Syntax Literate: Jurnal Ilmiah Indonesia p�ISSN: 2541-0849 e-ISSN: 2548-1398

Vol. 7, No. 8, Agustus 2022

�WORK FROM HOME ANALYSIS OF EMPLOYEE PRODUCTIVITY MEDIATED BY WORK LIFE BALANCE AND JOB SATISFACTION FOR FEMALE EMPLOYEES DURING THE COVID-19 PANDEMIC� (STUDYING AT BANK MANDIRI SAMARINDA AREA)�

Panji Sukma Nugraha

Universitas Mulawarman Samarinda, Indonesia

Email: [email protected]

Abstract

Work From Home (WFH) is referred to as a modern work concept where employees do not have to commute to work and can complete work remotely. WFH has a major influence on work results or employee productivity, especially for women. The purpose of this paper is to identify, analyze and prove the effect of work from home on work life balance and job satisfaction to achieve employee productivity of Bank Mandiri employees in the Samarinda area. This research was conducted using quantitative methods using population instruments and a sample of 110 employees of Bank Mandiri Samarinda. Data collection is done by using a questionnaire. The analysis method uses the Structural Equation Model with the Partial Least Square (SEM-PLS) approach, which is tested with the outer model (measurement model) and inner model test (structural model). The results of the study show that work-life balance has a non-significant positive effect on employee productivity, meaning that it increases work-life balance and slightly increases employee productivity. Employees get a balance of satisfaction, namely job satisfaction and satisfaction in being able to do household work, but this does not produce good performance.

Keywords: Work from Home, work life balance, job satisfaction, employee productivity

Introduction

Since February 2020, the country of Indonesia has been designated as a country affected by the COVID-19 virus pandemic. The entry of the COVID-19 outbreak into Indonesia is a national disaster that has caused many Indonesian people to be affected by the COVID-19 virus with various symptoms ranging from mild to death. The speed of the spread of the COVID-19 virus is caused by the main factor, namely the gathering of many people in one area, which makes the COVID-19 virus easily and quickly spread to anyone.

To prevent the spread of the COVID-19 virus, the majority of them have locked down or closed the area in a tight and limited manner. This method is recommended by the WHO, which directly supervises and closely monitors all countries in the world in order to control the spread of the COVID-19 virus. The Indonesian government does not implement a lockdown for several considerations, one of which is economic considerations.

Almost all sectors of the Indonesian economy in 2020 experienced a significant decline, and not even a few were able to continue their activities. The trade, manufacturing, tourism, export-import, banking, mining, plantation, oil and gas sectors, as well as other sectors, experienced losses and decreased productivity. Companies are forced to impose restrictions on working hours and reduce the number of employees who are allowed to work in one area or room.

The banking sector is the most stringent in implementing social distancing because, as a financial services company, it always provides various services to customers for various purposes. In addition, the majority of offices in the banking industry are closed rooms with air conditioner (AC) cooling, which creates a great potential for the rapid transmission of the COVID-19 virus. For this reason, since February 2020, many banks have implemented a work-from-home system on a scheduled basis for their employees.

Working at home is the dream of many employees where they can work while doing household tasks that cannot be done if they work in the office from morning to evening. On the other hand, working at home is not easy for some employees, especially female employees or female employees who are married. At the same time they have to do their job duties is not an easy thing to do. Women workers who are not able to balance work and family life while working from home will feel dissatisfied with their work and even experience work stress because of the demands of the role that must be carried out at the same time. However, the positive side of working from home is the fulfillment of a good quality of family life and personal life and can make female workers more motivated to work, (Dua & Hyronimus, 2020).

Many female employees hope that they can have a balance in their lives where they can stay committed to the work they do at home and, at the same time, get time to care for and be close to their families. WFH is a dream for career women who have multiple roles, because they think that with WFH they will be able to balance their roles, apart from being a supporter of the family economy, but also carry out their nature as housewives (Komalasari et al., 2020).

Based on the description and phenomena above, researchers are interested in conducting a Work from Home Analysis of Employee Productivity Mediated by Work Life Balance and Job Satisfaction for Female Employees During the COVID-19 Pandemic, with a specific case study at Bank Mandiri in the Samarinda area.

