You are on page 1of 11

Effect of Workstation and Virtual

Engagement During Covid-19 on Work


Life Balance

Submitted to: Ma’am Naziya Saeed.


Course: Methods Of Business Research.
Institute of Business Management, Karachi Pakistan.
Submitted From:
Syed Muhammad Ali. (20181-24540),
Affan Sami.
Uzair Sakrani
|4| Data Analysis and Interpretation.

A total of 202 responses were obtained out of which 54.95% were males and 45.05% were
females. The demographic profile was also analyzed on marital status, age, level of education
and number of employees. The demographic results of the respondents is shown in Table 4.1.

Table 4.1 Descriptive Statistics


Frequenc
Percentage
y
Gender
Male 111 54.95%
Female 91 45.05%
Marital Status
Single 149 73.76%
Married 53 26.24%
Age
20-29 111 54.95%
30-39 67 33.17%
40-49 19 9.41%
50 or above 5 2.48%
Level of Education
Intermediate 9 4.48%
DAE 1 0.50%
Bachelors 83 41.29%
Masters 104 51.74%
pHD 4 1.99%
No. of Employees
10-35 34 16.83%
36-99 57 28.22%
100 or above 111 54.95%
Note: The above table indicates the response of 202 respondents from Pakistan.

For judging the normality of the data of each variable we have analyzed the excess kurtosis
and skewness of the data which should be within the limit of +3 and -3. If the value lies
within this range we can verify the normality of our data otherwise it will be a reflected as a
katotic or skewed data. Excess kurtosis and skewness results are mentioned in Table 4.2.
Table 4.2 Univariate Normality
Excess
Skewness
Kurtosis
G -1.98 0.2
MS -0.824 1.088
A 0.868 1.177
Ed 2.823 -1.343
Emp -0.859 -0.765
WS1 1.522 -1.225
WS2 -1.048 -0.145
WS3 -0.596 -0.524
WS4 -0.311 -0.684
VE1 -0.481 -0.582
VE2 -0.328 -0.597
VE3 -0.943 -0.244
VE4 -0.421 -0.69
PW1 -0.679 -0.436
PW2 -0.857 -0.327
PW3 -0.891 -0.169
PW4 -0.27 -0.659
PR1 -0.842 -0.252
PR2 -0.875 -0.296
PR3 -0.407 -0.559
PR4 -0.655 -0.366
WLB1 -0.717 -0.432
WLB2 -0.617 -0.501
WLB3 -0.992 -0.308
WLB4 0.557 -0.998
MV1 -0.964 -0.249
MV2 -0.685 -0.514
MV3 -0.026 -0.84
MV4 0.609 -0.869
MV5 -1.652 0.604
MV6 -1.083 0.401
MV7 -0.735 0.815
MV8 -1.011 0.415

Note: The above table indicates the values of kurtosis and skewness, all of which lies within
range of +3 and -3.
Since a variable has relation with other variables, we can also validate their normality of one
variable with other variables through Mardia’s test. With p-values less then zero which
reflects the normality of data in multivariate dimension. The results of Mardia’s test is
mentioned in Table 3.

Table 4.3 Multivariate Normality (Mardia’s Test)


p-
b z
value
Skewnes 270.429 9104.46
0
s 7 7
1270.78 17.1193
Kurtosis 0
4 7
Note: The above table indicates values of kurtosis and skewness.
Proposed model was shows the relationship between dependent and independent variables.
The dependent variable is work life balance (WLB) while the independent variables are
workstation (WS) and virtual engagement (VE) during Covid-19. Psychological well-being
(PW) serves to be the mediating variable while peer relationship (PR) serves to be the
moderator. Figure 4.1 shows the conceptual model drawn in Smart PLS.

FIGURE4.1 shows the Conceptual model shows the relationships between dependent and
independent variable
Figure 4.2 and Table 4.4 shows the values of outer loading. The values of outer loading
should be greater than 0.708. All values of outer loading are greater than 0.708 except WS1
and VE4 but we will retain them as they are greater than 0.5. The values that are greater than
0.708 shows the accuracy of research questions. The outer loading values reflects the
representation of item from its constructs.
FIGURE 4.2 conceptual model with outer loading values

Table 4.4 Outer Loadings of reflective constructs


PR PW VE WL WS
B
0.82
PR1
7
0.80
PR2
5
0.86
PR3
9
0.82
PR4
6
0.74
PW1
6
0.85
PW2
2
0.84
PW3
3
0.83
PW4
5
0.80
VE1
7
VE2 0.84
0.76
VE3
3
0.70
VE4
2
WLB 0.81
1 3
WLB 0.88
2 2
WLB 0.82
3 6
WLB 0.82
4 1
0.60
WS1
2
0.76
WS2
5
0.81
WS3
2
0.75
WS4
6
Note: The above table indicates the value of outer loadings, and all the values which are
greater than 0.5
The technique used for measurement model includes PLS-SEM 3, the model was tested
through algorithms value for convergent validity and also Heterotrait and Monotrait ratio.
This was assessed using Composite Reliability (CR) which reflects the consistency in data if
collected from same group after specific interval of time, and Average Variance Extracted
(AVE) which shows whether all questions measure a construct or not. Table 4.5 shows that
all Composite reliability values which are greater 0.7 which reflects the consistency of our
data, and average variance values greater than 0.5 which reflects that all questions measure
the associated construct.

