1

REFFERED BY PROF. UMAR ALIYU

Job Satisfaction And Employee Motivation Mediates The Productivity Of
COVID-19 Inducted Work From Home
Andrews Tardie 1 & Lawal Aliyu Umar 2
Department Of Business Administration
SMC University Switzerland
Abstract- COVID-19 lockdown has tremendously effected and change the world in different
aspects and dimensions. Working from home became inevitable after the Ghanaian government
enforced a full lockdown in Accra, the capital, and Kumasi, the second largest city and this
resulted too many workers beginning to work from home for the first time. Furthermore,
reactions of employees varied in the midst of the pandemic as a result of many variables. The
thrust of the study is to examine the mediating role of job satisfaction and motivation on
productivity of working from home during the COVID-19 pandemic and its aftermath using
Structural Equation Modelling. The researcher sampled 355 respondents in order to examine the
mediating role. The findings however showed that job satisfaction and motivation mediate the
relationship between organizational factors, employee engagement, and technical support on one
hand, and productivity on the other hand. To reach higher levels of performance, the study
recommends that practitioners establish techniques to foster good work attitudes and boost
perceived organizational support.
I. INTRODUCTION
Many workers in Ghana were forced to work from home and significant numbers are still
working from home. Some corporate organizations are considering the viability of an extended
WfH. A typical example is USAID Ghana and West Africa; whose staff are all still working
from home as of March 2022. The role of job satisfaction and employee motivation in mediating
the relationship between organizational factors, employee engagement, and technical support on
one hand and COVID-19 induced Work from Home (WfH) productivity on the other hand has
not been explored to its fullest. Employee attitudes are incredibly important to management and
organizations because they determine how employees behave in the workplace. Employees who
are satisfied are regarded to be more productive than their unsatisfied colleagues (Bhardwaj,
Mishra, & Jain, 2021).
Although several studies have looked at the relationship between job satisfaction and employee
performance, empirical research on the factors that influence job satisfaction and its impact on
performance, particularly in COVID-19 induced WfH, is limited. Deficiencies in productivity
issues have a negative impact on profitability, resulting in billions of dollars in yearly reporting
losses across industries (De Winne, S., Marescaux, Sels, Van Beveren, & Vanormelingen, 2019).
Understanding what motivates staff to fulfill metric-based goals is critical to improving overall
performance. There is a positive relationship between job satisfaction, employee motivation, and
job performance (Rožman, Tominc, & Milfelner, 2020). The goals of this study, which was
based on the motivation and job satisfaction theory, is to model and analyze the relationships
between employee engagement, organizational factors, technical support, job satisfaction,
employee motivation, and productivity of WfH.
 2

II. LITERATURE REVIEW

The COVID-19 pandemic has stunned our thought and perception of work and it continues to
redefine work globally. The world economy has not fully recovered from the shocks of COVID-
19 in the midst of threatening new waves. According to the International Monetary Fund (IMF),
COVID-19 will make close to 90 percent of the world’s population economically worse off
(Chebly, Schiano, & Mehra, 2020). The shock of COVID-19 lockdown is the worst to hit the
globe since World War II (Dabla-Norris, Vitor, & Kalpana, 2020). The continent of Africa has
had its fair share of the lockdown shock. Africa has the lowest numbers of workers who can
work from home (Saltiel, 2020). The mediating role of job satisfaction and employee motivation
on WfH productivity has not received much research in Ghana. In particular, the mediating role
of job satisfaction and employee motivation on WfH productivity has not been well explored to
its fullest.
2.1. Synthesis / Critique Of Previous Research
Firms that are productive are growing faster than the rest. Baily and Montalbano (2016) defined
productivity as the efficiency of converting input to output. The most important determinant of
growth and living standard is productivity (Du & Temouri, 2015). Magnus (2018) described
productivity as the output added per employees. During the 1980s, the idea of working from
home started, which was necessitated by technology (Faulds & Raju, 2020). Both small and
large corporations have been making efforts since the mid-80s at making work from home
possible with the mainstream media dominant use of phrases such as “the growing
telecommuting movement (Streitfeld, 2020). Globally, 52 percent of all employees work from
home at least once a week and 56 percent of employers allow workers to work from home
(OWLLabs, 2018).
The mediating effect of motivation on productivity was evaluated by Al Banin et al (2020) in
their study of enhancing employee performance with work in Bumiayu Hospital in Indonesia.
The researchers concluded that motivation mediates the relationship between organizational
support and employee performance. It is, however, contrary to the woks of Saltson and Nsiah
(2015) in their study of the mediating relationship of motivation between perceived
organizational support and work performance of Ghanaian logistic staff. The researchers came
to a conclusion that the relationship between perceived organizational support and employee
work performance was not explained by motivation.
Berry and Morris (2018) in their quest to examine the hypothesized relationship between the
selected work-related employee engagement factors, and the outcome variable, turnover intent,
mediated by job satisfaction; examined then related literature for evidence. They discovered that
their study filled a gap in the literature because employee engagement was a relatively new term
that had not before been linked to both job satisfaction. Nas and Suriah (2020) conducted a
correlational study with 187 respondents to examine the effects of work engagement and job
satisfaction on nurse performance in Syekh Yusuf Regional Hospital of Gowa Regency. They
found that employee engagement improved nurse performance and job satisfaction improved
nurse performance as well. Li et al (2019) in a study involving 250 private sector textile workers
in Pakistan to find intervening mediators between high performance work systems and employee
performance, identified job satisfaction as a mediating factor. These findings thus, complement
the literature; and may not necessarily confirm the mediating role of job satisfaction on
productivity in WfH context.
 3

