of the instrument, correlation coefficient (R), coefficient of determination (R2), f test, t test, and multiple linear regression analysis. From the results of the validity analysis, all statement items are said to be valid because r count > r table and reliability test results, all statement items are said to be reliable because the Cronbach's alpha value is > 0.06. For the correlation test the R value is 0.945 meaning that there is a strong relationship between the independent variables and the dependent variable, the R square results are able to explain 89.2% of the dependent variable, while the remaining 10.8% is influenced by other variables not discussed in this study, the results F test with F count> F table (77.353> 1.99) then Ho is rejected, meaning that all independent variables simultaneously (simultaneously) affect the dependent variable. (2.717>1.663), training techniques (3.503>1663), and training assessment methods (10.262>1.663), H0 is rejected instructor variables, training techniques and assessment methods partially affect work productivity (dependent).
(JessicaDeboraSibarani1, 2022) the regression coefficient of job training has a t count of 4.041 > the t table value of 1.97612. then for a significance level of 0.000 greater than 0.05, the results above can be explained for the research hypothesis to be accepted. Motivation has a t count of 4.265 > of t table of 1.97612 and a significance level of 0.000 < of 0.05, this means that the research hypothesis is accepted. Work stress has a t count of -0.902> from t table 1.97612 and a significance level of 0.365 <0.05, this means that the research hypothesis is accepted. F test results The F test is carried out with the aim of testing the effect of the independent variables on the dependent and independent variables simultaneously. Effect of job training (X1) on work productivity (Y).
The results of the t test can be obtained by calculating the t value of 4.041 > 1.976 and the significance value is 0.000 > 0.05 showing positive and significant t test results for the work productivity variable, meaning that the hypothesis of the study is accepted. The results of the job training test (X1) obtained a coefficient value of 0.365, statistically the results of this test also provide evidence of a significant and positive effect of job training on work productivity. The results of the obtained t count are 4.262 > 1.976 and the significance value is 0.010 <0.05 indicating that there is a facility for motivation in a significant and positive way, meaning that the hypothesis of the study is accepted. The results of the work motivation test (X2) obtained a coefficient value test of 0.356, statistically the test results have a positive and significant effect of motivation on work productivity. Effect of job training (X1), motivation (X2) and work stress (X3) on work productivity (Y). The results of f are obtained at the calculated F value of 21.573 and the F table is worth 3.06, it can be seen for the large calculated F value of the F table then for the significance value get that is 0.000 this significance is certainly lower and smaller than what is obtained at alpha of 0 .05, therefore the conclusion drawn is that Ho was then rejected and Ha was accepted (Alfis Vernando1, 2021) The Effect of Job Training and Workload on Employee Work Productivity at the XYZ Mother and Child Hospital. Test of the Coefficient of Determination (R2) This test is determined by the magnitude of the adjusted R2 value to determine the magnitude of the variation in work productivity variables that can be explained by job training and workload variables. Based on the results of the t test for the job training variable, t count = 4.646 with a significance value of 0.000 <0.05, then Ho is rejected and Ha is accepted. This shows that partially H1 which states that there is an effect of job training on employee work productivity at the xyz Jombang mother and child hospital. Based on the results of the t test for the job training variable, t count = 4.392 with a significance value of 0.000 <0.05, then Ho is rejected and Ha is accepted. This shows that partially H1 which states that there is an effect of workload on the work productivity of employees at the xyz Jombang mother and child hospital. Statistical test F Statistical test F or hypothesis testing simultaneously aims to determine the effect of two independent variables.