Research Method

This research is quantitative research. Sugiyono (2010) suggested that quantitative research is research based on the philosophy of positivism, used to examine certain populations or samples. Sampling techniques are generally carried out homogeneously, data collection uses research instruments, and data analysis is quantitative in nature with the aim of testing hypotheses that have been set.

This research includes correlation research. Correlation research is research conducted by researchers to determine the level of relationship between two or more variables without making changes, additions, or manipulations to existing data (Arikunto, 2010). This study was conducted to determine the effect of working from home on employee productivity, mediated by work-life balance and job satisfaction. The method used is the Population and Sample method.

1. Population

The population is the entire research subject (Arikunto, 2010). The population is a generalization area consisting of objects or subjects that have certain qualities and characteristics determined by researchers to be studied and then drawn conclusions from (Sugiyono, 2010). The population in this study were employees of Bank Mandiri in the Samarinda area, totaling 350 employees.

2. Sample

In this study, the sampling technique used is non-probability sampling with a purposive sampling technique. According to Sugiyono (2016), purposive sampling is a sampling technique of data sources with certain considerations. The reason for using the purposive sampling technique is that not all samples have criteria that match the phenomenon under study. Therefore, the authors chose the purposive sampling technique, which stipulates certain considerations or criteria that must be met by the samples used in this study.

In this study, the sample consists of employees who meet certain criteria. The criteria used as research samples are:

a. Employees of Bank Mandiri in the Samarinda area

b. Employee status is married or married.

According to Hair et al. (2010), the number of samples is taken to be a minimum of 5�10 times the number of indicators. The number of indicators in this study is 22 indicators, so the sample in this study is 5 x 22, namely 110 respondents, who are felt to be sufficient to represent the population.

Results of Analysis and Discussion

1. Results of Analysis

a. Data Analysis

Data analysis was performed using Smart PLS 3.0 software. The analysis was carried out using the outer model (measurement model) and inner model (structural model). The outer model test (measurement model) is to assess the validity and reliability of the research instrument, while the inner model (structural model) is to test the fit, goodness of fit model, and hypothesis testing.

1) Outer model (Measurement model)

The results of the outer model test are as follows:

Gambar 1 Outer Model step 1

Source: primary data processed

From the results of the outer model with Smart PLS, the loading factor value of JS5 is 0.300. The valid loading factor value is 0.5, ideally 0.7, and the AVE value should be greater than 0.5 (Hair et al., 2010). For this reason, the JS5 indicators are not relevant to the research model and are excluded from it. Next, the second stage of analysis is carried out:

Gambar 2 Outer Model tahap 2

Sumber: data primer diolah

2) Convergen validity test

Convergent validity is a number of indicators measuring or representing one latent variable and which underlies the existence of the latent variable. The purpose of the convergent validity test is to ensure that the items used in this study can be understood by the respondents so that there are no errors in filling them out. Convergent validity is tested using outer loading, namely by looking at the coefficients between the variables and their items, provided that the loading value is said to be valid if > 0.5 (Hair et al., 2010). The loading value of each variable is as follows:

Tabel 1

�Loading Factor score

Variables and Indicators Relationship	Original Sample (O)	Sample Mean (M)	Standard Deviation (STDEV)	T Statistics� (|O/STERR|)	P Value
EP1 <- EP	0,862	0,860	0,027	31,498	0,000
EP2 <- EP	0,840	0,838	0,031	27,343	0,000
EP3 <- EP	0,854	0,854	0,025	34,105	0,000
EP4 <- EP	0,836	0,835	0,029	28,923	0,000
EP5 <- EP	0,896	0,893	0,022	40,792	0,000
JS1 <- JS	0,861	0,860	0,026	32,565	0,000
JS2 <- JS	0,878	0,877	0,022	39,820	0,000
JS3 <- JS	0,877	0,878	0,023	38,823	0,000
JS4 <- JS	0,818	0,818	0,032	25,567	0,000
JS6 <- JS	0,757	0,756	0,046	16,314	0,000
WFH1 <- WFH	0,734	0,732	0,055	13,267	0,000
WFH2 <- WFH	0,831	0,829	0,035	23,569	0,000
WFH3 <- WFH	0,800	0,799	0,037	21,455	0,000
WFH4 <- WFH	0,806	0,808	0,030	26,826	0,000
WFH5 <- WFH	0,749	0,746	0,048	15,656	0,000
WLB1 <- WLB	0,822	0,822	0,038	21,479	0,000
WLB2 <- WLB	0,819	0,817	0,032	25,525	0,000
WLB3 <- WLB	0,818	0,817	0,035	23,238	0,000
WLB4 <- WLB	0,809	0,814	0,039	21,000	0,000
WLB5 <- WLB	0,841	0,841	0,032	26,467	0,000
WLB6 <- WLB	0,795	0,793	0,043	18,407	0,000

