Escolar Documentos
Profissional Documentos
Cultura Documentos
Smumn^ - f^^^n/^.^^nal^ II
CHAPTER - V
TRAINING - AN ANALYSIS II (STATISTICAL ANALYSIS)
Result:
Hence the y^ value is greater than the table value we reject the hypotheses.
There is association between age and overall training programme in BHEL. This test
shows that sig is below.05 there is association between age and training. In the case phi
= .264, which shows a positive relationship between two variables. The correlation is
flagged as significant, with same p-value that was given in the chi-square test.
173
Table - 58
Hypotheses:
Result:
Hence the x2 value is less than the table value we accept the hypotheses. There
is no association between Sex and Overall Training Programme in BHEL. This test
shows that sig is above.05 there is no association between age and training. In the case
phi - .083, which shows a weak relationship between two variables. The correlation is
flagged as no significant, with same p-value that was given in the chi-square test.
174
Table - 59
Hypotheses:
Result:
Hence the x2 value is less than the table value we accept the hypotheses. There
This test shows that sig is above.05 there is no association between marital status and
training. In the case phi = .094, which signifies a weak relationship between two
variables. The correlation is flagged as no significant, with same p-value that was given
175
Table - 60
Hypotheses:
Result:
Hence the x2 value is greater than the table value we reject the hypotheses.
This test shows that sig is below.05 there is association between qualification and
training. In the case phi = .302, which reveals a strong relationship between two
variables. The correlation is flagged as significant, with same p-value that was given in
176
Table - 61
Hypotheses:
in BHEL.
Result:
Hence the x2 value is greater than the table value we reject the hypotheses.
This test shows that sig is below.05 there is association between qualification and
training. In the case phi = .210, which shows a mild relationship between two variables.
The correlation is flagged as significant, with same p-value that was given in the chi-
square test.
177
Table - 62
Hypotheses:
There is no association between Work Experience and Overall Training
Programme in BHEL.
Overall Training Programme in BHEL
Cross Tabulation Highly Total
Satisfied Moderate Dissatisfied
Satisfied
5 - 10 0 12 0 0 12
11 - 15 13 7 1 0 21
Work
16-20 7 9 0 0 16
Experience
21 - 30 179 204 15 0 398
31 and above 81 100 6 6 193
Total 280 332 22 6 640
Degrees of Freedom =12 Chi Square Value = 29.100 Table Value =28.3
The Distribution is significant.
Result:
Hence the x2 value is greater than the table value we reject the hypotheses.
BHEL. This test shows that sig is below.05 there is association between qualification
and training. In the case phi = .213, which signifies a mild relationship between two
variables. The correlation is flagged as significant, with same p-value that was given in
178
Table - 63
Hypotheses:
Result:
Hence the x2 value is greater than the table value we reject the hypotheses.
There is association between Age and Growth & Result. This test shows that sig is
below.05 there is association between qualification and training. In the case phi = .218,
which reveals a mild relationship between two variables. The correlation is flagged as
significant, with same p-value that was given in the chi-square test.
179
Table - 64
Hypotheses :
Result:
Hence the ^2 value is greater than the table value we reject the hypotheses.
There is association between Sex and Growth & Result. This test shows that sig is
below.05 there is association between qualification and training. In the case phi = .143,
which shows a weak relationship between two variables. The correlation is flagged as
significant, with same p-value that was given in the chi-square test.
180
Table - 65
Hypotheses :
Result:
Hence the x2 value is less than the table value we accept the hypotheses. There
is no association between Marital Status and Growth & Result. This test shows that sig
is below.05 there is association between qualification and training. In the case phi=.125,
which signifies a weak relationship between two variables. The correlation is flagged as
significant, with same p-value that was given in the chi-square test.
181
Table - 66
Hypotheses :
Degrees of Freedom =12 Chi Square Value = 91.383 Table Value - 28.3
The Distribution is significant.
Result:
Hence the x2 value is greater than the table value we reject the hypotheses.
There is association between Qualification and Growth & Result. This test shows that
sig is below.05 there is association between qualification and training. In the case
phi = .378, which shows a strong relationship between two variables. The correlation is
flagged as significant, with same p-value that was given in the chi-square test.
182
Table - 67
Hypotheses :
There is no association between Designation and Growth & Resuh.
Result:
Hence the %1 value is greater than the table value we reject the hypotheses.
There is association between Designation and Growth & Result. This test shows that sig
phi = .294, which shows a mild relationship between two variables. The correlation is
flagged as significant, with same p-value that was given in the chi-square test.
183
Table - 68
Hypotheses:
Result:
Hence the ^2 value is greater than the table value we reject the hypotheses.
There is association between Work Experience and Growth & Result. This test shows
that sig is below.05 there is association between qualification and training. In the case
phi = .375, which reveals a strong relationship between two variables. The correlation is
flagged as significant, with same p-value that was given in the chi-square test.
184
Table - 69
Hypotheses :
Result:
Hence the ^2 value is greater than the table value we reject the hypotheses.
There is association between Age and Effects & Benefits. This test shows that sig is
below.05 there is association between Age and Effects & Benefits. In the case phi=.180,
which shows a weak relationship between two variables. The correlation is flagged as
significant, with same p-value that was given in the chi-square test.
185
Table - 70
Hypotheses :
Result:
Hence the x2 value is greater than the table value we reject the hypotheses.
There is association between Sex and Effects & Benefits. This test shows that sig is
below.05 there is association between Sex and Effects & Benefits. In the case phi =
.157, which shows a weak relationship between two variables. The correlation is
flagged as significant, with same p-value that was given in the chi-square test.
186
Table - 71
Hypotheses :
There is no association between Qualification and Effects & Benefits.
Result:
Hence the %L value is greater than the table value we reject the hypotheses.
There is association between Qualification and Effects & Benefits. This test shows that
sig is below. 05 there is association between Qualificafion and Effects & Benefits. In the
case phi = .215, which shows a mild relationship between two variables. The correlation
isflaggedas significant, with same p-vaiue that was given in the chi-square test.
187
Table - 72
Hypotheses:
Result:
Hence the x2 value is greater than the table value we reject the hypotheses.
There is association between Designation and Effects & Benefits. This test shows that
sig is below.05 there is association between Designation and Effects & Benefits. In the
case phi = .195, which shows a weak relationship between two variables. The
correlation is flagged as significant, with same p-value that was given in the chi-square
test.
188
Table - 73
Hypotheses:
Result:
Hence the x2 value is greater than the table value we reject the hypotheses.
There is association between Work Experience and Effects & Benefits. This test shows
that sig is below.05 there is association between Work Experience and Effects &
Benefits. In the case phi = .220, which shows a mild relationship between two variables.
The correlation is flagged as significant, with same p-value that was given in the chi-
square test.
189
Table - 74
Hypotheses :
Result:
Hence the yl value is greater than the table value we reject the hypotheses.