METHODOLOGY
This type of research used in this research is multiple regression research. The research method uses meta-analysis, namely by collecting similar journals using primary data. What is meant by meta-analysis according to (Nindrea, 2016) is a method that conducts an in-depth analysis of a topic from several valid studies that are put together so that it resembles a large study using statistical analysis. Where the author does a gradual screening of the journal. to find the data using publish Perish. Initially, we searched for 1000 data on Google Scholar with the key words influence, training and productivity. then screening data is carried out according to the required variables, namely training (Independent) and productivity (Dependent) variables. The Publish perish application is linked to the Vosviewer application, so that 48 journals are obtained. After opening one by one, relatively similar data was obtained, the same methods and variables were taken and 10 journals were obtained. then recapitulated into one format. You can see the flow in the image below; Picture 1. Data Picture 2. Data 2 Picture 3. Data 3 The analysis in this meta-analysis method is to see the heterogeneity and effect size of the 10 journals which have relatively the same variables. also in this method, the researcher reviews, examines, criticizes, so that he can find out what writings are not in accordance with research theory both from abstracts, introductions, methods, results, discussion, and conclusions. The gap was added by researchers according to research theories according to research experts. Ingredients used are several journals that are in accordance with the topics that researchers take. After that, inputs were given and recommendations were given. What is meant by Meta-Analysis according to Sutjipto (1995) is: "an attempt to summarize various research results quantitatively. in other words, meta-analysis as a technique is intended to re-analyse the results of the research which are processed statistically based on primary data collection. 10 screening data were processed using the Jamovi application. References use Mendeley, so that they are neatly arranged, so that they are structured in their placement together.  Forest Plot

DISCUSSION
Meta-analysis of the correlation coefficient based on the standard error value, obtained data 0.149, where the interpretation of the results of the metaanalysis test, the smaller the standard error value, the better, meaning that there is a correlation coefficient relationship after merging 10 journal data with similar data. Meanwhile, if you look at the P value, there is a value of < 0.001, meaning that the combined p value is based on the results of meta-analysis testing using the correlation coefficient with the Jamovi application, a value of < 0.001 is obtained. the value has a significant relationship if it does not exceed 0.05, then, it can be interpreted that there is a significant relationship combined data from 10 journals that have passed the Meta Analysis calculation using Jamovi. The observed Fisher r-to-z correlation coefficients ranged from 0.2877 to 1.7828, with mostly positive estimates (100%). Fisher's estimated mean r-to-z change correlation coefficient based on the random-effects model is \hat{\mu} = 0.8005 (95% CI: 0.5078 to 1.0931). Therefore, the mean result is significantly different from zero (z = 5.3610, p < 0.0001). According to the Q-test, the actual results appear to be heterogeneous (Q(8) = 149.8552, p <0.0001, tau² = 0.1870, I² = 93.8284%). The 95% prediction interval for the actual result is given by -0.0961 to 1.6971. Therefore, even if the average result is expected to be positive, in some studies the actual result may be negative.
Meta analysis based on effect size for mean estimated regression coefficient of 84.0, changing regression coefficient 95% CL: 65.307 to 1.093. Z mean results differed significantly 8.79, P < 0.0001. According to the Q-test the actual results appear Q= 9155.667, P < 0.001, tau2 30.188, 12 = 430.9786. The 95% prediction interval for the actual outcome is given by 65,737 to 102,788. interpretation of the regression coefficient of heterogeneous data if the Tau2 value > 0 regardless of Q, it can be said that the regression coefficient data is stated to be heterogeneous Meta-analysis based on effect size for T arithmetic estimates Metaanalysis based on effect size for the average estimated regression coefficient of 84. interpretation of the correlation coefficient of heterogeneous data if the value of Tau2 > 0 regardless of Q, it can be said that the data correlation coefficient is stated to be heterogeneous Of the ten journals that were analyzed by conducting this Meta-Analysis, it was found that all of these journals used a quantitative statistical approach. Where almost all studies tested the regression coefficient, T count, significance test, F count, and correlation coefficient. This means that most of it, namely as much as 90% did the test, only 10% did not do the statistical test, that is, did not do the F count test. Based on manual analysis per journal item, it can be seen whether there is a correlation and the effect of variable x (training) on variable Y (productivity). but by looking at the results of the significance test it can be seen whether training has a significant effect on productivity, if Sig < 0.05, then the result is significant. If you look at the significance test data for all journals <0.05, it means that all of the studies conducted by the meta-analysis are all significant, meaning that there is a significant effect between training on work productivity. Meanwhile, based on the correlation test, there is a relationship between training and work productivity, this relationship shows how strong the independent variable and the dependent variable are. The value is between -1 to 1. The correlation results are increasingly showing the number 1, meaning that the relationship is strong, but if it is close to -1, then the relationship is weak.
CONCLUSIONS AND RECOMMENDATIONS 1. The results of the Meta-analysis of the correlation coefficient, there is a significant relationship between the training variable (Independent) and the dependent variable (Productivity), with a standard error value of 0.149 and a significance value of <0.05. results appear heterogeneous, the average result is positive. This means that there is a significant relationship between the independent variable and the dependent variable with positive results 2. The results of the meta-analysis test with the effect size of the regression coefficient, it is known that the data is heterogeneous and there is a significant relationship between the independent variable (Training) and the dependent variable (Productivity) 3. The results of the meta-analysis test with the effect size of the T test, it is known that the data is heterogeneous and there is a significant relationship between the independent variable (Training) and the dependent variable (Productivity) 4. The results of the meta-analysis test with the effect size F test, it is known that the data is heterogeneous and there is a significant relationship between the independent variable (Training) and the dependent variable (Productivity) 5. The results of the meta-analysis test with the effect size R Square, it is known that the data is heterogeneous and there is a significant relationship between the independent variable (Training) and the dependent variable (Productivity) 6. The results of the meta-analysis test with the correlation coefficient effect size, it is known that the data is heterogeneous and there is a significant relationship between the independent variable (training) and the dependent variable (productivity)