Source: primary data processed (2021)

Based on table 5.1, the relationship between variables and indicators can be described as follows:

a) Work from home (WFH) Variable

The WFH₂ loading factor value of 0.831 is the indicator that is most influenced by the Work from Home (WFH) variable, and WFH₁, with a loading factor value of 0.734, is the smallest indicator influenced by the Work from Home (WFH) variable. It can be seen that all indicators on the Work from Home (WFH) variable in this study have a loading factor value greater than 0.5. This shows that the indicators WFH₁, WFH₂, WFH₃, WFH₄ and WFH₅ have a high level of validity, thus meeting convergent validity and can then be used for testing research hypotheses.

b) Work life balance (WLB)Variable

The WLB₅ loading factor value of 0.841 is the indicator that is most influenced by the work life balance (WLB) variable, and WLB₆, with a loading factor value of 0.795, is the indicator that is least influenced by the work life balance (WLB) variable. It can be seen that all indicators of the work-life balance (WLB) variable in this study have a loading factor value greater than 0.5. This shows that the indicators WLB₁, WLB₂, WLB₃, WLB₄, WLB₅ and WLB₆ have a high level of validity, so they meet convergent validity and can then be used for testing research hypotheses.

c) Job satisfaction (JS)Variabel

The value of loading factor JS₂ of 0.878 is the indicator that is most influenced by the variable Job satisfaction (JS) and JS₆, with a value of loading factor of 0.757, is the indicator that is least influenced by the variable Job satisfaction (JS). It can be seen that all indicators of the job satisfaction (JS) variable in this study have a loading factor value greater than 0.5. This shows that the indicators JS₁, JS₂, JS₃, JS₄ and JS₆ have a high level of validity, thus meeting convergent validity and can then be used for testing research hypotheses.

3) �Employee productivity (EP)Variabel

The EP₅ loading factor value of 0.896 is the indicator that is most influenced by the employee productivity (EP) variable, and EP₄ with a loading factor value of 0.836, is the smallest indicator influenced by the employee productivity (EP) variable. It can be seen that all indicators on the employee productivity (EP) variable in this study have a loading factor value greater than 0.5. This shows that the indicators EP₁, EP₂, EP₃, EP₄, and EP₅ have a high level of validity, thus meeting convergent validity and can then be used for testing research hypotheses.

4) Discriminant validity test

Discriminant validity is a concept that states that two different variables should be able to show adequate differences. This discriminant validity is measured by cross loading, namely the value of loading items on the variable must be greater than the loading stated in table 5.2 as follows.

Tabel 2

Cross Loading

Indicator	EP	JS	WFH	WLB
EP1	0,862	0,764	0,728	0,702
EP2	0,840	0,740	0,750	0,713
EP3	0,854	0,805	0,644	0,733
EP4	0,836	0,773	0,713	0,774
EP5	0,896	0,811	0,711	0,741
JS1	0,839	0,861	0,732	0,750
JS2	0,822	0,878	0,712	0,708
JS3	0,792	0,877	0,658	0,734
JS4	0,704	0,818	0,649	0,685
JS6	0,633	0,757	0,649	0,841
WFH1	0,531	0,518	0,734	0,569
WFH2	0,648	0,627	0,831	0,705
WFH3	0,620	0,594	0,800	0,601
WFH4	0,696	0,716	0,806	0,687
WFH5	0,720	0,698	0,749	0,694
WLB1	0,647	0,696	0,661	0,822
WLB2	0,693	0,711	0,669	0,819
WLB3	0,724	0,695	0,704	0,818
WLB4	0,763	0,750	0,765	0,809
WLB5	0,633	0,757	0,649	0,841
WLB6	0,712	0,712	0,631	0,795

Source: primary data processed (2021)

Based on table 5.2, it can be analyzed as follows:

a) Analysis of Discriminant Validity indicator variable Work from home (WFH)