III. METHODOLOGY
3.1. Research Design
Data for the study was gathered from workers in Ghana's Greater Accra and Ashanti regions who
worked from home during the lockdown. For the purposes of this study, an employee is a person
with recognized rights and obligations who works part-time or full-time under an oral or written,
express or implicit contract of employment (Aliyu, 2019). Using a nonprobability snowball
sampling technique, 355 COVID-19 induced WfH participants were sampled using a semi-
structured questionnaire. Data was collected for one month, from January 22nd to February
22nd, 2021. The research model is shown in Figure 1.

Figure 1: Research Model (Source: IBM AMOS Ver. 26)

3.2 Research Questions And Hypotheses

The study's main research question was: Does job satisfaction and employee motivation mediate
the productivity of WfH? From the research questions, the following mediation hypothesis was
formulated to guide the researcher to address the key critical questions that relates to mediating
role of job employee motivation.
HO1A: Employee motivation does not mediate the relationship between organizational factors
and productivity of WfH.
HO1B: Employee motivation mediates the relationship between organizational factors and
productivity of WfH.
HO2A: Employee motivation does not mediate the relationship between employee engagement
and productivity of WfH.
HO2B: Employee motivation mediates the relationship between employee engagement and
productivity of WfH.
HO3A: Employee motivation does not mediate the relationship between technical support and
productivity of WfH.
HO3B: Employee motivation mediates the relationship between technical support and
productivity of WfH.
HO4A: Job satisfaction does not mediate the relationship between organizations factors and
productivity of WfH.
 4

HO4B: Job satisfaction mediates the relationship between organizations factors and productivity
of WfH.
HO5A: Job satisfaction does not mediate the relationship between employee engagement and
productivity of WfH.
HO5B: Job satisfaction mediates the relationship between employee engagement and
productivity of WfH.
IV. DATA ANALYSIS
Prior to analysis, a check for Structural Equation Model (SEM) assumption was undertaken to
ensure the data meet the requirements for SEM modelling. All missing data were distributed
randomly and was below the 5% threshold. This is a very low percentage of missing data, and it
can be considered acceptable (Leyrat, Carpenter, Bailly, & Williamson, 2021). There were no
terrible influential outliers at the multivariate level. A total of 56 variables was used in the SEM
measurement model. This yields a Mahalanobis distance critical value of 94.46 at a probability
of .001 (Hair et al, 2010 as cited in Tarhini, 2018). Thus, all Mahalanobis distance value of more
than 94.46 were excluded in the final analysis to achieve multivariate normality. All the observed
variables met the necessary condition for achieving multicollinearity except for three measured
variables (OFW3, OFW6, and TSW2). The data also did not violate the assumption of
homoscedasticity.
4.1 Exploratory Factor Analysis
Exploratory factor analysis was carried out to ascertain the number of unique constructs that
corresponds to the dataset. The value of the determinant was close to, but was not equal to zero
(2.261x10-22), establishing the fact that the dataset does not violate the assumption of positive
definiteness (Kline 2016, p. 67). The Kaiser-Meyer-Olkin (KMO) measure of sampling adequacy
value was .908; establishing that the sample size is adequate (Civelek, 2018, p. 34). Further, the
Bartlett's test of sphericity is significant (p = .00); meaning that at least two of the variables are
strongly correlated and that factor analysis is warranted (Little, 2013, p. 139). The pattern matrix
is shown in Table 1.
4.2 Confirmatory Factor Analysis
Confirmatory Factor Analysis (CFA) was used in to analyze the relationships between the
various structures within the conceptual model. The aim of CFA was to ensure that additional
samples of data that match the model confirm the hypothesized model's validity (Schumacker, &
Lomax, 2010, p. 164). The model fit and the validity of the measurement model was evaluated.
All the latent variables were linked together with the measured variables; represented by a
rectangular shape (see Figure 2). In all, a total of 56 measured variables were used in the CFA
which was derived from the EFA.
 5