There is association between Designation and individual goal attainment. This test
shows that sig is below .05 there is association between Designation and individual goal
attainment. In the case phi = .237, which shows a mild relationship between two
variables. The correlation is flagged as significant, with same p-value that was given in
190
Table - 75
Hypotheses:
There is no association between Designation and Group Goal Attainment.
Result:
Hence the x2 value is less than the table value we accept the hypotheses. There
is no association between Designation and group goal attainment. This test shows that
sig is below.05 there is association between Designation and group goal attainment. In
the case phi^.lOl, which shows a weak relationship between two variables. The
correlation is flagged as significant, with same p-value that was given in the chi-square
test.
191
Table - 76
Hypotheses:
Goal.
Result:
Hence the ^2 value is less than the table value we accept the hypotheses. There
shows that sig is below.05 there is association between Designation and organziation
goal attainment. In the case phi=.136, which shows a weak relationship between two
variables. The correlation is flagged as significant, with same p-value that was given in
192
FACTOR ANALYSIS
Table - 77
based on 8 individual statements and the reliability of the subsequent factor structures
was then tested for internal consistency of the grouping of the items.
indicates that factor analysis is appropriate for the given data set. KMO measure of
values between 0.5 and 1.0 indicate factor analysis is appropriate. Values below 0.5
Bartlett's Test of Sphericity is used to examine the hypothesis that the variables
correlation matrix. A large value of the test statistic will favour the rejection of the null
the 8 statements are correlated and hence as concluded in KMO, factor analysis is
193
Communalities
Initial Extraction
The norms and values of the organisation are clearly
1.000 .435
explained to the new employees during induction
The new recruits finds induction training very useful in your
1.000 .732
organisation
The Induction training is periodically evaluated and
1.000 .642
improved
The duration provided for training at present is adequate 1.000 .594
The training equipment, lab, model rooms, etc are 1.000 .700
How do you rate the tools / study materials provided for
1.000 .752
training
How do you feel about the contents covered in the training
1.000 .633
program
Extraction Method: Principal Component Analysis.
explained by all the factors. That is, the communality is the squared multiple correlation
for the variable using the factors as predictors. Communality for a variable is the sum of
squared factor loadings for that variable (row), and thus is the percent of variance in a
given variable explained by all the factors. For full orthogonal PCA, the communality
will be 1.0 and all of the variance in the variables will be explained by all of the factors,
which will be as many as there are variables. The "extracted" communalities is the
percent of variance in a given variable explained by the factors which are extracted,
which will usually be fewer than all the possible factors, resulting in coefficients less
than 1.0.
194
Total Variance Explained
Eigen Value represents the total variance explained by each factor. Percentage
of the total variance attributed to each factor. One of the popular methods used in
Exploratory Factor Analysis is Principal Component Analysis, Where the total variance
in the data is considered to determine the minimum number of factors that will account
Rotation of factors is transferred through rotation into a simpler one that is easier
to interpret. It does not affect the percentage of total variance explained. However, the
commonly used method is Varimax rotation procedure. This procedure maximizes the
variance of the loadings on each factor, thus minimizing the complexity of the factors.
195
Rotated Component Matrix
Component
1 2
The norms and values of the organisation are clearly explained
.660 .001
to the new employees during induction
The new recruits finds induction training very useful in your
.855 .038
organisation
The Induction training is periodically evaluated and improved .705 .381
The duration provided for training at present is adequate .649 .417
The infrastructure for training is sufficient .536 .520
The training equipment, lab, model rooms, etc are .345 .763
How do you rate the tools / study materials provided for
-.042 .866
training
How do you feel about the contents covered in the training
.715 .349
program
Extraction Method: Principal Component Analysis.
Rotation Method: Varimax with Kaiser Normalization.
large loadings in the same factor. The factors can be interpreted in terms of the
statements are (1) The new recruits find induction training very useful in organization,
The 2 statements accounted for 63.080 percent of the variance in the original 8
196
Table - 78
analysis based on 4 individual statements and the reliability of the subsequent factor
structures was then tested for internal consistency of the grouping of the items.
indicates that factor analysis is appropriate for the given data set. KMO measure of
values between 0.5 and 1.0 indicate factor analysis is appropriate. Values below 0.5
Bartlett's Test of Sphericity is used to examine the hypothesis that the variables
correlation matrix. A large value of the test statistic will favour the rejection of the null
the 4 statements are correlated and hence as concluded in KMO, factor analysis is
197
Communalities
Initial Extraction
I feel training is not very difficult 1.000 .640
Problem faced in training can be easily .569
1.000
expressed to higher authority
Is there any provision to convey the difficulties
1.000 .614
faced (if any) at the time of training
Response to suggestions 1.000 .626
Extraction Method: Principal Component Analysis.
explained by all the factors. That is, the communality is the squared multiple correlation
for the variable using the factors as predictors. Communality for a variable is the sum of
squared factor loadings for that variable (row), and thus is the percent of variance in a
given variable explained by all the factors. For full orthogonal PCA, the communality
will be 1.0 and all of the variance in the variables will be explained by all of the factors,
which will be as many as there are variables. The "extracted" communalities is the
percent of variance in a given variable explained by the factors which are extracted,
which will usually be fewer than all the possible factors, resulting in coefficients less
than 1.0.
Eigen Value represents the total variance explained by each factor. Percentage
of the total variance attributed to each factor. One of the popular methods used in
Exploratory Factor Analysis is Principal Component Analysis, Where the total variance
198
in the data is considered to determine the minimum number of factors that will account
Rotation of factors is transferred through rotation into a simpler one that is easier
to interpret. It does not affect the percentage of total variance explained. However, the
commonly used method is Varimax rotation procedure. This procedure maximizes the
variance of the loadings on each factor, thus minimizing the complexity of the factors.
Component
1 2
I feel training is not very difficult .790 .125
Problem faced in training can be easily expressed
-.062 .752
to higher authority
Is there any provision to convey the difficulties
-.773 .131
faced (if any) at the time of training
Response to your suggestions .057 .789
extraction Method: Principal Component Analysis.
Rotation Method: Varimax with Kaiser Normalization.
Interpretation of factors is facilitated by identifying the statements that have
large loadings in the same factor. The factors can be interpreted in terms of the
statements in problem faced, 2 statements contribute more towards problem faced. The
statements are (1) I feel training is not very difficult, (2) Response to suggestions.
The 2 statements accounted for 61.244 percent of the variance in the original 4
199
Table - 79
examined using factor analysis based on 10 individual statements and the reliability of
the subsequent factor structures was then tested for internal consistency of the grouping
of the items.
indicates that factor analysis is appropriate for the given data set. KMO measure of
values between 0.5 and 1.0 indicate factor analysis is appropriate. Values below 0.5
Bartlett's Test of Sphericity is used to examine the hypothesis that the variables
correlation matrix. A large value of the test statistic will favour the rejection of the null
the 10 statements are correlated and hence as concluded in KMO, factor analysis is
200
Communalities
Initial Extraction
Employees are sponsored for training programmes on the basis
1.000 .506
of genuine training needs
Those who are sponsored for the training programmes take the
1.000 .722
training sincerely
Employees are sponsored for training programmes on the basis
1.000 .584
of carefully identified developmental needs
Employees sponsored for training go with a clear understanding
of the skills and knowledge they are expected to acquire from 1.000 .750
the training
How do you feel about the planning / execution of the entire
1.000 .294
training programme
Employees returning from the training are given adequate free
1.000 .771
time to reflect and plan improvements in the organisation
Training and Development programmes are handled by
1.000 .806
experienced and competent faculty
The performance appraisal of your superior after training is 1.000 .563
The quality of the training programmes in your organisation is
1.000 .646
excellent
Training programmes are selected, after collecting enough
1.000 .549
information about their suitability.