The loading value of each item on the work from home (WFH) variable is greater than the cross-loading value of each item on the other variables. The value of loading items WFH₁ (0.734), WFH₂ (0.831), WFH₃ (0.800), WFH₄ (0.806), and WFH5 (0.749) is greater than the value of cross loading items for each item variable WLB, JS, and EP. It can be concluded that all constructs or latent variables already have good discriminant validity where the indicators in the construct indicator block are better than indicators in other blocks. So it can be said that WFH₁, WFH₂, WFH₃, WFH₄ and WFH₅ are indeed items formed by the work from home (WFH) variable. Based on the cross-loading test, this research has met the criteria for discriminant validity, which can then be used for hypothesis testing.

b) Analysis of Discriminant Validity indicator variabel Work life balance (WLB).

The loading value of each item on the work life balance (WLB) variable is greater than the cross-loading value of each item on the other variables. The value of loading items WLB₁ (0.822), WLB₂ (0.819), WLB₃ (0.818), WLB₄ (0.809), WLB₅ (0.841) and WLB₆ (0.795) is greater than the value of cross loading items for each item variable EP, JS, and WFH. It can be concluded that all constructs or latent variables already have good discriminant validity where the indicators in the construct indicator block are better than indicators in other blocks. So it can be said that WLB₁, WLB₂, WLB₃, WLB₄, WLB₅ and WLB₆ are indeed items formed by the work life balance (WLB) variable. Based on the cross-loading test, this research has met the discriminant validity criteria, which can then be used for hypothesis testing.

c) Analysis of Discriminant Validity indicator variabel Job satisfaction (JS).

Based on table 5.2, the loading value of each item on the job satisfaction (JS) variable is greater than the cross-loading value of each item on the other variables. The value of loading items JS₁ (0.861), JS₂ (0.778), JS₃ (0.877), JS₄ (0.818) and JS₆ (0.757) is greater than the cross-loading item value of each item variable EP, WLB, and WFH. It can be concluded that all constructs or latent variables already have good discriminant validity where the indicators in the construct indicator block are better than indicators in other blocks. So it can be said that JS₁, JS₂, JS₃, JS₄ and JS₆ are indeed items formed by the job satisfaction (JS) variable. Based on the cross loading test, this research has met the criteria for discriminant validity, which can then be used for hypothesis testing.

d) Analysis of Discriminant Validity indicator variabel Employee productivity (EP).

The loading value of each item on the employee productivity (EP) variable is greater than the cross-loading value of each item on the other variables. The value of item loading EP₁ (0.862), EP₂ (0.839), EP₃ (0.854), EP₄ (0.836), and EP₅ (0.896) is greater than the cross loading item value of each item variable WFH, WLB, and JS. It can be concluded that all constructs or latent variables already have good discriminant validity where the indicators in the construct indicator block are better than indicators in other blocks. So it can be said that EP₁, EP₂, EP₃, EP₄, and EP₅ are indeed items formed by the employee productivity (EP) variable. Based on the cross loading test, this research has met the criteria for discriminant validity, which can then be used for hypothesis testing.

Then, the discriminant validity test was carried out using the AVE (average variance extracted) value. The AVE value describes the variance or diversity of the manifest variables that the latent construct can have. The greater the variance or diversity of the manifest variables that can be contained by the latent construct, the greater the representation of the manifest variable on the latent construct.

The test is carried out by comparing the AVE roots with the influence between latent variables and the AVE value must be greater than 0.5.

Tabel 3

Average Variance Extracted

Variabel	EP	JS	WFH	WLB	AVE	√ AVE
EP	0,858				0,736	0,858
JS	0,908	0,839			0,705	0,839
WFH	0,827	0,811	0,785		0,616	0,785
WLB	0,854	0,882	0,835	0,817	0,668	0,817

Source: primary data processed (2021).

From Table 5.3, it is known that the AVE value of each construct is above 0.5. Therefore, there is no problem with convergent validity in the model being tested, so the constructs in this research model can be said to have good discriminant validity. The square root value of AVE for each construct is greater than the correlation value, so that the construct in this research model can still be said to have good discriminant validity.

