Professional Documents
Culture Documents
CHAPTER 6
ANALYZING THE MEDIATING VARIABLE
IN A MODEL
Sometimes the research questions intend to address the effect of a mediating variable in the
relationship between an independent variable and its corresponding dependent variable in a
model. Diagram below illustrates the position of a mediator in the relationship between
independent variable and its corresponding dependent variable.
Direct Effect
from IV to DV
Indirect Effect
Through Mediator
First of all, the direct effect of independent variable on dependent variable is significant. When
the mediator variable M enters the model, the direct effect would be reduced since some of the
effect has shifted through the mediator. If it is reduced but still significant, the mediation effect
here is called “partial mediation”. However, if the direct effect is reduced and no longer
significant, then the mediation is called “complete mediation”.
101
SEM Made Simple
Zainudin Awang (2015). MPWS Publisher
When analyzing the mediator, there are two effects involved namely direct effect and indirect
effect. The direct effect is the effect from independent variable directly to dependent variable,
while the indirect effect is the effect from independent variable to dependent variable that goes
indirectly through the mediating variable.
Variable: M
Indirect Effect
Direct Effect
Independent Dependent
Variable: X1 Variable: Y
In this model, the researchers will examine the direct effect and indirect effect of X1 on Y. If
the direct effect of X1 on Y is reduced, and the indirect effect (through M) is significant, then
M is said to play a mediating role in linking X1 to Y indirectly.
102
SEM Made Simple
Zainudin Awang (2015). MPWS Publisher
Firstly, we begin by modeling the simple effect of X1 on Y as shown in Figure 2. We test the
direct effect of X1 on Y as shown in Figure 3. The output in Table 1 shows B1 is 0.36 and it has
a significant effect on Y (p-value < 0.001).
Secondly, we enter the mediator variable M into the model as shown in Figure 4. Now we test
the direct effect of X1 on Y as shown in Figure 5. The output in Table 2 shows the coefficient
value for B1 is reduced from 0.360 to 0.103 when M enters the model, and the direct effect of
B1 on Y is no longer significant (p-value = 0.062). Here, the requirement for complete
mediation is met. Finally, we need to test the hypothesis for B2 and B3. If these two hypotheses
are significant, then the type of mediation is complete mediation.
e1
1
X1 Y
e1
1
4.90, 1.36 3.87
.36
X1 Y
103
SEM Made Simple
Zainudin Awang (2015). MPWS Publisher
As has been said earlier, the test for mediator is only meaningful only if the direct effect is
statistically significant. The mediator variable is included into the model as shown in Figure 4.
Bear in mind to draw in triangular form so that the direct effect and indirect effect could easily
be identified. Once the model is executed, the result for path coefficient (standardized
coefficient) is shown in Figure 5.
Mediator M 1
enters the model M e2
e1
1
X1 Y
104
SEM Made Simple
Zainudin Awang (2015). MPWS Publisher
3.10
1
M e2
.53 .31
e1
1
4.90, 1.36 3.01
.10
X1 Y
105
SEM Made Simple
Zainudin Awang (2015). MPWS Publisher
The type of mediation here is a called a “complete mediation” since the direct effect of X1 on
Y is no longer significant after M entered the model (Hypothesis 1). Instead, the indirect
effect is significant. Thus, X1 has an indirect effect on Y through the mediator variable M
106
SEM Made Simple
Zainudin Awang (2015). MPWS Publisher
e14
1
Y1 e10
1 1
1
Y2 e11
Y
X1 1
Y3 e12
1 1
Y4 e13
X15 X14 X13 X11
1 1 1 1
e5 e4 e3 e1
107
SEM Made Simple
Zainudin Awang (2015). MPWS Publisher
.78
e14 .11
1
Y1 e10
1 1.00
.06
.47 1.03
1
Y2 e11
.87 Y .17
1.10
X1 1
Y3 e12
1.15
1.61 1.00 .87
1.42 1.53
1
Y4 e13
Fitness Indexes
X15 X14 X13 X11 1.ChiSq = 55.676
2.df = 19
1 1 1 1 3.ChiSq/df = 2.930
.19 .28 .23 .88 4.GFI = .956
5.AGFI = .917
e5 e4 e3 e1 6.CFI = .986
7.RMSEA = .078
Figure 7: The result shows the direct effect of X1 on Y (Beta Coefficient 0.87)
The analysis for mediation begins by showing that the direct effect of X1 on Y is significant.
The direct effect is measured through beta coefficient. In this case, B1 is significant.
When the mediating variable M enters the model, the value of path coefficient for X1 is
expected to reduce, or in other words the direct effect of X1 on Y would be reduced when the
mediator enters the model.