Extraction Method: Principal Component Analysis.
explained by all the factors. That is, the communality is the squared multiple correlation
for the variable using the factors as predictors. Communality for a variable is the sum of
squared factor loadings for that variable (row), and thus is the percent of variance in a
given variable explained by all the factors. For full orthogonal PCA, the communality
will be 1.0 and all of the variance in the variables will be explained by all of the factors,
which will be as many as there are variables. The "extracted" communalities is the
percent of variance in a given variable explained by the factors which are extracted,
which will usually be fewer than all the possible factors, resulting in coefficients less
than 1.0.
201
Total Variance Explained
Eigen Value represents the total variance explained by each factor. Percentage
of the total variance attributed to each factor. One of the popular methods used in
Exploratory Factor Analysis is Principal Component Analysis, Where the total variance
in the data is considered to determine the minimum number of factors that will account
Rotation of factors is transferred through rotation into a simpler one that is easier
to interpret. It does not affect the percentage of total variance explained. However, the
commonly used method is Varimax rotation procedure. This procedure maximizes the
variance of the loadings on each factor, thus minimizing the complexity of the factors.
202
Rotated Component Matrix
Component
1 2
Employees are sponsored for training programmes on the basis of
.622 .344
genuine training needs
Those who are sponsored for the training programmes take the
.793 .305
training sincerely
Employees are sponsored for training programmes on the basis of
.244 .724
carefully identified developmental needs
Employees sponsored for training go with a clear understanding of
the skills and knowledge they are expected to acquire from the .326 .802
training
How do you feel about the planning / execution of the entire
.398 .369
training programme
Employees returning from the training are given adequate free
.819 .318
time to reflect and plan improvements in the organisation
Training and Development programmes are handled by
.889 .128
experienced and competent faculty
The performance appraisal of your superior after training is .712 .236
The quality of the training programmes in your organisation is
.513 .619
excellent
Training programmes are selected, after collecting enough
.121 .731
information about their suitability.
Extraction Method: Principal Component Analysis.
Rotation Method: Varimax with Kaiser Normalization.
large loadings in the same factor. The factors can be interpreted in terms of the
towards Selection, Evaluation and feed back. The statements are (1) Training
sponsored for training go with a clear understanding of the skills and knowledge they
The 2 statements accounted for 61.920 percent of the variance in the original 10
Evaluation and feed back in training, (i.e) 38.08 percent of variance only.
203
Table - 80
was examined using factor analysis based on 8 individual statements and the reliability
of the subsequent factor structures was then tested for internal consistency of the
indicates that factor analysis is appropriate for the given data set. KMO measure of
values between 0.5 and 1.0 indicate factor analysis is appropriate. Values below 0.5
Bartlett's Test of Sphericity is used to examine the hypothesis that the variables
correlation matrix. A large value of the test statistic will favour the rejection of the null
the 8 statements are correlated and hence as concluded in KMO, factor analysis is
204
Communalities
Initial Extraction
The training programs were relevant to my growth and
1.000 .614
development
The training programmes were helpful in upgrading my skills 1.000 .751
Training programmes helped me to remain updated with the
1.000 .609
latest technology / practices / policies
Training programs contribute in improving the overall
1.000 .595
performance of the Organization
After the training programme, I am likely to get more
1.000 .816
incentives than before
Training programs help in bridging the gap between
1.000 .645
supervisor and subordinate
After attending the training programme, I think that I can be
of great help in maintaining a cordial and orderly atmosphere 1.000 .648
in the organisation
Training contributes to a large extent in improving the
1.000 .646
confidence and commitment of an employee.
Extraction Method: Principal Component Analysis.
explained by all the factors. That is, the communality is the squared multiple correlation
for the variable using the factors as predictors. Communality for a variable is the sum of
squared factor loadings for that variable (row), and thus is the percent of variance in a
given variable explained by all the factors. For full orthogonal PCA, the communality
will be 1.0 and all of the variance in the variables will be explained by all of the factors,
which will be as many as there are variables. The "extracted" communalities is the
percent of variance in a given variable explained by the factors which are extracted,
which will usually be fewer than all the possible factors, resulting in coefficients less
than 1.0.
205
Total Variance Explained
Eigen Value represents the total variance explained by each factor. Percentage
of the total variance attributed to each factor. One of the popular methods used in
Exploratory Factor Analysis is Principal Component Analysis, Where the total variance
in the data is considered to determine the minimum number of factors that will account
Rotation of factors is transferred through rotation into a simpler one that is easier
to interpret. It does not affect the percentage of total variance explained. However, the
commonly used method is Varimax rotation procedure. This procedure maximizes the
variance of the loadings on each factor, thus minimizing the complexity of the factors.
206
Rotated Component Matrix
Component
1 2
The training programs were relevant to my growth and
.610 .492
development
The training programmes were helpful in upgrading my skills .866 -.012
Training programmes helped me to remain updated with the latest
.567 .536
technology / practices / policies
Training programs contribute in improving the overall
.731 .245
performance of the Organization
After the training programme, I am likely to get more incentives
-.017 .903
than before
Training programs help in bridging the gap between supervisor
.803 -.007
and subordinate
After attending the training programme, I think that I can be of
great help in maintaining a cordial and orderly atmosphere in the .756 .276
organisation
Training contributes to a large extent in improving the confidence
.770 .233
and commitment of an employee.
extraction Method: Principal Component Analysis.
Rotation Method: Varimax with Kaiser Normalization.
large loadings in the same factor. The factors can be interpreted in terms of the
relationships. The statements are (1) The training programmes were helpful in
upgrading my skills, (2) After the training programme, I am likely to get more
The 2 statements accounted for 66.536 percent of the variance in the original 8
207
Table - 81
Dimension : Attitude
based on 4 individual statements and the reliability of the subsequent factor structures
was then tested for internal consistency of the grouping of the items.
indicates that factor analysis is appropriate for the given data set. KMO measure of
values between 0.5 and 1.0 indicate factor analysis is appropriate. Values below 0.5
Bartlett's Test of Sphericity is used to examine the hypothesis that the variables
correlation matrix. A large value of the test statistic will favour the rejection of the null
the 4 statements are correlated and hence as concluded in KMO, factor analysis is
208
Communalities
Initial Extraction
The training programmes have reinforced my belief in
1.000 .716
the usefulness of training
I will not like to miss the training programs conducted
1.000 .807
by my organisation in the future
Training programmes should be taken more seriously 1.000 .767
Training is an ongoing process 1.000 .994
Extraction Method: Principal Component Analysis.
explained by all the factors. That is, the communality is the squared multiple correlation
for the variable using the factors as predictors. Communality for a variable is the sum of
squared factor loadings for that variable (row), and thus is the percent of variance in a
given variable explained by all the factors. For full orthogonal PCA, the communality
will be 1.0 and all of the variance in the variables will be explained by all of the factors,
which will be as many as there are variables. The "extracted" communalities is the
percent of variance in a given variable explained by the factors which are extracted,
which will usually be fewer than all the possible factors, resulting in coefficients less
than 1.0.