5) Cronbach Alpha dan Composite reliability Test

The Cronbach alpha test is used to show how reliable and accurate a measuring instrument is in assessing what should be assessed which is usually used to assess construct reliability. The Cronbach alpha value is said to be reliable if the value is > 0.7. Composite reliability test is a tool used to test the reliability between indicator blocks of the latent variables that make up it. The composite reliability value is said to be reliable if the value is > 0.7.

Tabel 4

Composite Reliability and Cronbachs Alpha score

Variabel	Cronbachs Alpha	Composite Reliability
Employee productivity (EP)	0,910	0,933
Job satisfaction (JS)	0,894	0,922
Work from home (WFH)	0,844	0,889
Work life balance (WLB)	0,901	0,924

Source: primary data processed (2021).

Based on table 5.4 the value of Cronbach's alpha and composite reliability of all variables > 0.7 so that it has fulfilled the reliability test, then all variables can be used to test the hypothesis.

6) Inner model test (structural model)

a. Variant Analysis (R²) or Determination Test

The determination test aims to test the influence model of each independent variable on the dependent variable. The value of each dependent variable R² is as follows:

Tabel 5

R Square Value(R²)

Variabel	�R Square
Employee productivity (EP)	0,851
Job satisfaction (JS)	0,658
Work life balance (WLB)	0,698

Source: primary data processed

In table 5.5, the value of R² for the dependent variable work life balance (WLB) is 0.698, meaning that 69.8% of changes in the work life balance variable are influenced by the independent work from home (WFH) variable, while the remaining 30.2% is influenced by other variables not discussed in this study. The value of R² for the dependent variable job satisfaction (JS) is 0.658, meaning that 65.8% of changes in the job satisfaction (JS) variable are influenced by the independent work from home (WFH) variable, while the remaining 34.2% is influenced by other variables not discussed in this research.

The R² value for the dependent variable employee productivity (EP) of 0.851 means that 85.1% of changes in employee productivity (EP) are influenced by work from home (WFH), work life balance (WLB), and job satisfaction (JS), while the remaining 14.9% are influenced by other variables not discussed in this study.

b. Hypothesis Test

The results of Smart PLS 3 after boothstrapping are as followst:

Pitcure 3 Full Model Boothstrapping

Source: primary data processed

Hypothesis testing is used to test the hypothesis proposed in this study.

Tabel 6

Path coefficient

Relationship Between Variablesl	Original Sample (O)	Sample Mean (M)	Standard Deviation (STDEV)	T Statistics (|O/STERR|)	P Value	Ket.
JS -> EP	0,622	0,625	0,094	6,604	0,000	Sig.
WFH -> EP	0,220	0,216	0,084	2,616	0,009	Sig.
WFH -> JS	0,811	0,812	0,033	24,268	0,000	Sig.
WFH -> WLB	0,835	0,836	0,030	27,700	0,000	Sig.
WLB -> EP	0,122	0,123	0,090	1,357	0,175	-
WFH -> JS -> EP	0,505	0,508	0,086	5,861	0,000	Sig.
WFH -> WLB -> EP	0,102	0,102	0,075	1,358	0,175	-

Source: primary data processed

Based on table 5.6 with a significance level of 5% and a value of df = 106, the hypothesis testing can be explained as follows:

1) Testing the first hypothesis (H₁), namely Work from home (WFH) has a positive and significant effect on increasing Employee productivity (EP). The value of the coefficient of influence between work from home on employee productivity is positive, which is 0.220 which indicates that the direction of the relationship between work from home and employee productivity is positive. The t-count value of 2.616 > 1.98 means that it is significant, then the first hypothesis (H¹) is proven true and accepted, which means that an increase in work from home can significantly increase employee productivity.

2) Testing the second hypothesis (H₂), working from home (WFH) has a positive and significant effect on increasing work-life balance (WLB). The value of the coefficient of influence between work from home and work life balance is positive, which is 0.835, which indicates that the direction of the relationship between work from home and work life balance is positive. The t-count value of 27.700 > 1.98 means that it is significant. Then the second hypothesis (H²) is proven true and accepted, which means that increasing work from home can significantly improve work-life balance.

3) Testing the third hypothesis (H₃), working from home (WFH) has a positive and significant effect on increasing job satisfaction (JS). The coefficient value of the influence of work from home on job satisfaction is positive, which is 0.811, which indicates that the direction of the relationship between work from home and job satisfaction is positive. The t-statistical value of 24,268 > 1.98 means significant, then the second hypothesis (H₃) is proven true and accepted, which means that increasing work from home can significantly increase job satisfaction.