108
SEM Made Simple
Zainudin Awang (2015). MPWS Publisher
e6 e7 e8 e9
1 1 1 1
1
e15 M
e14
1
Y1 e10
1 1
1
Y2 e11
Y
X1 1
Y3 e12
1
1
Y4 e13
X15 X14 X13 X11
1 1 1 1
e5 e4 e3 e1
In the above diagram, X1, M, and Y are latent constructs. The study is interested to prove that
the construct M is mediating the relationship between construct X1 and construct Y.
The Regression Weight estimates for the model are presented in Figure 9 while the text output is
presented in Table 4.
109
SEM Made Simple
Zainudin Awang (2015). MPWS Publisher
.45
.19 .09 1.39
e6 e7 e8 e9
1 1 1 1
1.00 .81
1.00
.64
1 .75
.77 e15 M
e14 .11
1
.94 .79 Y1 e10
1 1.00
.06
.48 1
1.03 Y2 e11
.68 Y .17
1.10
X1 1
Y3 e12
1.59 1.15
1.00 .87
1.40 1.51
1
Y4 e13
Fitness Indexes
X15 X14 X13 X11 1.ChiSq = 151.309
2.df = 50
1 1 1 1 3.ChiSq/df = 3.026
.19 .28 .23 .87 4.GFI = .929
5.AGFI = .919
e5 e4 e3 e1 6.CFI = .972
7.RMSEA = .079
***Observed that the value of direct effect linking X1 to Y is reduced from 0.87 to 0.68.
110
SEM Made Simple
Zainudin Awang (2015). MPWS Publisher
The type of mediation here is a called “partial mediation” since the direct effect of X1 on Y is
still significant after the mediator variable M entered the model even though the path
coefficient for X1 is reduced from 0.87 (in Figure 7) to 0.68 (in Figure 9). In this case, X1 is
both significant direct effect on Y and also significant indirect effect on Y through the mediator
variable namely M.
111
SEM Made Simple
Zainudin Awang (2015). MPWS Publisher
The direct effect is the effect that goes directly from exogenous construct to endogenous
construct, while the indirect effect is the effect from exogenous construct to endogenous
construct that goes indirectly through the mediator in the model.
Figure 10 illustrate the position of mediator construct namely customer satisfaction in the
model. In this example, the researcher is interested to assess the role of Customer Satisfaction
as a mediator in the following relationships:
112
SEM Made Simple
Zainudin Awang (2015). MPWS Publisher
Figure 10: The Standardized Regression weights for every path in the model
First of all, obtain the standardized regression weights and the probability values which
indicate the significance for the respective path (Figure 10). The required information is given
in Table 9. Draw the triangle as shown in Figure 11.
113
SEM Made Simple
Zainudin Awang (2015). MPWS Publisher
Table 9: The standardized regression weights and its significance for each path
Standardized
Construct Path Construct P-Value Result
Estimate
Customer_Satisfaction <--- Service_Quality 0.51 0.001 Significant
Customer_ Loyalty <--- Customer_Satisfaction 0.58 0.001 Significant
Customer_ Loyalty <--- Service_Quality 0.22 0.001 Significant
The main hypothesis statement for testing a mediator: Ha: Customer Satisfaction mediates
the relationship between Service Quality and Customer loyalty
114
SEM Made Simple
Zainudin Awang (2015). MPWS Publisher
Ha1: Service Quality has significant effect on customer satisfaction 0.51 0.001 Supported
Ha2: Customer Satisfaction has significant effect on Customer Loyalty 0.58 0.001 Supported
Ha3: Service Quality has significant effect on Customer Loyalty 0.22 0.001 Supported
First of all, obtain the standardized regression weights and the probability values which
indicate the significance for the respective path (Figure 10). The required information is given
in Table 10. Draw the triangle as shown in Figure 12.
Table 11: The standardized regression weights and its significance for each path
Standardized
Construct Path Construct P-Value Result
Estimate
Customer_Satisfaction <--- Corporate_Image 0.27 0.001 Significant
Customer_ Loyalty <--- Customer_Satisfaction 0.58 0.001 Significant
Customer_ Loyalty <--- Corporate_Image 0.08 0.106 Not Significant
115
SEM Made Simple
Zainudin Awang (2015). MPWS Publisher
The main hypothesis statement for testing a mediator: Hb: Customer Satisfaction mediates
the relationship between Corporate Image and Customer loyalty
Hb1: Corporate Image has significant effect on Customer Satisfaction 0.202 0.001 Supported
Hb2: Customer Satisfaction has significant effect on Customer Loyalty 0.609 0.001 Supported
Hb3: Corporate Image has significant effect on Customer Loyalty 0.065 0.106 Not Supported
The results of hypothesis testing in Table 12 indicate that Customer Satisfaction does mediate
the relationship between Corporate Image and Customer Loyalty. Thus the type of mediation
here is full mediation since the direct effect is no longer significant after the mediator enters
the model.