209
Rotation Sums of Squared Loadings
Total Vo of Variance Cumulative %
2.272 56.792 56.792
1.012 25.306 82.098
Rotation of factors is transferred through rotation into a simpler one that is easier
to interpret. It does not affect the percentage of total variance explained. However, the
commonly used method is Varimax rotation procedure. This procedure maximizes the
variance of the loadings on each factor, thus minimizing the complexity of the factors.
Component
1 2
The training programmes have reinforced my belief in the
.843 .065
usefulness of training
I will not like to miss the training programs conducted by my
.893 .093
organisation in the future
Training programmes should be taken more seriously .873 -.077
Training is an ongoing process .030 .997
Extraction Method: Principal Component Analysis.
Rotation Method: Varimax with Kaiser Normalization.
large loadings in the same factor. The factors can be interpreted in terms of the
statements, 2 statements contribute more towards attitude. The statements are (1) I will
not like to miss the training programs conducted by my organization in the future, (2)
The 2 statements accounted for 82.098 percent of the variance in the original 4
210
Table - 82
factor analysis based on 6 individual statements and the reliability of the subsequent
factor structures was then tested for internal consistency of the grouping of the items.
indicates that factor analysis is appropriate for the given data set. KMO measure of
values between 0.5 and 1.0 indicate factor analysis is appropriate. Values below 0.5
Bartlett's Test of Sphericity is used to examine the hypothesis that the variables
correlation matrix. A large value of the test statistic will favour the rejection of the null
the 6 statements are correlated and hence as concluded in KMO, factor analysis is
211
Communalities
Initial Extraction
Training programme helps to increase the production 1.000 .661
Grievance is minimized after training 1.000 .427
Do you agree absenteeism is minimized after training 1.000 .387
Does your morale and efficiency increased after training 1.000 .405
Training helped you to minimize mistakes and increase the
1.000 .387
job satisfaction
Can you make your job more interesting after training 1.000 .384
Extraction Method: Principal Component Analysis.
explained by all the factors. That is, the communality is the squared multiple correlation
for the variable using the factors as predictors. Communality for a variable is the sum of
squared factor loadings for that variable (row), and thus is the percent of variance in a
given variable explained by all the factors. For full orthogonal PCA, the communality
will be 1.0 and all of the variance in the variables will be explained by all of the factors,
which will be as many as there are variables. The "extracted" communalities is the
percent of variance in a given variable explained by the factors which are extracted,
which will usually be fewer than all the possible factors, resulting in coefficients less
than 1.0.
212
Eigen Value represents the total variance explained by each factor. Percentage
of the total variance attributed to each factor. One of the popular methods used in
Exploratory Factor Analysis is Principal Component Analysis, Where the total variance
in the data is considered to determine the minimum number of factors that will account
Rotation of factors is transferred through rotation into a simpler one that is easier
to interpret. It does not affect the percentage of total variance explained. However, the
commonly used method is Varimax rotation procedure. This procedure maximizes the
variance of the loadings on each factor, thus minimizing the complexity of the factors.
Component
1 2
213
Interpretation of factors is facilitated by identifying the statements that have
large loadings in the same factor. The factors can be interpreted in terms of the
effectiveness and benefit. The statements are (1) Morale and efficiency increased after
The 2 statements accounted for 44.190 percent of the variance in the original 6
214
Table - 83
analysis based on 18 individual statements and the reliability of the subsequent factor
structures was then tested for internal consistency of the grouping of the items.
indicates that factor analysis is appropriate for the given data set. KMO measure of
values between 0.5 and 1.0 indicate factor analysis is appropriate. Values below 0.5
Bartlett's Test of Sphericity is used to examine the hypothesis that the variables
correlation matrix. A large value of the test statistic will favour the rejection of the null
the 18 statements are correlated and hence as concluded in KMO, factor analysis is
215
Communalities
Initial Extraction
Your top managers see training as an important way of helping
1.000 .685
to achieve organisation's mission
Your top managers show commitment to Training by spending
1.000 .658
time promoting and delivering it.
The managers strongly support the development of new skills
1.000 .657
and knowledge among all levels of employees
The Training activities are encouraged clearly, what top
1.000 .700
managers are trying to accomplish
There are some Training activities (e.g., diversity, ethics, or
computer security training) that everyone in the organisation 1.000 .687
participates in, regardless of position
The managers help their employees meet personal Training
1.000 .685
goals and needs
Your top managers are closely involved in determining the
1.000 .534
direction and goals of the Training activities
Training activities meets the requirement of the employees 1.000 .449
Employees are held accountable for using what they've learned
1.000 .721
in their Training activities
Managers are held accountable for following up and
encouraging their employees to apply what they've learned 1.000 .617
through their Training activities
Training activities provide learning that is practical for use on
1.000 .621
the job
Managers personally provide Training for their employees 1.000 .732
Training gives employees an opportunity to learn the skills and
1.000 .585
behaviors that will help them to get rewarded and promoted
After employees receive Training, they are asked to provide
1.000 .513
feedback how much they learned
Subordinates are asked to provide feedback on the effectiveness
1.000 .612
of the Training received by their managers
Managers are asked to provide feedback on the effectiveness of
1.000 .604
the Training received by their subordinates
Individuals are publicly recognized for their Training
1.000 .734
accomplishments
In general, I am satisfied with the range of Training
opportunities available to me 1.000 .733
Extraction Method: Principal Component Analysis.
explained by all the factors. That is, the communality is the squared multiple correlation
for the variable using the factors as predictors. Communality for a variable is the sum of
squared factor loadings for that variable (row), and thus is the percent of variance in a
216
given variable explained by all the factors. For full orthogonal PCA, the communality
will be 1.0 and all of the variance in the variables will be explained by all of the factors,
which will be as many as there are variables. The "extracted" communalities is the
percent of variance in a given variable explained by the factors which are extracted,
which will usually be fewer than all the possible factors, resulting in coefficients less
than 1.0.