4) Testing the fourth hypothesis (H₄), work-life balance (WLB) has a positive and significant effect on increasing employee productivity (EP). The coefficient value of the effect of work life balance on employee productivity is positive, namely 0.122, which indicates that the direction of the relationship between work life balance and employee productivity is positive. The t-count value of 1.357 > 1.98 means that it is not significant. Then the second hypothesis (H⁴) is not proven true and is rejected, which means that an increase in work-life balance cannot significantly increase employee productivity.

5) Testing the fifth hypothesis (H₅) Job satisfaction (JS) has a positive and significant effect on increasing employee productivity (EP). The coefficient value of the effect of job satisfaction on employee productivity is positive, which is 0.622, which indicates that the direction of the relationship between job satisfaction and employee productivity is positive. The t-count value of 6.604 > 1.98 means that it is significant, then the second hypothesis (H₅) is proven true and accepted, which means that an increase in job satisfaction can significantly increase employee productivity.

6) Testing the sixth hypothesis (H₆) Work life balance (WLB) mediates the effect of Work from home (WFH) on Employee productivity (EP). This hypothesis aims to test whether work life balance is a mediating variable or not by using the VAF (variance account for) method (Hair et al., 2014:225) with the following steps:

Figure 4. Simple mediation model

�������������

The VAF value is calculated by the following formula:

����������������� ������ a x b

�VAF = -------------------���

����� ������������� ����(a x b) + c

If the VAF value <20% means Y1 is not a mediating variable, VAF between 20% - 80% means partial mediation and a VAF value> 80% means full variable.

����������������������� 0,835 x 0,122

VAF = ---------------------------------

����������� ��� (0,835 x 0,122) + 0,220

�

����������� = 0,317

The VAF value of 0.31 is still in the range of values from 0.2 to 0.8, meaning work-life balance is a variable that mediates the effect between working from home (WFH) and employee productivity (EP), with the nature of the mediation being partial mediation, which means the effect of the variable independent of the dependent variable is not significant when the mediating variable is included in the model.

After work life balance (WLB) was entered as a mediating variable, there was a decrease in the coefficient of work from home (WFH) on employee productivity (EP) from 0.220 to 0.102 and the t value decreased from 27.700 to 1.358 which indicates that work life balance is capable of being a mediating variable. which reduces the indirect effect between Work from home (WFH) and Employee productivity (EP).

7) Testing the seventh hypothesis (H7) Job satisfaction (JS) mediates the effect of Work from home (WFH) on Employee productivity (EP). Testing the seventh hypothesis to determine whether job satisfaction (JS) is a mediating variable or not, the VAF (variance account for) method is used.

Based on table 5.6, the VAF value can be found as follows:

����������� ����������� 0,811 x 0,622

VAF �������� = ---------------------------------

(0,811 x 0,622) + 0,220

����������� = 0,696

�The VAF value of 0.696 is still in the range of values of 0.2-0.8, meaning that job satisfaction is a variable that mediates the effect of work from home (WFH) and employee productivity (EP), with the nature of the mediation being partial mediation, which means the influence of the independent variable on the variable. The dependent is still significant when the mediating variable is included in the model.

After job satisfaction was included as a mediating variable, there was a decrease in the coefficient of work from home (WFH) on employee productivity (EP) from 0.220 to 0.505 and the t-value increased from 2.616 to 5.861 which indicates that job satisfaction is a mediating variable that increases the indirect effect between Work from home (WFH) with Employee productivity (EP).

2. Discussion

From the results of the analysis and hypothesis testing as well as from the theoretical basis and empirical evidence, the following discussion can be carried out:

a. The effect of work from home on employee productivity

The results of the analysis show that work from home has a significant positive effect on employee productivity. This means that if the effect of work from home is increased, it will significantly increase employee productivity.

If the organization wants to increase employee productivity through the work from home variable, it starts by providing a sufficient work duration for employees who work from home. The dominant indicator affecting work from home is that the company provides sufficient time for me to complete office work from home, and the lowest effect is the house where the respondent feels comfortable doing office work.

The results of this study are relevant to research conducted by Rahman & Arif, (2021) on 100 employee respondents in Bangladesh showing that although they work according to a regular schedule, most employees feel that they have completed more work at home compared to the location.