Table 1
Pattern Matrix
Factor
1 2 3 4 5 6 7 8
CHW6 .914
CHW7 .884
CHW8 .856
CHW5 .847
CHW3 .809
CHW9 .784
CHW4 .739
CHW2 .737
CHW1 .522
PRW8 .876
PRW5 .837
PRW7 .837
PRW6 .836
PRW4 .834
PRW2 .831
PRW3 .801
PRW1 .646
OFW5 .907
OFW3 .904
OFW2 .885
OFW4 .880
OFW6 .876
OFW7 .848
OFW1 .776
MTW7 .926
MTW6 .913
MTW5 .902
MTW4 .847
MTW8 .809
MTW3 .704
MTW2 .624
MTW1 .570
TSW2 .870
TSW6 .861
TSW7 .846
TSW5 .828
TSW1 .754
TSW8 .699
TSW3 .682
TSW4 .630
JSW2 .924
JSW6 .879
JSW5 .873
JSW3 .867
JSW4 .772
JSW1 .701
JSW7 .687
EEW2 .909
EEW6 .897
EEW4 .886
EEW3 .843
EEW5 .819
EEW1 .757
FOW1 .893
FOW3 .871
FOW2 .664
Note.
Extraction Method: Maximum Likelihood.
Rotation Method: Promax with Kaiser Normalization.
 6

a: Rotation converged in 7 iterations.

Figure 2: Measurement Model / Confirmatory Factor Analysis

4.3 Goodness Of Fit Indices

The maximum-likelihood approach was used to estimate the model's parameters in this analysis,
with all tests being done on variance-covariance matrices. The levels of initial fit indices from
the survey data are provided (see Table 2). These results showed that there was still room for
improving the fitness of the model to the data as some parameters were still terrible.
 7

Table 2
Goodness of Fit Indices (Initial Measurement Model)
Measure Model Estimate Threshold Interpretation
CMIN (χ2) 3428.453 -- --
Degrees of freedom 1456 -- --
DF
CMIN/DF 2.355 Between 1 and 3 Excellent
GFI .750 >.95 Terrible
AGFI .726 >95 Terrible
CFI .877 >.95 Terrible
SRMR .058 <.08 Excellent
RMSEA .062 <.06 Acceptable
PClose -- >.05 Not Estimated
RMR .050 <.10 Excellent
PNFI .761 >.60 Excellent
TLI .870 >.095 Terrible
IFI .878 >.95 Terrible

The data did not fit well with the initial modelling. There was the need to use adjustment indices
add or remove paths in the model in order to arrive at the final best model (Schumacker &
Lomax, 2010 p. 73). All standardized regression wights less than .50 were first removed exept
for FOW2. FOW2 was retained for further observation since it is a measure of frequency of
WfH during the lockdown. The variables TSW8, OFW7, TSW6, EEW6, TSW4, OFW2, JSW5,
and CHW1 (see Questionaire), were removed from the model one at a time in the listed order to
improve the fit indices. Table 3 and Figure 3 show the model fit indices of the final measurement
model after the improvement in the model fit with an additional run of the model using the
Gaskin and Lim (2017) model fit measures AMOS plugin; and Hu and Bentler (1999) cut-off
criteria.
Table 3
Goodness of Fit Indices (Final Measurement Model)
Measure Model Estimate Threshold Interpretation
CMIN (χ 2) 1747.847 -- --
Degrees of freedom 1006 -- --
DF
CMIN/DF 1.737 Between 1 and 3 Excellent
CFI .942 >.95 Acceptable
SRMR .056 <.08 Excellent
RMSEA .046 <.06 Excellent
PClose .979 >.05 Excellent
PNFI .813 >.60 Excellent
TLI .937 >.95 Acceptable
IFI .942 >.95 Acceptable
 8