116
SEM Made Simple
Zainudin Awang (2015). MPWS Publisher
117
SEM Made Simple
Zainudin Awang (2015). MPWS Publisher
Figure 13: The Standardized Regression weights for every path in the model
118
SEM Made Simple
Zainudin Awang (2015). MPWS Publisher
Table 14: The Summary – The Significance of direct and indirect effects
Based on the results in Table 14, the researcher can conclude that the results of bootstrapping is
consistent with the results of mediation test in Table 10
119
SEM Made Simple
Zainudin Awang (2015). MPWS Publisher
***The significance indirect effects indicate the moderation exists between CI and CL
120
SEM Made Simple
Zainudin Awang (2015). MPWS Publisher
Table 16: The Summary – The Significance of direct and indirect effects
Based on the results in Table 16, the researcher can conclude that the result of bootstrapping is
consistent with mediation test result in Table 12.
121
SEM Made Simple
Zainudin Awang (2015). MPWS Publisher
Figure 1: The Standardized Path Coefficient for every path in the model
122
SEM Made Simple
Zainudin Awang (2015). MPWS Publisher
Figure 2: The Regression Path Coefficient for every path in the model
Note: KDS, KPS, KTP, KMF, OP1, and OP2 are six exogenous constructs while Job Stress is
a mediator and Job Satisfaction is an endogenous construct in the model.
123
SEM Made Simple
Zainudin Awang (2015). MPWS Publisher
Table 1: The Regression Path Coefficient and also the Standardized Path Coefficient
between constructs in the model
Std
Construct Construct Estimate S.E. C.R. P Result
Estimate
Job_Stress <--- KTP -.066 -.176 .033 -2.016 0.044 Significant
Job_Stress <--- KDS -.073 -.239 .036 -1.990 0.047 Significant
Job_Stress <--- KPS -.099 -.180 .032 -3.078 0.002 Significant
Job_Satisfaction <--- KTP -.039 -.049 .059 -.655 0.513 Not Significant
Job_Satisfaction <--- KDS -.052 -.140 .066 -.780 0.436 Not Significant
Job_Satisfaction <--- KPS -.150 -.041 .060 -2.517 0.012 Significant
Job_Satisfaction <--- Job_Stress -2.237 -.868 .148 -15.159 0.000 Significant
Table A: Testing STRESS as a Mediator in the relationship between KTP and SATISFACTION
Based on the above test, the mediation test is supported and the type of mediation is Full
Mediation since the direct effect is not significant. However, the researcher needs to confirm
with the results of direct effect and indirect effect from Bootstrapping. The Bootstrapping
result is shown in Table A1.
124
SEM Made Simple
Zainudin Awang (2015). MPWS Publisher
Table A1: The Bootstrapping Result Shows the Significance of Direct and Indirect Effect
The bootstrapping result also indicates the Full Mediation, and it is consistent with the normal
testing procedure.
Table B: Testing STRESS as a Mediator in the relationship between KDS and SATISFACTION
Based on the above test, the mediation test is supported and the type of mediation is Full
Mediation since the direct effects is not significant. However, the researcher needs to confirm
with the results of direct effect and indirect effect from Bootstrapping. The Bootstrapping
result is shown in Table B1.
Table B1: The Bootstrapping Result Shows the Significance of Direct and Indirect Effect
Indirect Effect P-Value Direct Effect P-Value
Bootstrapping P-Value 0.050 0.495
Result Significant Not Significant
Type of Mediation Full Mediation since direct effect is not significant
The bootstrapping result indicates the Full Mediation, and it is consistent with the normal
testing procedure which indicates Full Mediation.
125
SEM Made Simple
Zainudin Awang (2015). MPWS Publisher
Table C: Testing STRESS as a Mediator in the relationship between KPS and SATISFACTION
Based on the above test, the mediation test is supported and the type of mediation is Partial
Mediation since the direct effect is also significant. However, the researcher needs to confirm
with the results of direct effect and indirect effect from Bootstrapping. The Bootstrapping
result is shown in Table C1.
Table C1: The Bootstrapping Result Shows the Significance of Direct and Indirect Effect
Indirect Effect P-Value Direct Effect P-Value
Bootstrapping P-Value 0.007 0.035
Result Significant Significant
Type of Mediation Partial Mediation since direct is also significant
The bootstrapping result indicates the Partial Mediation; this is consistent with the result
using normal testing procedure which also indicates Partial Mediation.
126
SEM Made Simple
Zainudin Awang (2015). MPWS Publisher
Supposed we are working with the following model (Figure 1). In this model, supposed the
researcher is interested to carry out the following analysis:
1) To determine whether Job Attitude mediates the relationship between Skills & Training
and Career Advancement.
2) To measure the effect size for every single path in the model
3) To measure the mediated effect of the mediator in the model
r2XMY= 0.67
Figure 1: The model containing all constructs namely Skill & Training (X), Job Attitude (M), and
Career Advancement (Y).
The coefficients and their probability values are summarized in Table A. As discussed earlier,
these values can be used to determine the significance of a mediator in the model.
127
SEM Made Simple
Zainudin Awang (2015). MPWS Publisher
From the results in Figure 1, the indirect effect is 0.537 (0.88 * 0.61) higher than the direct
effect of 0.22. Thus, we can conclude that the construct Job Attitude is a mediator in the
relationship between Skills & Training and Career Advancement. The type of mediation here is
full mediation since the direct effect is not significant.
Now, we focus on computing the two effect sizes (objective 2 and 3). The computation will be
based on figures obtained in Figure 1, Figure 2, and Figure 3.
r2XY = 0.59
Figure 2: The model containing construct Skills and Training (X) and Career Advancement (Y)
128
SEM Made Simple
Zainudin Awang (2015). MPWS Publisher
Table B: The Regression Weights when Skills & Training is a sole predictor
r2MY = 0.66
Figure 3: The model containing Job Attitude (M) and Career Advancement (Y)
Effect size is the amount of variance explained in the mediation model contributed by every
single path. The researcher might also be interested to know the relative contribution of
individual paths in the mediation model, especially the effect of a mediator on the dependent
variable. The researcher could assess the effect size for every path in the model, and also the
mediated effect of the mediator variable itself. Specifically, the two effect sizes are:
129
SEM Made Simple
Zainudin Awang (2015). MPWS Publisher
(1) The r2 Measures the effect size of individual path in the model (3 paths)
(2) The r2 Measures the size of mediated effect of a mediator in the model
The range of values of r2 and its relative effect size based on definition by Cohen (1988) is
given in Table C
Table C: The Cohen's (1988) benchmark range of effect sizes:
1) r2 measures for the individual path in the mediation model is defined as follows
i) r2XM represents the squared partial correlation between the X and M variables in
the model,
ii) r2XY denotes the squared partial correlation between X and Y variables in the
model,
iii) r2MY.X corresponds to the squared partial correlation between the M and Y
variables when the influence of variable X removed,
For the above example - the following measures are obtained:
i) The effect size of Skills & Training (X) on Job Attitude (M) is 0.78 (Figure 1)
ii) The effect size of Skills & Training (X) on Career Advancement (Y) is 0.59 (Figure 2)
iii) The effect size of Job Attitude (M) on Career Advancement (Y), controlling for Skills
& Training. This is equivalent to r2XMY - r2XY = 0.67 – 0.59 = 0.08
(Note: r2XMY=0.67 - Figure 1)
Conclusion: The effect size of a mediator on DV is in small range
130
SEM Made Simple
Zainudin Awang (2015). MPWS Publisher
Thus, based on Cohen (1988), the mediated effect size of the mediator (Job Attitude) in
the relationship between Skills & Training and Career Advancement is large.
In Class Exercises:
The power to detect the significance of mediation effects is always lower than the power to
detect the significance of the main effects because the magnitude of the mediated effect is
bounded by the individual coefficients from which it is formed. Recent research has shown that
causal steps tests for mediation and normal theory point estimators of the mediated effect are
underpowered (MacKinnon et al., 2002).
Methodological work has shown that asymmetric confidence limits based on the distribution of
the product and re-sampling methods such as the percentile bootstrap and the bias-corrected
131
SEM Made Simple
Zainudin Awang (2015). MPWS Publisher
bootstrap give the best combination of low Type 1 error rates and power to detect effects in
mediation models (MacKinnon et al., 2002). Consequently, these methods should be
implemented to test mediation hypotheses in substantive research.
In some studies the structural model consists of many mediator constructs, and it is very
challenging to test the hypothesis for mediation effects in this model. The example in Figure 1
consists of five latent constructs (three second order and two first order constructs).
132
SEM Made Simple
Zainudin Awang (2015). MPWS Publisher
133