Eigen Value represents the total variance explained by each factor. Percentage
of the total variance attributed to each factor. One of the popular methods used in
Exploratory Factor Analysis is Principal Component Analysis, Where the total variance
in the data is considered to determine the minimum number of factors that will account
217
Rotation Sums of Squared Loadings
Total Vo of Variance Cumulative %
6.309 35.052 35.052
3.689 20.497 55.549
1.529 8.494 64.043
Rotation of factors is transferred through rotation into a simpler one that is easier
to interpret. It does not affect the percentage of total variance explained. However, the
commonly used method is Varimax rotation procedure. This procedure maximizes the
variance of the loadings on each factor, thus minimizing the complexity of the factors.
218
After employees receive Training, they are asked to provide
.698 .161 -.023
feedback how much they learned
Subordinates are asked to provide feedback on the
.650 .179 .398
effectiveness of the Training received by their managers
Managers are asked to provide feedback on the effectiveness
.749 .173 .117
of the Training received by their subordinates
Individuals are publicly recognized for their Training
.253 .806 .139
accomplishments
In general, I am satisfied with the range of Training
.697 -.293 .402
opportunities available to me
Extraction Method: Principal Component Analysis.
Rotation Method: Varimax with Kaiser Normalization.
large loadings in the same factor. The factors can be interpreted in terms of the
employee perception. The statements are (1) Top managers show commitment to
Training by spending time promoting and delivering it (2) The managers help their
employees meet personal Training goals and needs (3) Managers personally provide
The 3 statements accounted for 60.043 percent of the variance in the original 18
219
CORRELATION
Table - 84
Improve Effecti
Training ment in veness Growtl
Selection Attitude Perception Training
Inputs Relations and & Resu
hip Benefit
Training
1.000
Inputs
Improvement
in .499 .754 1.000
Relationship
Employee
.621 .454 .479 .439 1.000
Perception
Effectiveness
.367 .456 .383 .345 .450 1.000
and Benefit
Growth &
.559 .506 .463 .385 .460 .402 .648 1.000
Result
Training Inputs
The correlation between the two variables training inputs and selection was
.486; both the variables were positively correlated and significant. The correlation
between training inputs and improvement in relationship was .499, which has the
positive correlation between two variables and significant. The correlation between
training inputs and attitude of the respondents was .342, found to be weak correlation
among the correlation between other variables of training. The correlation between
training inputs and employee perception was found to be .621, which has the positive
correlation and significant. The correlation between training inputs and effectiveness
220
and benefits was .367, both the variables were mildly correlated and significant. The
correlation between training inputs and training was .668, which has the positive
correlation and significant. The correlation between training inputs and growth and
results was .559, both the variables are highly correlated and significant.
Selection
The correlation between selection and improvement in relations was .754, both
the variables are highly correlated and significant. The correlation between selections
and attitude, employee perception, effectiveness and benefits. Training, growth and
results were .647, .454, .456, .710, .506 respectively. The correlations between the
Improvement in relationships
and employee perception was .479, positively correlated and significant. The correlation
in relationships and training was found to be highly correlated .673 and significant. The
correlation between improvement in relationships and growth and results was found to
Attitude
The correlation between attitude and employee perception was .439, moderately
correlated and significant. The correlation between attitude and effectiveness and
benefits was .345, mildly correlated and significant. The correlation between variables
221
attitude and training was .580, highly correlated and significant. The correlation
between attitude and growth and result was .385, mildly correlated and significant.
Perception:
training was .450, moderately correlated and significant. The correlation between
perception and training was .605, highly correlated and significant. The correlation
between perception and growth, results was .460, moderately correlated and significant.
The correlation between effectiveness, benefit and training was .518, highly
correlated and significant. The correlation between effectiveness and benefits and
Training
The correlation between training and growth and result was found to be (.648)
222
MULTIPLE REGRESSIONS
Table - 85
Model Summary
The multiple regressions are shown in the above table. The model summary
table shows R-Square for this model is .730. This means that 73 percent of the variation
variables. The table also shows the adjusted R-square for the model as .727.
the R-square will increase (even if only slightly). Consequently, it becomes difficult to
determine which models do the best job of explaining variation in the same dependent
variable. The adjusted R-Square does just what its name implies. It adjusts the R-square
by the number of predictor variables in the model. This adjustment allows the easy
variable. It also helps us to decide how many variables to include in our regression
model.
223
ANOVA
Sum of Mean
Model df F Sig.
Squares Square
Regression 169.007 8 21.126
1 Residual 62.437 631 .099 213.501 .000
Total 231.444 639
Predictor s: (Constan t), Perception of Training , Problem Facec , Attitude towards
training, Effects & Benefits, Growth & Result, Improvement in
Relationships, Training Inputs, Selection.
Dependent Variable: Training
The ANOVA table, as displayed in the above table shows the F ratio for the
regression model that indicates the statistical significance of the overall regression
model. The F ratio is calculated the same way for regression analysis as it was for the
of the total variance in Independent variable that is not associated with dependent
The larger the F ratio that more variance in the dependent variable that is
associated with the independent variable. The F ratio = 213.501. The statistical
significance is .000 - the "Sig". So we can reject the null hypothesis that no relationship
exists between the two variables. There is relationship between independent and
dependent variables.
224
Coefficients
table. All the above five independent statements are statistically significant.
The standardized coefficient beta column reveals that training inputs has a beta
coefficient .242, which is significant (.000). Problem faced has a beta coefficient -.085,
which is significant (.000) but it creates a negative influence towards overall training.
Selection has a beta coefficient .185, which is significant (.000). Growth and result has
beta coefficient .158, which is significant (.000). Atfitude towards training has a beta
coefficient .121, which is significant (.000). Effects and benefits have a beta coefficient
.114, which is significant (.000). Perception has a beta coefficient .110, which is
significant (.000).
225
To assess multicollinearity we look at the size of Tolerance and VIF. For the
tolerance small value indicate the absence of coUinearity. The VIF is the inverse of
tolerance, we look for large values. If the tolerance value is smaller then .10, we
multicollinearity is a problem.
Since the tolerance value is substantially above .10 and the VIF is smaller than
226
MODELING OF SUPERVISOR'S TRAINING,
GROWTH AND RESULTS
.38
/21 Selection
14
.15 I Attitude
.16
Perception
227
Table - 86
Regression Weights:
The above table explains the various factors of training. It is inferred that from
the above table when training inputs goes by one standard deviation, there will be .204
units of changes in overall training. Similarly when selection changes by one unit of
0.06.
228
Level of significance for regression weight
less than 0.001. In other words, the regression weight for Training Inputs in the
level (two-tailed).
less than 0.001. In other words, the regression weight for Selection in the
level (two-tailed).
the prediction of Overall Training is not significantly different from zero at the
.129. In other words, the regression weight for Attitude in the prediction of
Overall Training is not significantly different from zero at the 0.05 level (two-
tailed).
less than 0.001. In other words, the regression weight for Perception in the
level (two-tailed).
less than 0.001. In other words, the regression weight for Effectiveness Benefit
229
7. The probability of getting a critical ratio as large as 10.094 in absolute value is
less than 0.001. In other words, the regression weight for Overall Training in
the prediction of Growth Result is significantly different from zero at the 0.001
level (two-tailed).
Covariances:
Improvement in ***
Attitude <> .232 .025 9.187
Relations
Improvement in ***
<-> Selection .365 .034 10.668
Relations
Selection <> Training Inputs .227 .028 8.062 ***
Improvement in
Perception <> .179 .026 6.946 ***
Relations
Perception <> Selection .143 .027 5.339 ***
Improvement in ***
Effectiveness Benefit <> .146 .023 6.332
Relations
Effectiveness Benefit <> Selection .147 .025 5.958 ***
Improvement in ***
<> Training Inputs .211 .026 8.056
Relations
Estimate of covariance
be. 147.
.104.
230
3. The covariance between Attitude and Improvement in Relations is estimated
to be .232.
to be .365.
estimated to be . 179.
.216.
is estimated to be .146.
.147.
estimated to be .147.
14. The covariance between Attitude and Training Inputs is estimated to be . 131.
estimated to be .211.
231
Variances:
Estimate of variance
232
MODELING OF EXECUTIVES TRAINING,
.24
.64 .38
^^ 1
.22
/
1 Selection
.15
_ _ _ Jt_
11
i
.13
^4
'43
1
Improvement in Relations
.49 .20
.16
I Overall Training
.80
28 ,24 21 .45
r
.30 :
.15 1 Attitude '
.16
^
>, 1 1 7 .05 Growth & Result
,24 .13
.32
233
Table - 87
Regression Weights:
less than 0.001. In other words, the regression weight for Training Inputs in the
level (two-tailed).
234
2. The probability of getting a critical ratio as large as 3.485 in absolute value is
less than 0.001. In other words, the regression weight for Selection in the
level (two-tailed).
less than 0.001. In other words, the regression weight for Improvement in
less than 0.001. In other words, the regression weight for Attitude in the
level (two-tailed).
.312. In other words, the regression weight for Perception in the prediction of
Overall Training is not significantly different from zero at the 0.05 level (two-
tailed).
less than 0.001. In other words, the regression weight for Effectiveness Beneflt
less than 0.001. In other words, the regression weight for Overall Training in
the prediction of Grov'th Result is significantly different from zero at the 0.001
level (two-tailed).
235
Covariances:
Estimate of covariance
be.177
.131
236
3. The covariance between Attitude and Improvement in Relations is estimated
to be .270.
to be .429.
estimated to be .238.
.279.
is estimated to be . 151.
.243.
estimated to be . 110.
14. The covariance between Attitude and Training Inputs is estimated to be . 134.
estimated to be .220.
237
Variances:
Estimate of variance
238
MODELING OF OVERALL TRAINING, GROWTH AND RESULT
.11
Reaction
e2.
Selection
Training
1
Attitude towards training Growth & Result
239
Table -88
Hypotheses:
I. Regression Weights:
The above table explains the impact of various factors on training. From the
above table it is inferred that if the perception of training goes up by one standard
deviation, there will be .102 units of change in standard deviation in the training.
Similarly effects and benefits changes up in one standard deviation, there will be .146
units of change in the training. Like wise Attitude towards training goes up by one
If Training inputs goes up in one standard deviation, there will be .295 changes
training changes in one standard deviation, there will be .197 units of change in training.
When the effectiveness of training goes up in one standard deviation, there will be .699
towards training.
240
Total Effects of Training towards growth and result:
The total (direct and indirect) effect of training inputs on Growth and Result is
.206. That is, due to both direct (unmediated) and indirect (mediated) effects of training
inputs on growth and result, when training inputs goes up by 1, growth and result goes
up by 0.206.
The total (direct and indirect) effect of Selection on Growth Result is .137. That
is, due to both direct (unmediated) and indirect (mediated) effects of Selection on
growth and result, when selection goes up by 1, growth and result goes up by 0.137.
and result is .090. That is, due to both direct (unmediated) and indirect (mediated)
The total (direct and indirect) effect of Attitude towards training on Growth
Result is .086. That is, due to both direct (unmediated) and indirect (mediated) effects of
Attitude towards training on growth and result, when attitude towards training goes up
The total (direct and indirect) effect of Perception of Training on Growth Result
is .072. That is, due to both direct (unmediated) and indirect (mediated) effects of
241
The total (direct and indirect) effect of Effects Benefits on Growth and Result is
.102. That is, due to both direct (unmediated) and indirect (mediated) effects of effects
and benefits on growth result, when effects and benefits goes up by 1, growth and result
goes up by 0.102.
The total (direct and indirect) effect of TRAINrNG on growth and result is .699.
That is, due to both direct (unmediated) and indirect (mediated) effects of TRAINING
on growth and result, when TRAINING goes up by I, growth and result goes up by
0.699.
1. The regression weight estimate,. 102, has a standard error of about .029.
2. The regression weight estimate, .146, has a standard error of about .027.
3. The regression weight estimate,. 123, has a standard error of about .029.
4. The regression weight estimate, .295, has a standard error of about .028.
5. The regression weight estimate,. 129, has a standard error of about .030.
6. The regression weight estimate,. 197, has a standard error of about .029.
7. The regression weight estimate, .699, has a standard error of about .033.
242
c. Critical ratio for regression weight
1. Dividing the regression weight estimate by the estimate of its standard error gives
z = .102/.029 = 3.573.
In other words, the regression weight estimate is 3.573 standard errors above zero
2. Dividing the regression weight estimate by the estimate of its standard error gives
z = .1467.027 = 5.456.
In other words, the regression weight estimate is 5.456 standard errors above zero
3. Dividing the regression weight estimate by the estimate of its standard error gives
z = . 1237.029 = 4.320.
In other words, the regression weight estimate is 4.32 standard errors above zero.
4. Dividing the regression weight estimate by the estimate of its standard error gives
z = .2957.028 = 10.513.
In other words, the regression weight estimate is 10.513 standard errors above zero
5. Dividing the regression weight estimate by the estimate of its standard error gives
z = . 1297.030 = 4.252.
In other words, the regression weight estimate is 4.252 standard errors above zero
6. Dividing the regression weight estimate by the estimate of its standard error gives
z = . 1977.029 = 6.731.
In other words, the regression weight estimate is 6.731 standard errors above zero
7. Dividing the regression weight estimate by the estimate of its standard error gives
z = .6997.033 = 21.510.
In other words, the regression weight estimate is 21.51 standard errors above zero
less than .001. In other words, the regression weight for Perception of Training
243
in the prediction of Training is significantly different from zero at the .001 level
(two-tailed).
less than .001. In other words, the regression weight for Effects & Beneflts in
the prediction of Training is significantly different from zero at the .001 level
(two-tailed).
3. The probability of getting a critical ratio as large as 4.32 in absolute value is less
than .001. In other words, the regression weight for Attitude towards Training
in the prediction of Training is significantly different from zero at the .001 level
(two-tailed).
less than .001. In other words, the regression weight for Training Inputs in the
prediction of Training is significantly different from zero at the .001 level (two-
tailed).
less than .001. In other words, the regression weight for Improvement in
less than .001. In other words, the regression weight for Selection to undergo
less than .001. In other words, the regression weight for Training in the
prediction of Growth and Result is significantly different from zero at the .001
level (two-tailed).