Figure 3: Final Good Fit Measurement Model

4.4 Model Validity Measure

The AVEs for the data under consideration were all greater than .50, and greater than.80 for CR,
as shown in Table 4. As a result, all factors have sufficient reliability and convergent validity.
Furthermore, the total AVE of the average value of variables used in the proposed model is
greater than their correlation value (AVE > MSV), indicating that there were no discriminant
validity issues.
From Table 4, the construct with the least CR is FOW, with a CR value of .784 (greater
than .7); thus reliability of the constructs is well established. The construct with the least AVE
from Table 4 is also that of FOW with an AVE value of .567 (greater than .5); all CRs are
greater than AVE for each construct; thus establisheshing constuct validiy. On discriminant
validity, all MSVs from Table 4 are less than AVE for each construct and the corresponding
ASVs are less than AVE; indicationg discriminant validity is well established.
 9

Table 4
Validity Analysis of Measured Variables (MV)
MaxR
MV CR AVE MSV CHW PRW MTV JSW OFW EEW TSW FOW
(H)
CHW .933 .636 .307 .938 .798

PRW .929 .622 .270 .937 .056 .788

MTV .918 .618 .270 .937 .137 .519 .786

* ***
JSW .904 .612 .195 .914 .270 .412 .388 .782
*** *** ***
OFW .932 .735 .149 .990 .133 .266 .381 .386** .857
* *** *** *
EEW .936 .746 .307 .944 .554 .238 .282 .442** .258 .864
*** *** *** * ***
TSW .879 .608 .212 .983 .185 .460 .353 .320** .233 .303 .779
** *** *** * *** ***
FOW .784 .567 .008 .883 .009 .019 -.032 .004 .091 -043 -.087 .753

Note.
Significance of Correlations:
† p < .100
* p < .050
** p < .010
*** p < .001
4.5 Heterotrait-Monotrait Ratio Of Correlations.
Discriminant validity is considered established if the HTMT is obviously less than one.
Henseler, Ringle, and Sarstedt (2015) new criteria for discriminant validity was adopted (see
Table 5).
Table 5
HTMT Analysis
CHW PRW MTW JSW OFW EEW TSW FOW
CHW
PRW .038
MTW .140 .562
JSW .285 .435 .429
OFW .142 .289 .425 .406
EEW .549 .256 .326 .461 .27
TSW .219 .477 .397 .394 .289 .372
FOW .036 .044 .007 .008 .066 .017 .053

From Table 5, the paired correlations of all constructs are shown. The largest paired correlation
was .562 (less than one) between MTW and PRW, further establishing discriminant validity.

4.6 The Structural Model

The goodness of fit structural model was tested. This is based on the same parameters used for
measuring the goodness-of-fit for the proposed model (see Figure 4 and Table 6).
 10

Figure 4: Structural Model

Table 6
Goodness of Fit Indices (Final Structural Model)
Measure Model Estimate Threshold Interpretation
CMIN (χ ) 2
7.699 -- --
Degrees of freedom DF 5 -- --
CMIN/DF 1.540 Between 1 and 3 Excellent
CFI .996 >.95 Excellent
RMSEA .039 <.06 Excellent
GFI .995 >.90 Excellent
AGFI .961 >.80 Excellent
RMR .006 <.10 Excellent
NFI .989 >.90 Excellent
PNFI .177 >.60 Excellent
TLI .977 >.95 Excellent
IFI .996 >.95 Excellent