Table 4.5 Composite Reliability & Convergent Validity


Composite
AVE
Reliability
PR 0.9 0.692
PW 0.891 0.673
VE 0.861 0.608
WL
0.903 0.699
B
WS 0.826 0.545
Note: The above table indicates the value of composite reliability and average variance
extracted

Figure 4.3a Composite reliability chart


Figure 4.3b Average variance extracted chart
Now we move towards the analysis of Discriminant validity (Heterotrait-Monotrait ratio of
correlations), also called HTMT ratio. Discriminant validity talks about the discrimination
that exist between the constructs. Below is the matrix which shows five constructs. Since the
value of HTMT is below = 0.85 hence it is assured that discriminant validation is established.
The HTMT ratio analysis is given in Table 4.6.

Table 4.6 Discriminant Validity


PR PW VE WLB WS
PR
PW 0.771
VE 0.698 0.691
WLB 0.751 0.706 0.642
WS 0.686 0.699 0.747 0.642
Note: The above table indicates discriminant validity of the constructs
Figure 4.4 Hetro Trade Mono Trade (HTMT) ratio
Moving towards the collinearity which refers to the overlapping between the constructs, the
threshold for collinearity is 3.3, and values above this threshold will indicate overlapping
between the constructs. Table 4.7 provides the results of collinearity, all the values are below
3.3 which indicate that there is no overlapping between the constructs, all the constructs are
independent and collinearity issue does not exist.

Table 4.7 Collinearity Via VIF Values


P V W W
PW
R E LB S
2.0
PR
32
2.0
PW
59
1.4 1.8
VE
69 03
W
LB
1.4 1.7
WS
69 24
Note: The above table indicates the collinearity between the constructs
Moving forward with the structure model analysis, this table shows whether the hypothesis
between constructs is supported or not supported. In order to support the hypothesis, it is
necessary that the p-value must be less than 0.05 and according to the Table 4.8 all the values
are less than 0.05 except one (WS  WLB), indicating that all the hypothesis are supported
except one.
Table 4.8 Structure Model
Beta Standard T P
Decision
Coefficient Error Statistics Values
PR -> WLB 0.308 0.082 3.77 0 Supported
PW -> WLB 0.238 0.074 3.228 0.001 Supported
PW*PR -> WLB -0.139 0.054 2.553 0.005 Supported
VE -> PW 0.375 0.069 5.407 0 Supported
VE -> WLB 0.124 0.07 1.761 0.039 Supported
WS -> PW 0.341 0.07 4.901 0 Supported
Not
WS -> WLB 0.085 0.064 1.332 0.091
Supported
VE -> PW ->
0.089 0.031 2.844 0.002 Supported
WLB
WS -> PW ->
0.081 0.032 2.577 0.005 Supported
WLB

Data analysis would not be completed until advanced analysis is conducted. IPMA also called
Importance Performance Map Analysis reflects on the importance and performance of
constructs. It is found that the Peer relationship has the most importance and maximum
performance among all other variables. However psychological well-being has high
importance but significantly lower performance, hence we need to find ways in order to
increase the performance of it. Spreading awareness of mental health, ways to accommodate
stressful conditions and setting routine with break intervals will surely play their part in
increasing the performance of phycological wellbeing.

Figure 4.5 Importance Performance Map Analysis

The importance of R2 and adjusted R2 plays an important part in statistical analysis. If reflects
how much the independent variable explains the variance of dependent variable. It’s better if
the R2 value is closer to 1. Although 0.75 is substantial, 0.50 is moderate, 0.25 is weak and
according to the following results as given in Table 4.9 the level of R square is moderate.
Table 4.9 Coefficient of Determination (R2)
R R Square
Square Adjusted
PW 0.401 0.395
WLB 0.499 0.489
Note: The above table indicates the values of R and adjusted R2
Effect Size, f2 verifies the effect size that a construct has on another construct. The effect size
is measured accordingly, 0.02 (small), 0.15 (medium) and 0.35 (large effect). The Table 4.10
shows the results which reflects medium effect size of virtual engagement (VE) on
phycological wellbeing (PW). Followed by small effect size of virtual engagement (VE) and
workstation (WS) on work life balance (WLB).

Table 4.10 Effect Size (f2)


P V W W
PW
R E LB S
0.1
PR
21
0.0
PW
56
0.1 0.0
VE
59 18
W
LB
0.1 0.0
WS
32 16
Note: The above table indicates the values of f2
Moving towards the predictive relevance of variables which is showcased by Q 2. It basically
explains about out of sample prediction power, whether it is generalizable or not. Higher the
value from 0, more generalize the research would be. The results in Table 4.11 shows that all
values of Q square are greater than 0 therefore, the predictive relevance of all variables are
high.
Table 4.11 Predictive Relevance (Q2)
SSO SSE Q² (=1-SSE/SSO)
PR 808 808
PW 808 598.339 0.259
VE 808 808
WLB 808 536.412 0.336
WS 808 808
Note: The above table indicates the values of Q2 while retaining all the constructs
The relative impact of predictive relevance can be compared by means of measure to the
effect size (q2), which is defined as follows:
q2 = (Q2included – Q2excluded) / (1- Q2included)
In this table 4.12 we check the effect on work life balance by removing phycological
wellbeing. After plugging in the values in the above formula we get the final value of 0.029
which comprehends small effect size on the predictive relevance.
0.336 - 0.317 / 1-0.336= 0.029 (Small)

Table 4.12 Effect Size on Predictive Relevance (q2)


SSO SSE Q² (=1-SSE/SSO)
PR 808 808
VE 808 808
WLB 808 552.17 0.317
WS 808 808
Note: The above table indicates the values of Q2 after removing PW construct from the model

You might also like