244
II. Standardized Regression Weights:
Estimate
Training < Perception of Training .108
Training < Effects & benefits .140
Training < Attitude towards Training .128
Training < Training Inputs .310
Training < Improvement in Relationships .148
Training < Selection to undergo training .251
Growth and result < Training .648
245
Covariances:
Effects & benefits <> Perception of training .165 .016 10.365 ***
Improvement in ***
Effects & Benefits <> .153 .017 9.050
Relationships
Selection to undergo ***
Effects & Benefits <> .202 .019 10.493
Training
Effects & Benefits <> Training Inputs .134 .015 8.707 **
Improvement in ***
<--> Training Inputs .217 .019 11.289
Relationships
a. Estimate of covariance
estimated to be .165.
2. The covariance between Effects & Benefits and Attitude towards Training is
estimated to be .125.
246
4. The covariance between Improvement in Relationships and Selection to
is estimated to be .236.
estimated to be .249.
11. The covariance between Effects & Benefits and Selection to undergo Training
is estimated to be .202.
12. The covariance between Effects & Benefits and Training Inputs is estimated
to be. 134.
13. The covariance between Attitude towards Training and Selection to undergo
14. The covariance between Attitude towards Training and Training Inputs is
estimated to be .135.
is estimated to be .217.
247
b. Standard error of covariance
10. The covariance estimate .153, has a standard error of about .017. '
11. The covariance estimate .202, has a standard error of about .019.
12. The covariance estimate .134, has a standard error of about .015.
13. The covariance estimate .312, has a standard error of about .023.
14. The covariance estimate .135, has a standard error of about .017.
15. The covariance estimate .217, has a standard error of about .019.
Dividing the covariance estimate by the estimate of its standard error gives
z = .125/.015 = 8.251.
In other words, the covariance estimate is 8.251 standard errors above zero.
Dividing the covariance estimate by the estimate of its standard error gives
z = .1657.016=10.365.
In other words, the covariance estimate is 10.365 standard errors above zero.
Dividing the covariance estimate by the estimate of its standard error gives
z = .2547.020= 12.799.
In other words, the covariance estimate is 12.799 standard errors above zero.
248
4. Dividing the covariance estimate by the estimate of its standard error gives
z = .400/.026= 15.225.
In other words, the covariance estimate is 15.225 standard errors above zero.
5. Dividing the covariance estimate by the estimate of its standard error gives
z=.236/.021 = 11.054.
In other words, the covariance estimate is 11.054 standard errors above zero.
6. Dividing the covariance estimate by the estimate of its standard error gives
z = .175/.017= 10.156.
In other words, the covariance estimate is 10.156 standard errors above zero.
7. Dividing the covariance estimate by the estimate of its standard error gives
z = .2097.019=10.916.
In other words, the covariance estimate is 10.916 standard errors above zero.
8. Dividing the covariance estimate by the estimate of its standard error gives
z = .2227.021 = 10.455.
In other words, the covariance estimate is 10.455 standard errors above zero.
9. Dividing the covariance estimate by the estimate of its standard error gives
z = .2497.019 =13.340.
In other words, the covariance estimate is 13.34 standard errors above zero.
10. Dividing the covariance estimate by the estimate of its standard error gives
z = .1537.017 = 9.050.
In other words, the covariance estimate is 9.05 standard errors above zero.
11. Dividing the covariance estimate by the estimate of its standard error gives
z = .2027.019 =10.493.
In other words, the covariance estimate is 10.493 standard errors above zero.
249
12. Dividing the covariance estimate by the estimate of its standard error gives
z=.134/.015 = 8.707.
In other words, the covariance estimate is 8.707 standard errors above zero.
13. Dividing the covariance estimate by the estimate of its standard error gives
z = .312/.023 = 13.739.
In other words, the covariance estimate is 13.739 standard errors above zero.
14. Dividing the covariance estimate by the estimate of its standard error gives
z=.1357.017 = 8.169.
In other words, the covariance estimate is 8.169 standard errors above zero.
15. Dividing the covariance estimate by the estimate of its standard error gives
z = .217/.019=11.289.
In other words, the covariance estimate is 11.289 standard errors above zero.
less than .001. In other words, the covariance between Effects & Benefits and
(two-tailed).
less than .001. In other words, the covariance between Effects & Benefits and
Attitude towards Training is significantly different from zero at the .001 level
(two-tailed).
less than .001. In other words, the covariance between Attitude towards
250
4. The probability of getting a critical ratio as large as 15.225 in absolute value is
less than .001. In other words, the covariance between Selection to Undergo
Training and Training Inputs is significantly different from zero at the .001
level (two-tailed).
less than .001. In other words, the covariance between Perception of Training
and Attitude towards Training is significantly different from zero at the .001
level (two-tailed).
less than .001. In other words, the covariance between Perception of Training
less than .001. In other words, the covariance between Perception of Training
less than .001. In other words, the covariance between Perception of Training
and Training Inputs is significantly different from zero at the .001 level (two-
tailed).
10. The probability of getting a critical ratio as large as 9.05 in absolute value is less
than .001. In other words, the covariance between Effects & Benefits and
251
Improvement in Relationships is significantly different from zero at the .001
level (two-tailed).
11. The probability of getting a critical ratio as large as 10.493 in absolute value is
less than .001. In other words, the covariance between Effects & Benefits and
level (two-tailed).
12. The probability of getting a critical ratio as large as 8.707 in absolute value is
less than .001. In other words, the covariance between Effects & Benefits and
Training Inputs is significantly different from zero at the .001 level (two-
tailed).
13. The probability of getting a critical ratio as large as 13.739 in absolute value is
less than .001. In other words, the covariance between Attitude Towards
14. The probability of getting a critical ratio as large as 8.169 in absolute value is
less than .001. In other words, the covariance between Attitude towards
Training and Training Inputs is significantly different from zero at the .001
level (two-tailed).
15. The probability of getting a critical ratio as large as 11.289 in absolute value is
252
Correlations:
Hypotheses Statements Estimate
Effects & Benefits <> Perception of Training .450
Effects & Benefits <> Attitude towards Training .345
Attitude towards Training <> Improvement in Relationships .587
Improvement in Relationships <> Selection to undergo Training .754
Selection to undergo Training <> Training Inputs .486
Perception of Training <> Attitude towards Training .439
Perception of Training <-> Improvement in Relationships .479
Perception of Training <> Selection to undergo Training .454
Perception of Training <> Training Inputs .621
Effects & Benefits <> Improvement in Relationships .383
Effects & Benefits <> Selection to undergo Training .456
Effects & Benefits <-> Training Inputs .367
Attitude towards Training <> Selection to undergo Training .647
Attitude towards Training <> Training Inputs .342
Improvement in Relationships <--> Training Inputs .499
Estimate of correlation
1. .450 is the estimated correlation between Effects & Benefits and Perception of
Training.
2. .345 is the estimated correlation between Effects & Benefits and Attitude
towards Training.
Improvement in Relationships.
Training Inputs.
towards Training.
253
7. .479 is the estimated correlation between Perception of Training and
Improvement in Relationships.
to undergo Training.
Inputs.
10. .383 is the estimated correlation between Effects & Benefits and Improvement
in Relationships.
11. .456 is the estimated correlation between Effects & Benefits and Selection to
undergo Training.
12. .367 is the estimated correlation between Effects & Benefits and Training
Inputs
13. .647 is the estimated correlation between Attitude towards Training and
14. .342 is the estimated correlation between Attitude towards Training and
Training Inputs.
Training Inputs.
254
a. Estimate of variance
1. The variance of Effects & Benefits is estimated to be .333.
2. The variance of Perception of Training is estimated to be .403.
3. The variance of Attitude towards Training is estimated to be .393.
4. The variance of Improvement in Relationships is estimated to be .475.
5. The variance of Selection to undergo Training is estimated to be .591
6. The variance of Training Inputs is estimated to be .398.
1. Dividing the variance estimate by the estimate of its standard error gives
z=.3337.019= 17.875.
In other words, the variance estimate is 17.875 standard errors above zero.
2. Dividing the variance estimate by the estimate of its standard error gives
z = .403/.023 = 17.875.
In other words, the variance estimate is 17.875 standard errors above zero.
3. Dividing the variance estimate by the estimate of its standard error gives
z = .393/.022= 17.875.
In other words, the variance estimate is 17.875 standard errors above zero.
255
4. Dividing the variance estimate by the estimate of its standard error gives
z=.475/.027= 17.875.
In other words, the variance estimate is 17.875 standard errors above zero.
5. Dividing the variance estimate by the estimate of its standard error gives
z = .591/.033 = 17.875.
In other words, the variance estimate is 17.875 standard errors above zero.
6. Dividing the variance estimate by the estimate of its standard error gives
z = .398/.022= 17.875.
In other words, the variance estimate is 17.875 standard errors above zero.
Training
Perception of Training
Improvement in Relationships
Training Inputs
e2
el
256
Variable counts
RMR, GFI
RMR, the root mean square residual, is an index of the amount by which the
estimated (model) variances and covariances differ from the observed variances and
GFI, the goodness of fit index, tells the proportion of the variance in the sample
variance-covariance matrix is accounted for by the model. This should exceed .9 for a
good model. For the full model it will be a perfect 1. Here in the model it is .984.
AGFI (adjusted GFI) is an alternate GFI index in which the value of the index is
adjusted for the number of parameters in the model. The fewer the number of
parameters in the model relative to the number of data points (variances and covariances
in the sample variance-covariance matrix), the closer the AGFI will be to the GFI. The
PGFI (P is for parsimony), the index is adjusted to reward simple models and penalize
models in which few paths have been deleted. Note that for our data the PGFI is larger
257
Baseline Comparisons
Goodness of fit indices compare model to the independence model rather than to
the saturated model. The Normed Fit Index (NFI) is simply the difference between the
two models' chi-squares divided by the chi-square for the independence model. For our
data, that is (.337 -. 98.411)/ 1441.725= .932. Values of .9 or higher (some say .95 or
higher) indicate good fit. The Comparative Fit Index (CFI) uses a similar approach
(with a noncentral chi-square) and is said to be a good index for use even with small
samples. It ranges from 0 to 1, like the NFI, and .95 (or .9 or higher) indicates good fit.
RMSEA
The Root Mean Square Error of Approximation (RMSEA) estimates lack of fit
compared to the saturated model. RMSEA of .05 or less indicates good fit, and .08 or
less adequate fit. LO 90 and HI 90 are the lower and upper ends of a 90% confidence
interval on this esfimate. PCLOSE is the p value testing the null that RMSEA is no
258
MODELING THE PERCEPTION OF TRAINING AND
GOAL ATTAINMENTS
.40
e1
.47 e2
e3 .79
.53 e4
e6
.32
.32
.49
e7
1
Perception of Training
259
Table -89
Regression Weights:
Individual goal attainment < Perception of Training .461 .041 11.131 ***
The above table explains the perception of training and goal attainment. From
the above table it is inferred that if training inputs changes up by one standard deviation,
will be .318 units of changes in effects and benefits of training. Similarly if the effects
and benefits go up by one standard deviation, there will be .494 units of changes in
perception of training.
.522 units of changes in group goal attainment, similarly there will be .461 units of
changes in individual goal attainment, like wise there will be -.124 units of changes in
260
CLUSTER DISTRIBUTION
seeks to identify a set of groups which both minimize within-group variation and
Two-step cluster analysis is often preferred for large datasets, since hierarchical
and k-means clustering do not scale efficiently when n is very large. The two-step
by identifying pre-clusters in a first step, then treating these as single cases in a second
step which uses hierarchical clustering. The two-step method is also the one chosen
when categorical variables with three or more levels are involved. The researcher can let
the two-step algorithm determine the number of clusters, or the researcher may set the
number of clusters.
Table - 90
N % of Combined % of Total
2 28 4.4% 4.4%
Cluster
3 268 41.9% 41.9%
From the above table it is inferred that the total number of respondents (640)
were group in three clusters. In cluster I, 344 respondents were grouped, cluster II, 28
261
Training
The clusters were group on the perception level of training. Cluster I implies the
respondents.
Selection
kiprovement in Relations
Growth &Resut
10 15 20 25
The above table explains the perception of highly satisfied respondents towards
training. According to this group of respondents are highly satisfied towards the
selection for training programme. Due to training programme they attain improvement
in relations. These cluster groups of respondents attain growth and result. The
perceptions of training were found to be good. These groups of respondents were not
262
PERCEPTION OF SATISFIED RESPONDENTS TOWARDS TRAINING
(nprovement in Relations . . . . -. -. 1
Selection ey>.v**MMM.: ^ ^ ^ ^ ^ ^ ^ H
Attitude ..-::lW;|;Wiy;;;;.>x:';;:;:x.:-:-:->:;>:J
FTobtem Faced I
Fterception gjSjggSj^
0 1 2 3 4 5 6 7 8
The above table shows the perception of satisfied respondents towards training.
These groups of respondents attain more growth and result due to the impact of training.
As this group attained more growth and result, it impacted in effects and benefits of
individuals and improvement in relation. The perception level towards training found to
Irrprovement in Retations /
AtHude
Selectnn 1
( 5 10 15 20 25
The above table explains the resuhs of training rated moderate and dis-satisfied.
These groups of respondents have a negative perception towards training. The members
failed to interact and develop their relations among the respondents. Comparatively
263
ANOVA FOR CLUSTERED GROUP
Hypotheses:
Training
Sum of
df Mean Square F Sig.
Squares
Between Groups 215.148 2 107.574
The ANOVA table, as displayed in the above table shows the F ratio 4205.081.
The statistical significance is .000 - the "Sig". Hence we can accept the hypotheses;
264