4.7 Mediation Analysis

Since social phenomena are fundamentally complex and as such a direct relationship cannot
always provide a complete picture of an inter-relations; mediators are needed to add a layer of
uncertainty to aid a researcher in coming close to explaining a specific phenomenon in the social
sciences (Boateng, 2021). In performing the mediation analysis, the bootstrap approach, which
is fast becoming the method of choice in SEM mediation effects analysis (Coutts, Hayes, &
Jiang, 2019), was adopted.
First, the total and specific indirect effects from the model were obtained by running the model
in AMOS and recording the values from the text output. From the model, all single headed
arrows are considered direct effects (see Figure 5). Beyond direct effect, the focus of the research
was on indirect effects that determine mediation. As an example, the path from OFW through
MTV to PRW is an indirect effect; whilst the path from OFW to PRW is a direct effect.
 11

Figure 5: Specific Direct and Indirect Effects

Indirect effects are obtained by calculating path coefficients (Dempsey, O’Brien, Tiamiyu, &
Elhai, 2019). Essentially, the indirect effect of MTW on the path from OFW to PRW is
OFW_MTW x MTW_PRW. From the AMOS output tab, the direct, indirect and total effects
check boxes were checked. Bootstrapping was also used with a bootstrap sample of 2000 and a
bias-corrected interval confidence interval of 95%.

SPSS AMOS plugins and estimands (Gaskin & Lim, 2018) were used to generate the indirect
effects of employee motivation and Job satisfaction on WfH productivity. In bootstrapping,
2,000 bootstrapped samples were taken from the data using Gaskin and Lim's (2018) AMOS
plugin and estimands to estimate coefficients, and measure the indirect effects of job satisfaction,
and motivation on WfH productivity. This was used in AMOS to first test for direct effect of the
independent variables (OFW, TSW, and EEW) on WfH productivity (PRW). The results of the
generated output from the AMOS plugin and the estimands are as shown in Table 7. The null
hypothesis is that the indirect effect of the mediating variables is zero. Thus, zero must exist in
between the lower and the upper bound. From Table 7, it can be concluded that the effect of the
mediating variables (MTV and JSW) is significant since zero does not fall in the lower and the
upper bound of all the observed paths. The hypothesized mediation effects on the dataset are as
summarized in Table 8. Job satisfaction and employee motivation are mediators in the SEM
model.
 12

Table 7
Indirect Effects
Indirect Path Unstandardize Lower Upper P-Value Standardized
OFW --> MTV --> PRW d Estimate
.067 .045 .097 .001 Estimate
.111***
EEW --> MTV --> JSW .024 .009 .046 .005 .032**
TSW --> MTV --> PRW .089 .057 .129 .001 .091***
OFW --> JSW --> PRW .031 .015 .053 .001 .051***
EEW --> JSW --> PRW .056 .034 .083 .001 .078***
TSW-->MTV--> JSW .055 .029 .096 .001 .054***
TSW-->MTV--> JSW --> PRW .012 .005 .024 .000 .054***
EEW-> MTV->JSW-> PRW .005 .002 .012 .004 .032**
EEW --> MTW --> PRW .039 .017 .066 .007 .055**
OFW --> MTV --> JSW .042 .025 .066 .001 .066***
OFW --> MTV --> JSW --> PRW .009 .004 .017 .000 .066***
MTV --> JSW --> PRW .038 .018 .068 .001 .050***
Note.
Significance of Estimates:
*** p < .001
** p < .010

Table 8.
Hypotheses Testing (Mediation Effects)
Relationship Alternative Standardized Estimate P-Value Study Results
Hypotheses
OFW --> MTV --> PRW 1 .001 .111*** Supported
EEW --> MTV --> PRW 2 .007 .055** Supported
TSW --> MTV --> PRW 3 .001 .091*** Supported
OFW --> JSW --> PRW 4 .001 .051*** Supported
EEW --> JSW --> PRW 5 .001 .078*** Supported
Note.
Significance of Estimates:
*** p < .001
** p < .010

4.8 Mediation Effects Of Employee Motivation

The null hypotheses 1 states that employee motivation does not mediate the relationship between
organizational factors and productivity of WfH. Based on Baron and Kenny criteria, there must
be a significant relationship between motivation and productivity in order for a mediation to be
validated. This criterion was met as the relationship between motivation and productivity of
WfH is significant ( = .371 (***)). Recent research, however, has revealed that observing a
direct relationship is not required to create a mediational path (Hayes, 2009). Both the Baron and
Kenny criteria and the estimands with plugins all established mediation of employee motivation
with WfH productivity. The relationship between employee motivation and productivity of WfH
is significant ( = .371***). Alternative hypotheses 1 state that employee motivation mediates
the relationship between organizational factors and productivity of WfH. Table 8 shows that
mediation effect has a significant p-value (p < .001). Thus, we reject the null hypotheses 1.
The finding is in sharp contrast to the work of Saltson and Nsiah (2015), who concluded that
there is neither moderating nor mediating effect on motivation on the relationship between
organizational factors and employee performance in their study of 130 logistic company workers
in Ghana. Their study was, however, not related to WfH staff but the theory can be extended to
contrast the mediating effect of motivation on WfH scenarios.
 13

In their study of enhancing employee performance with work using 120 respondents from
Bumiayu Hospital in Indonesia, Al Banin et al (2020) came to a conclusion that motivation
mediates the relationship between organizational support and employee performance
(productivity). This shows that motivation must be considered if organizational factors are to
have the desired productivity increase effect on WfH.
The null hypotheses 2 states that employee motivation does not mediate the relationship between
employee engagement and productivity of WfH. From Table 8, the mediation effect has a
significant p-value (p < .010); thus, establishing a mediation between employee engagement and
productivity of WfH by both the Baron and Kenny criteria and the estimands with plugins. We
thus, reject the null hypotheses 2.
Null hypotheses 3 states that employee motivation does not mediate the relationship between
technical support and productivity of WfH. From Table 8, the mediation effect has a significant
p-value (p < .001); thus, establishing a mediation between technical support and productivity of
WfH by both the Baron and Kenny criteria and the estimands with plugins. We therefore reject
the null hypotheses 3. This study of WfH workers in Ghana confirms the important role of
motivation as a mediator in the relationship between organizational characteristics and employee
productivity. It is not an exaggeration to argue that motivation is essential for success
(Geelmaale, 2019).
I. Mediation effects of job satisfaction.
Both the Baron and Kenny criteria and the estimands with plugins all established mediation of
job satisfaction with WfH productivity. The relationship between job satisfaction and
productivity of WfH is significant ( =.226***). Null hypotheses 4 states that job satisfaction
does not mediate the relationship between organizations factors and productivity of WfH. From
Table 8, the mediation effect has a significant p-value (p < .001); thus, establishing a mediation
between technical support and productivity of WfH by both the Baron and Kenny criteria and the
estimands with plugins. We therefore reject the null hypotheses 4.
Sharma and Biswakarma (2020) in their study of 158 hotel employees in Nepal found that the
influence of perceived organizational support on perceived work performance was mediated by
job satisfaction. Li, Naz. Khan, Kusi, and Murad (2019); identified job satisfaction as a
mediating factor in high performance work system on employee performance. Their study
findings, consistent with this study, provided clear proof that perceived organizational support
and perceived work performance can be enhanced when employees are satisfied with their jobs.
Job satisfaction, according to the null hypotheses 5, does not influence the relationship between
employee engagement and WfH productivity. The mediation effect has a significant p-value (p
< 001) in Table 8, indicating that both the Baron and Kenny criteria and the estimands with
plugins established a mediation between employee engagement and productivity of WfH. We
thus reject the null hypotheses 5. Null hypotheses 4 and 5 were rejected with estimands (p
= .051*** and p=.078***) respectively.
In a correlational study of 187 respondents to analyze the effects of work engagement and job
satisfaction on performance of nurses in Syekh Yusuf Regional Hospital of Gowa Regency in
Indonesia; Nas and Suriah (2020) demonstrated that job satisfaction played a key role to elevate
the performance of the nurses. Based on this study, employers need to pay more attention to job
satisfaction to increase the productivity of WfH staff.
 14

4.9 Summary Of Results

This study started by undertaking a factor analysis and the development of the measurement
model, coupled with the reliability, discriminant validity, and convergent validity of all the
constructs within the proposed research model. The results of the confirmatory factor analysis
revealed that eight items (TSW8, OFW7, TSW6, EEW6, TSW4, OFW2, JSW5, and CHW1); see
questionnaire for code description; have to be deleted from the initial measurement model to
achieve a good fit. The deletion criteria was based on indicators with high covariance and high
regression weight. Following the validation and reliability of the constructs, the structural model
was evaluated in order to assess the hypothesized relationships.
The refined model has a reasonably high explanatory capacity for the study, according to
the results of the squared multiple correlations (R2), which provide details about the degree to
which the model explains variance in the data set. The determinants OFW, TSW, EEW, MTW,
CHW, FOW, and JSW, in particular, accounted for 43% of the variance in WfH productivity.
V. CONCLUSION
Job satisfaction and employee motivation was firmly established as a mediating factor by both
the Baron and Kenny criteria and the estimands with plugins (see Table 8), in the relationship
between organizational factors, technical support, and employee engagement on one hand; and
WfH productivity on the other hand. The consistency of the current study of WfH staff in Ghana
further validates the key mediation role that motivation plays in the relationship between
organizational factors and employee productivity. A motivated employee is more likely to
willingly put in more effort to complete a task, and a good outcome is more likely. When
objectives are met, employees have a sense of fulfillment and happiness, which fosters a positive
attitude in the workplace.
On job satisfaction, the mediation effect of job satisfaction on the relationships between
organizational factors and WfH productivity has a significant p-value (p < .001). Further, the
mediation effect of job satisfaction on the relationships between employee engagement and WfH
productivity has a significant p-value (p < .001). Thus, job satisfaction was evaluated in this
study as a mediating factor by both the Baron and Kenny criteria and the estimands with plugins
(see Table 8), in the relationship between organizational factors and employee engagement on
one hand; and WfH productivity on the other hand.
References
Al Banin, Q., Eliyana, A., & Latifiyah, E. R. (2020). Enhancing employee performance with work motivation as a
mediation variable. Systematic Reviews in Pharmacy, 11(9), 333-346.
Aliyu, A. U. L. (2019). Effect of employee participation in decision making in an organization performance.
International Journal of Economics & Business, 3(2), 255–259.
Baily, M. N., & Montalbano, N. (2016) Why is US productivity growth so slow? Possible explanations and policy
responses. The Brookings Institution. Retrieved from http://www.nomurafoundation.or.jp/wordpress/wp-
content/uploads/2016/12/20161116_M_Baily-N_Montalbano.pdf
Berry, M. L., & Morris, M. L. (2018). The impact of employee engagement factors and job satisfaction on
withdrawal intent. Erick Collection Thesaurus.
Bhardwaj, A., Mishra, S., & Jain, T. K. (2021). An analysis to understanding the job satisfaction
of employees in banking industry. Materials Today, 37(2), 170-174,
https://doi.org/10.1016/J.MATPR.2020.04.783
Boateng, S.L. (2021). Structural Equation Modelling with SPSS and AMOS Session 4 - Practical Session. Retrived
from https://www.youtube.com/watch?v=16UJJNSdLUo
 15

Chebly, J., Schiano, A., & Mehra, D. (2020). The value of work: Rethinking labour productivity in times of COVID-
19 and automation. American Journal of Economics & Sociology, 79(4), 1345–1365.
https://doi.org/10.1111/ajes.12357
Civelek, M. E. (2018). Essentials of structural equation modelling. Zea Books. https://doi.org/10.13014/k2sj1hr5
Coutts, J. J., Hayes, A. F., & Jiang, T. (2019). Easy statistical mediation analysis with distinguishable dyadic data.
Journal of Communication, 69(6), 612–649. https://doi.org/10.1093/joc/jqz034
Dabla-Norris, E., Vitor G., & Kalpana, K. (2020). Preparing for an unknown world. Washington, DC: International
Monetary Fund. https://www.imf.org/exter nal/pubs/ft/fandd/ 2020/06/the-internationalorder-post-covid 19-
dabla.htm
Dempsey, A. E., O’Brien, K. D., Tiamiyu, M. F., & Elhai, J. D. (2019). Fear of missing out (FoMO) and rumination
mediate relations between social anxiety and problematic Facebook use. Addictive Behaviors Reports,
9(October 2018), 100150. https://doi.org/10.1016/j.abrep.2018.100150
De Winne, S., Marescaux, E., Sels, L., Van Beveren, I., & Vanormelingen, S. (2019). The impact
of employee turnover and turnover volatility on labor productivity: a flexible non-linear
approach. The International Journal of Human Resource Management, 30(21), 3049-
3079.
Du, J., & , Y. (2015). High-growth firms and productivity: Evidence from the United Kingdom. Small Business
Economics, 44(1), 123–143.
Faulds, D. J., & Raju, P. S. (2020). The work-from-home trend: An interview with Brian Kropp. Business horizons,
10(16). Advance online publication. https://doi.org/10.1016/j.bushor.2020.10.005
Gaskin, J., & Lim, J. (2018), "CFA Tool", AMOS Plugin. Gaskination's StatWiki.
Geelmaale, A. M. A. (2019). Impact of employee motivation on organizational performance. International Journal
of Advanced Research, 7(10), 166–172. https://doi.org/10.21474/ijar01/9818