The results of this study are also in accordance with research conducted by Ramos & Prasetyo, (2020) which resulted in the finding that there is a positive relationship between work from home and employee productivity for employees in the Philippines.

b. The effect of work from home on work life balance

The results of the analysis show that work from home has a positive and significant effect on work life balance. This means that if the effect of work from home is increased, it will increase work life balance.

If the organization wants to improve work-life balance through the work-from-home variable, it starts by providing sufficient work duration to employees who work at home. The dominant work from home indicator that affects work life balance is that the company provides sufficient time for me to complete office work from home, and the lowest effect is the home where the respondent feels comfortable doing office work.

This study is relevant to what Dua & Hyronimus, (2020) stated that the work from home variable has a positive and significant influence on the work life balance variable for female workers in the city of Ende who are married or single parents.

The results of this study are in accordance with research conducted by Putra et al., 2020 in 250 that employees do remote working so they can save their time and energy which is usually used to travel to the office, can be allocated to manage their personal lives.

c. The effect of work from home on job satisfaction.

The results of the analysis show that work from home has a positive and significant effect on job satisfaction. This means that if the effect of work from home is increased, it will increase job satisfaction.

If the organization wants to increase job satisfaction through the work from home variable, it starts by providing sufficient work duration to employees who work at home. The dominant work from home indicator that affects work life balance is that the company provides sufficient time for me to complete office work from home, and the lowest effect is the home where the respondent feels comfortable doing office work.

The results of the analysis relevant to research conducted by Bellmann & H�bler, (2020) reveal that work from home has a significant positive effect on job satisfaction for employees in the manufacturing and service industries and information providers in Germany.

The results of this study are not in accordance with research conducted by Novianti & Roz, (2020) that WFH has a positive but not significant effect on job satisfaction. These findings can be interpreted as the implementation of WFH makes a small contribution to job satisfaction.

d. The Effect of work life balance on employee productivity

The results of the analysis show that work-life balance has a positive and insignificant effect on employee productivity. This means that if the effect of work-life balance is increased, it will not have too much impact on increasing employee productivity.

If the organization wants to increase employee productivity through work-life balance variables, it starts from providing a balance of satisfaction, namely job satisfaction and satisfaction in being able to do household work. The desire of employees who expect a balance between office work and household affairs will be difficult to realize where if you want to provide high employee productivity, you have to sacrifice household affairs.

This study is in accordance with research by Ghareeb (2019), finding that poor performance in family roles leads to poor productivity at work. Poor work-life balance management will have an impact on family and work or the company.

e. The effect of job satisfaction on employee productivity

The results of the analysis show that job satisfaction has a positive and significant effect on employee productivity. This means that if the effect of job satisfaction is increased, it will increase employee productivity.

If the organization wants to increase employee productivity through variable job satisfaction, it starts with employees getting the opportunity to do something useful even though they are working from home.

The results of the data analysis are in line with research conducted by Adekanmbi, Ukpere & Adegoke, (2020) that job satisfaction significantly affects the stages of employee productivity in the manufacturing industry in Oyo State, Nigeria.

This research is in accordance with research by Soewito, (2020:2) that the ability of employee resources in encouraging effective work results is not only determined by professional attitudes, but must also be accompanied by high job satisfaction to encourage work results that satisfy all parties.

f. The effect of work from home on employee productivity through work life balance

The results of the analysis show that there is an indirect effect between work from home and employee productivity through work life balance, which means that work life balance is a mediating variable between work from home and employee productivity, although it has not been able to make a significant effect.

These results reveal that if employees achieve a high work life balance, it will sacrifice employee productivity where the demands and responsibilities of work in the company take up quite a lot of employee time and energy.

g. The effect of work from home on employee productivity through job satisfaction.

�The results of the analysis show that there is an indirect effect between work from home and employee productivity through job satisfaction, which means job satisfaction is a mediating variable between work from home and employee productivity.

If the organization wants to increase employee productivity through work from home variables mediated by job satisfaction, then it starts with employees getting the opportunity to do something useful even though they are working at home.

