7.6.2 Path-by-Path Moderation
In the second step, path-by-path moderation is examined to
check which paths are being moderated on an individual
level. Path-by-path moderation is calculated by using ‘regression weights’ of groups and ‘critical ratios’ for the differences between the parameters (group differences), using
tools within the ‘Stats Tools Package’ (see Stats Wiki and
Stats Tools Package http://statwiki.kolobkreations.com/
index.php?title=Main_Page). The estimates, p-values, and
z-scores are calculated for each group and every path. Following the same procedure, path-by-path tests for other
variables were run. However, path-by-path tests were executed only for the paths that are moderated by the moderating variables.
• Path-by-Path moderation of experience: Data were
collected in different groups of experience (high experience and low experience) and a comparison was made. In
this test, two paths were found to be moderated, that is,
FC ! BI and FC ! UB.
– FC ! BI: The experience has a moderating impact
between FC ! BI (t = 1.661*, p<0.1) such that for
high experience this relationship (0.374 at p<0.01) is
significant; also for low experience this relationship is
significant (0.155 at p<0.05). In other words, the
relationship between FC ! BI increases and is stronger (almost triple) for high-experience instructors.
– FC ! UB: There is a moderating effect of experience
between FC ! UB (t = 2.8***, p<0.01) such that for
Table 7.6 Summary of path coefficient of the constructs−path analysis
Paths
Regression Weights
S.E.
C.R.
P-Value
Estimate
Stand
Estimate
BI
PE
.259
.246
.055
4.711
***
BI
EE
.188
.198
.056
3.340
***
BI
SI
.144
.148
.056
2.569
.010
BI
FC
.268
.270
.062
4.292
***
BI
M
.169
.214
.043
3.926
***
UB
FC
.035
.034
.078
.447
.655
UB
BI
.470
.457
.083
5.687
***
Table 7.7 Moderation test at model-level
Moderation Test at Model-Level
Dimension
v
2
/d
Unconstrained
Fully
constrained
Difference
Number
of groups
p-value
Invariant
Interpretation
Age
Chi-square
1078.572
1091.427
12.855
2
0.169
Yes
No
moderation exists
Df
808
817
2
Experience
Chi-square
1042.94
1062.973
20.033
2
0.018
No
Groups are different
at the model level
Df
808
817
9
Awareness
Chi-square
757.578
769.616
12.038
2
0.099
No
Groups are different
at the model level
Df
656
663
7
Power
Distance
Chi-square
783.844
800.461
16.617
2
0.011
No
Groups are different
at the model level
Df
656
662
6
Uncertainty
Avoidance
Chi-square
1010.478
1038.346
27.868
2
0.001
No
Groups are different
at the model level
Df
808
817
9
Individualism
Chi-square
988.291
996.182
7.891
2
0.639
Yes
No moderation exists
Df
808
818
10
Masculinity
Chi-square
989.456
999.185
9.729
2
0.373
Yes
No moderation exists
Df
808
817
9
7.6 Tests for Moderation
69
In the second step, path-by-path moderation is examined to
check which paths are being moderated on an individual
level. Path-by-path moderation is calculated by using ‘regression weights’ of groups and ‘critical ratios’ for the differences between the parameters (group differences), using
tools within the ‘Stats Tools Package’ (see Stats Wiki and
Stats Tools Package http://statwiki.kolobkreations.com/
index.php?title=Main_Page). The estimates, p-values, and
z-scores are calculated for each group and every path. Following the same procedure, path-by-path tests for other
variables were run. However, path-by-path tests were executed only for the paths that are moderated by the moderating variables.
• Path-by-Path moderation of experience: Data were
collected in different groups of experience (high experience and low experience) and a comparison was made. In
this test, two paths were found to be moderated, that is,
FC ! BI and FC ! UB.
– FC ! BI: The experience has a moderating impact
between FC ! BI (t = 1.661*, p<0.1) such that for
high experience this relationship (0.374 at p<0.01) is
significant; also for low experience this relationship is
significant (0.155 at p<0.05). In other words, the
relationship between FC ! BI increases and is stronger (almost triple) for high-experience instructors.
– FC ! UB: There is a moderating effect of experience
between FC ! UB (t = 2.8***, p<0.01) such that for
Table 7.6 Summary of path coefficient of the constructs−path analysis
Paths
Regression Weights
S.E.
C.R.
P-Value
Estimate
Stand
Estimate
BI
PE
.259
.246
.055
4.711
***
BI
EE
.188
.198
.056
3.340
***
BI
SI
.144
.148
.056
2.569
.010
BI
FC
.268
.270
.062
4.292
***
BI
M
.169
.214
.043
3.926
***
UB
FC
.035
.034
.078
.447
.655
UB
BI
.470
.457
.083
5.687
***
Table 7.7 Moderation test at model-level
Moderation Test at Model-Level
Dimension
v
2
/d
Unconstrained
Fully
constrained
Difference
Number
of groups
p-value
Invariant
Interpretation
Age
Chi-square
1078.572
1091.427
12.855
2
0.169
Yes
No
moderation exists
Df
808
817
2
Experience
Chi-square
1042.94
1062.973
20.033
2
0.018
No
Groups are different
at the model level
Df
808
817
9
Awareness
Chi-square
757.578
769.616
12.038
2
0.099
No
Groups are different
at the model level
Df
656
663
7
Power
Distance
Chi-square
783.844
800.461
16.617
2
0.011
No
Groups are different
at the model level
Df
656
662
6
Uncertainty
Avoidance
Chi-square
1010.478
1038.346
27.868
2
0.001
No
Groups are different
at the model level
Df
808
817
9
Individualism
Chi-square
988.291
996.182
7.891
2
0.639
Yes
No moderation exists
Df
808
818
10
Masculinity
Chi-square
989.456
999.185
9.729
2
0.373
Yes
No moderation exists
Df
808
817
9
7.6 Tests for Moderation
69