Hayes, A.F. (2009). Beyond Baron and Kenny: Statistical mediation analysis in the new millennium.
Communication Monographs, 76(4), 408-420, DOI: 10.1080/03637750903310360
Henseler, J., Ringle, C. M., & Sarstedt, M. (2015). A new criterion for assessing discriminant validity in variance-
based structural equation modelling. Journal of the Academy of Marketing Science, 43 (1), 115-135.
Kline, R. B. (2016). Principles and practice of structural equation modelling (4th ed.). New
York: Guilford Press. ISBN: 9781609182304
Leyrat, C., Carpenter, J. R., Bailly, S., & Williamson, E. J. (2021). Common methods for handling missing data in
marginal structural models: What work and why. American Journal of Epidemiology, 190(4), 663–672.
https://doi.org/10.1093/aje/kwaa225
Li, C., Naz, S., Khan, M. A. S., Kusi, B., & Murad, M. (2019). An empirical investigation on the
relationship between a high-performance work system and employee performance:
Measuring a mediation model through partial least squares–structural equation modelling.
Psychology Research and Behavior Management, 12, 397–416.
https://doi.org/10.2147/PRBM.S195533
Little, T. D. (2013). The Oxford Handbook of Quantitative Methods. Vol 2. Statistical Analysis.

Magnus, C. B. (2018). Employment protection and labour productivity. Journal of Public Economics, 157, p. 138-
157.
Nas, A. R., & Suriah, I. (2020). The effects of work engagement and job satisfaction on
performance of nurses in Syekh Yusuf regional hospital of Gowa. European Journal of
Molecular & Clinical Medicine, 07(08), 4407–4418.
OWLLabs (2018). Global state of remote work. Retrieved from
https://www.owllabs.com/state-of-remote-work/2018
Rožman, M., Tominc, P., & Milfelner, B. (2020). A comparative study using two SEM
techniques on different samples sizes for determining factors of older employee’s
motivation and satisfaction. Sustainability (Switzerland), 12(6).
https://doi.org/10.3390/su12062189
Saltson, E., & Nsiah, S. (2015). The mediating and moderating effects of motivation in the relationship between
perceived organizational support and employee job performance. International Journal of Economics,
Commerce and Management, III(7), 654–667.
Saltiel, F. (2020). Who can work from home in developing countries? COVID Economics, 7,
104 –118. http://econweb.umd.edu/~saltiel/files/wfh_mostrecent.pdf
 16

Schumacker, R. E., & Lomax, R. G. (2010). A beginner’s guide to structural equation modelling (3rd éd.). New
York London: Routledge, Taylor & Francis Group. https://doi.org/10.1111/j.1467-985x.2012.01045_12.x
Sharma, D. R., & Biswakarma, G. (2020). Mediating role of job satisfaction in the relationship
between organizational support and job performance. Jornal of European Business &
Management, 6(4), 84–94. https://doi.org/10.11648/j.ebm.20200604.13
Streitfeld, D. (2020). The long, unhappy history of working from home. The New
York Times. Retrieved https://www.nytimes.com/2020/06/29/technology/working-from-home-failure.html
Tarhini, A. (2018). The effects of individual-level culture and demographic characteristics on e-learning acceptance
in Lebanon and England: A structural equation modelling approach. (Unpublished doctoral dissertation).
Brunel University, Brunnel.