Conclusion

Based on the results of data analysis and understanding, this research can be concluded as follows: 1). Work from home has a significant positive effect on employee productivity, meaning that increasing work from home will significantly increase employee productivity. The dominant work from home indicator is that the company provides sufficient duration for employees to complete office work from home. Sufficient duration of office work at home can achieve high employee productivity. 2). Work from home has a significant positive effect on work life balance, which means that increasing work from home will increase work life balance. The dominant work from home indicator is that the company provides sufficient duration for me to complete office work from home. Sufficient duration of working from home provides an opportunity for employees to achieve work-life balance. 3). Work from home has a significant positive effect on job satisfaction, meaning that increasing work from home will increase job satisfaction. The dominant work from home indicator is that the company provides sufficient duration for me to complete office work from home. Doing office work at home for a sufficient duration has a high job satisfaction impact for employees. 4). Work life balance has no significant positive effect on employee productivity, meaning that increasing work life balance will not necessarily increase employee productivity. The indicator of the work life balance that dominantly affects employee productivity is that employees get a balance of satisfaction, namely job satisfaction and satisfaction in being able to do household work. Employees feel that they have sufficient time to complete their duties as workers as well as housewives but this does not necessarily result in good performance. 5). Job satisfaction has a significant positive effect on employee productivity, meaning that increasing job satisfaction will increase employee productivity. The dominant indicator of job satisfaction affecting employee productivity is that employees get the opportunity to do something useful even though they are working from home. 6). Work life balance is a variable mediating the effect of work from home on employee productivity, meaning that an increase in work life balance will make work from home decrease employee productivity. Work life balance has failed to become a mediation that increases employee productivity. 7). Job satisfaction is a mediating variable for the effect of work from home on employee productivity, meaning that increasing job satisfaction will help work from home increase employee productivity. The dominant indicator of job satisfaction affecting employee productivity is that employees get the opportunity to do something useful even though they are working from home.

BIBLIOGRAPHY

Arikunto, S. (2010). Prosedur Penelitian: Suatu Pendekatan Praktik. Rineka Cipta.

Bellmann, L., & H�bler, O. (2020). Working from home, job satisfaction and work�life balance � robust or heterogeneous links? International Journal of Manpower. Novianti, K. R., & Roz, K. (2020). Teleworking and Workload Balance on Job Satisfaction: Indonesian Public Sector Workers During Covid-19 Pandemic. APMBA (Asia Pacific Management)

Dua, M. H. C., & Hyronimus. (2020). Pengaruh Work From Terhadap Work Work ife Balence Pekerja Perempuan di Kota Ende. Jurnal Ilmiah Manajemen Bisnis Dan Inovasi Universitas Sam Ratulangi.

Hair, J., William, B. C., Barry, B. J., & Rolph, A. E. (2010). Overview of Multivariate Methods (7th ed.). Pearson Prentice.

Komalasari, Y., Putu, N., & Krismawintari, D. (2020). Paradigma Work From Home Dalam Pengembangan Karir Wanita. Universitas Dhyana Pura, November.

Novianti, K. R., & Roz, K. (2020). Teleworking and Workload Balance on Job Satisfaction: Indonesian Public Sector Workers During Covid-19 Pandemic. APMBA (Asia Pacific Management and �, 1(December), 8997. Ghareeb, A. E. M. (2019). An Analysis of the Relationship Between Work-Life Balance and Productivity in Ghana. International Journal of Human Resource Studies.

Rahman, K. T., & Arif, Z. U. (2021). Working from Home during the COVID-19 Pandemic: Satisfaction, Challenges, and Productivity of Employees. International Journal of Trade and Commerce-IIARTC.

Ramos, J. P., & Prasetyo, Y. T. (2020). The Impact of Work-Home Arrangement on the Productivity of Employees during COVID-19 Pandemic in the Philippines: A Structural Equation Modelling Approach. International Conference on Industrial and Business Engineerin.

Soewito, D. S. M. (2020). Manajemen Sumber Daya Manusia dan Aplikasinya dalam Penelitian Ilmiah. RV Pustaka Horizon.

Sugiyono. (2010). Metode Penelitian Kuantitatif Kualitatif & RND. Alfabeta.

Sugiyono. (2016). Metode Penelitian Kuantitatif Kualitatif dan R&D. Alfabeta.

Copyright holder:

Panji Sukma Nugraha (2022)

First publication right:

Syntax Literate: Jurnal Ilmiah Indonesia

This article is licensed under: