Data type: Discrete (Categorical) Variable and Ordinal for all DV’s and IDV’s.

** For this research purpose, although the data is ordinal the variables are treated as continuous data
(interval) for simplification purpose. However, certain steps are taken as precaution to address this
situation, such as:

 Likert type scale with small to high range is used (with clear distinction), all items in the Google
Form is independent.
 Non-Parametric test are used for correlation test to back up the Parametric test.
 Once parametric procedures are used, the alpha α value is set to 0.01 for correlation test.
 For multiple regression both ANOVA (parametric) and Kruskal Wallis (non-parametric) are used
to determine the significance of the independent variables.

https://www.theanalysisfactor.com/can-likert-scale-data-ever-be-continuous/

https://www.theanalysisfactor.com/when-dependent-variables-are-not-fit-for-glm-now-what/

For your own reference pa….

Reliability Test
The Cronbach alpha basically measures the internal consistency of the set of items created for the
research. The total number of questions developed for this research questionnaire is 33 which includes 28
variables (7 dependent and 21 independent variable) which used the Likert scale. The remaining 5
questions are based on demographics which explains N of item = 28 (Refer Table 2). Since, the Cronbach
alpha value obtained is 0.898, it can be deduced that the internal consistency is good.
https://www.statisticshowto.datasciencecentral.com/cronbachs-alpha-spss/
A total of 112 samples are used to aid this research. The Case Processing Summary indicates that all 112
samples are used to contribute towards the reliability statistics (no samples are omitted for the data
analysis).

Table 1 Case Processing Summary

Case Processing Summary

N %
Cases Valid 112 100.0
Excludeda 0 .0
Total 112 100.0
a. Listwise deletion based on all variables in the
procedure.

Table 2 Reliability Statistics for 28 variables

Reliability Statistics
Cronbach's
Alpha Based on
Cronbach's Standardized
Alpha Items N of Items
.898 .908 28

However, according to the Item-Total Statistics, the Cronbach alpha value can be increased to 0.905 and
0.901 if item DV5 and IDV1C are deleted respectively (refer Table 3). Hence, an independent
correlation study is done to examine how do these two variables respond against other DV’s and
IDV’s.
Table 3 Possibilities to increase Cronbach alpha value

Item-Total Statistics
Squared Cronbach's
Scale Mean if Scale Variance Corrected Item- Multiple Alpha if Item
Item Deleted if Item Deleted Total Correlation Correlation Deleted
DV5. In Malaysia every effort 107.63 164.937 .102 .455 .905
is taken by the Ministry of
Higher Education (MOHE), to
improve and raise the
education system towards
producing competent
manpower.
IDV1C. The Ministry of 106.95 162.772 .204 .268 .901
Higher Education alleges that
the Institutions of Higher
Learning do not have
relevant conceptualized
curriculum for 21st century
employability skills.

Research Objective (RO) No.3 – Pedagogy Effectiveness in

HE. DISCUSS
DV5. In Malaysia every effort is taken by the Ministry of Higher Education (MOHE), to
improve and raise the education system towards producing competent manpower.
IDV1C. The Ministry of Higher Education alleges that the Institutions of Higher Learning do not
have relevant conceptualized curriculum for 21st century employability skills.
Correlation
The correlation coefficient test is carried out to establish the strength of the linear relationship
among the variables involved in the test in order to establish if the correlation test from the
statistical sample does exist in the population.

Table 4 Correlation Matrix of DV5 and IDV1C against other variables

Correlation Matrix
DV5. IDV1C.
Pearson Correlation DV1. -0.032 0.077
DV2. -0.153 0.093
DV3. 0.079 0.134
DV4. -0.200 0.187
DV5. 1.000 0.017
DV6. 0.059 0.041
DV7. -0.234 0.217
IDV1A. -0.177 0.113
IDV1B. 0.008 0.212
IDV1C. 0.017 1.000
IDV1D. 0.106 0.097
IDV1E. 0.001 0.125
IDV1F. 0.035 0.151
IDV1G. -0.043 0.163
IDV2A. -0.016 0.093
IDV2B. 0.120 -0.118
IDV2C. 0.071 0.057
IDV2D. 0.018 0.166
IDV2E. -0.170 0.064
IDV2F. 0.020 0.110
IDV2G. 0.041 0.193
IDV3A. 0.463 0.091
IDV3B. 0.428 0.189
IDV3C. 0.350 -0.065
IDV3D. 0.369 0.093
IDV3E. 0.000 0.273
IDV3F. 0.047 0.151
IDV3G. 0.078 0.091

As discussed in the previous page, the reliability test could have been improved if DV5 and
IDV1C is omitted from the questionnaire. Hence, the correlation is cross checked to understand
the justification for the increase in Cronbach alpha value. It can be observed that both DV5 and
IDV1C has a relatively low correlation against other variables. There is also negative correlation
that can be observed from the table in the previous page. The significance of these two variables
are also proved to be insignificant with a big ratio when compared against 95% confidence
interval (Refer to Appendix 1). Hence, it is unbiassed to say that the Cronbach alpha will
definitely increase if variable DV5 and IDV1C is omitted due to its weak and negative
correlation (in some cases) with other variable.

Now, the correlation between DV and other IDV’s (computed as single variable) are scrutinized
as an identity matrix to further study the correlation.

Correlations between DV and IDV1

DV IDV1
DV Pearson Correlation 1 .626**

Sig. (2-tailed) .000

N 112 112
IDV1 Pearson Correlation .626** 1

Sig. (2-tailed) .000

N 112 112
**. Correlation is significant at the 0.01 level (2-tailed).

Interpretation
First, the above output result of the bivariate, depicts a correlation coefficient value of .626
indicating a credible strength positive correlation because it is a positive value approximating
towards positive 1 which means perfect positive correlation. Second, the correlation significance
output value of .000 indicates that statistically the correlation is significant. If the significance
value had been greater than the chosen level of significance; the p value which is 0.01 then
statistically the correlation is insignificant. Since the output results indicate the significance
value is lesser than the p value, then it can be inferred that there is enough evidence to suggest
that the correlation observed does exist in the population.

Correlations between DV and IDV2

Correlations
DV IDV2
DV Pearson Correlation 1 .626**

Sig. (2-tailed) .000

N 112 112
IDV2 Pearson Correlation .626** 1

Sig. (2-tailed) .000

N 112 112
**. Correlation is significant at the 0.01 level (2-tailed).

Interpretation
First, the above output result of the bivariate, depicts a correlation coefficient value of .626
indicating a credible strength positive correlation because it is a positive value approximating
towards positive 1 which means perfect positive correlation. Second, the correlation significance
output value of .000 indicates that statistically the correlation is significant. If the significance
value had been greater than the chosen level of significance; the p value which is 0.01 then
statistically the correlation is insignificant. Since the output results indicate the significance
value is lesser than the p value, then it can be inferred that there is enough evidence to suggest
that the correlation observed does exist in the population.
Correlations between DV and IDV3

Correlations
DV IDV3
DV Pearson Correlation 1 .627**

Sig. (2-tailed) .000

N 112 112
IDV3 Pearson Correlation .627** 1

Sig. (2-tailed) .000

N 112 112
**. Correlation is significant at the 0.01 level (2-tailed).

Interpretation
First, the above output result of the bivariate, depicts a correlation coefficient value of .627
indicating a credible strength positive correlation because it is a positive value approximating
towards positive 1 which means perfect positive correlation. Second, the correlation significance
output value of .000 indicates that statistically the correlation is significant. If the significance
value had been greater than the chosen level of significance; the p value which is 0.01 then
statistically the correlation is insignificant. Since the output results indicate the significance
value is lesser than the p value, then it can be inferred that there is enough evidence to suggest
that the correlation observed does exist in the population.
Summary
Correlations
DV IDV1 IDV2 IDV3
DV Pearson 1 .626** .626** .627**
Correlation
Sig. (2-tailed) .000 .000 .000
N 112 112 112 112
**
IDV1 Pearson .626 1 1.000** .815**
Correlation
Sig. (2-tailed) .000 .000 .000
N 112 112 112 112
** **
IDV2 Pearson .626 1.000 1 .815**
Correlation
Sig. (2-tailed) .000 .000 .000
N 112 112 112 112
IDV3 Pearson .627** .815** .815** 1
Correlation
Sig. (2-tailed) .000 .000 .000
N 112 112 112 112
**. Correlation is significant at the 0.01 level (2-tailed).
Interpretation
The correlation output result of the multivariate (DV with IDV1, IDV2 and IDV3) depicts a
mean correlation coefficient value of .627 indicating credible strength positive correlation
because it is a positive value approximating towards positive 1 which means perfect positive
correlation. However, the correlation output result between IDV1 and IDV2 depicts a value of
1.000 which is a perfect positive correlation. Thus, the variance between these two independent
variables will further analyzed to identify the uniqueness using the regression model. Next, the
correlation output result between IDV1 and IDV3 depicts a value of .815 which is a strong
positive correlation. Lastly, the correlation output result between IDV2 and IDV3 depicts a value
of .815 which is a strong positive correlation.

The correlation significance output value of the multivariate (DV with IDV1, IDV2 and IDV3)
depicts a mean significant value of .000 indicating that statistically the correlation among the
multivariate is significant. If the significance value had been greater than the chosen level of
significance; the p value which is 0.01 then statistically the correlation is insignificant. Since the
output results indicate the significance value is lesser than the p value, then it can be inferred that
there is enough evidence to suggest that the multivariate correlation observed does exist in the
population.

Regression

Linear regression is the next step up after correlation. It is used when we want to predict the
value of a variable based on the value of another variable. The variable we want to predict is
called the dependent variable (or sometimes, the outcome variable)

In multiple regression we have set one (1) criterion (dependent variable, DV) and three (3)
predictors (independent variables, IDV’s). Hence, the regression model basically gives us an
insight of two distinct features:

 An analysis that investigates if the independent variables performed well enough to

predict the outcome which is the criterion (DV). (Helps to determine the strength of the
impact)
 An analysis that enables us to converge our focus to specific variables which are
significant predictors to the outcome. (helps us to determine the significance of the
relationship)

Thus, the regression model helps us to develop and explain the relationship between a dependent
variable along with several independent variable.

The regression model was generated for three different cases stated as below:

1. DV and IDV1
2. DV and IDV2
3. DV and IDV3
Case 1 - DV and IDV1

Model Summaryb

Std. Error of the

Model R R Square Adjusted R Square Estimate
1 .626a .391 .350 .429

a. Predictors: (Constant), IDV1G. The 21st century curriculum should be connected

with the real world to encourage student participation, learning and prepare them for
adult life., IDV1C. The Ministry of Higher Education alleges that the Institutions of
Higher Learning do not have relevant conceptualized curriculum for 21st century
employability skills., IDV1D. Academia recognizes curriculum as a determinant of the
quality of education and learning., IDV1B.Curricula determine the fate of individuals,
communities, countries and the world by determining what, why, when and how people
learn., IDV1A.The most critical factor affecting the academia in the 21st century is the
designing of appropriate curriculum which can satisfy the needs of an agile
employment industry., IDV1F. Malaysian Institutions of Higher Learning should adhere
to global normative guide produced by International Bureau of Education (IBE-
UNESCO), for competence based curricula that can help realize the education agenda
far into the future up to 2030., IDV1E. A well-designed Higher education curriculum
should include creativity, innovation, leadership and entrepreneurship to mitigate the
impacts of globalization.

b. Dependent Variable: DV
 ANOVAa
Model Sum of Squares df Mean Square F Sig.
1 Regression 12.285 7 1.755 9.552 .000b
Residual 19.108 104 .184
Total 31.393 111
a. Dependent Variable: DV
b. Predictors: (Constant), IDV1G. The 21st century curriculum should be connected with the real
world to encourage student participation, learning and prepare them for adult life., IDV1C. The
Ministry of Higher Education alleges that the Institutions of Higher Learning do not have relevant
conceptualized curriculum for 21st century employability skills., IDV1D. Academia recognizes
curriculum as a determinant of the quality of education and learning., IDV1B.Curricula determine
the fate of individuals, communities, countries and the world by determining what, why, when and
how people learn., IDV1A.The most critical factor affecting the academia in the 21st century is the
designing of appropriate curriculum which can satisfy the needs of an agile employment industry.,
IDV1F. Malaysian Institutions of Higher Learning should adhere to global normative guide
produced by International Bureau of Education (IBE-UNESCO), for competence based curricula
that can help realize the education agenda far into the future up to 2030., IDV1E. A well-designed
Higher education curriculum should include creativity, innovation, leadership and entrepreneurship
to mitigate the impacts of globalization.

Observation: F (7,104 = 9.552), p < 0.05, r2 = 0.40

Interpretation: Based on the obtained results, the overall regression model of the IDV1
predicting that the outcome for the DV accounts for 40% of the variance. Hence, this 40%
basically explains the percentage of variable variation that is fitted into a linear model. Since the
r2 > 0 and the significant, p < 0.05 then this model is a well fitted for this research. Although the
r2 is relatively low when compared to 0 to 1 scale, the ANOVA test includes a coefficient table
where significant predictors can be spotted and crucial conclusions can be made.
Therefore, the F-test and the r2 gives us the significance of the data provided and the strength of
the relationship between the generated model and the predictors respectively. The correlation
and coefficient table will further break down the uniqueness of the IDV’s.
The normality assumptions are also met for N= 112 as the plotted point follows the straight line
(Refer Figure 1). The normality assumption is deemed to fit for the next 2 cases as well as the
sample size is ≥ 25.

Figure 1 Normal P-P plot for Normality

Table 5 Coefficient Table for IDV1

Coefficientsa
Standardized
Unstandardized Coefficients Coefficients
Model B Std. Error Beta t Sig.
1 (Constant) 1.489 .329 4.533 .000
IDV1A. .119 .061 .179 1.940 .055
IDV1B. .190 .058 .300 3.293 .001
IDV1C. .038 .039 .077 .979 .330
IDV1D. .129 .051 .226 2.552 .012
IDV1E. .098 .084 .132 1.156 .250
IDV1F. -.128 .072 -.199 -1.786 .077
IDV1G. .129 .104 .153 1.246 .215
a. Dependent Variable: DV

The coefficient table indicates the significance of each independent variable that contributes towards
the outcome of DV. For case in, it can be observed that IDV1A and IDV1D are significant for p < 0.05.
Therefore, these two predictors are bound to have unique variance that could strongly influence the
criterion.

Case 2 - DV and IDV2

Model Summaryb
Adjusted R Std. Error of the
Model R R Square Square Estimate
1 .588a .346 .302 .444
a. Predictors: (Constant), IDV2G., IDV2B. , IDV2A. , IDV2F. , IDV2E. ,
IDV2C. , IDV2D.
b. Dependent Variable: DV

ANOVAa
Model Sum of Squares df Mean Square F Sig.
1 Regression 10.855 7 1.551 7.853 .000b
Residual 20.538 104 .197
Total 31.393 111
a. Dependent Variable: DV
b. Predictors: (Constant), IDV2G. , IDV2B. , IDV2A. , IDV2F. , IDV2E. , IDV2C. , IDV2D.

Observation: F (7,104 = 7.853), p < 0.05, r2 = 0.35

Interpretation: Based on the obtained results, the overall regression model of the IDV1
predicting that the outcome for the DV accounts for 35% of the variance. Since the r2 > 0 and the
significant, p < 0.05 then this model is deemed fit for this research.

Coefficientsa
Standardized
Unstandardized Coefficients Coefficients
Model B Std. Error Beta t Sig.
1 (Constant) 2.027 .288 7.047 .000
IDV2A. .059 .077 .083 .771 .442
IDV2B. -.019 .054 -.034 -.345 .731
 IDV2C. .059 .073 .088 .806 .422
IDV2D. .159 .068 .261 2.330 .022
IDV2E. .053 .072 .080 .738 .462
IDV2F. .194 .077 .287 2.506 .014
IDV2G. -.044 .058 -.078 -.749 .456
a. Dependent Variable: DV

For this case, by looking at each predictor individually, IDV2D and IDV2F seems to be
significant and hence could be a great variable to predict the dependent variables whereas other
IDV2’s is cohesively strongly related. Hence, the unique variance of IDV2D and IDV2F has
definitely contributed towards the outcome of the dependent variable.

Henceforth, the correlation matrix which yielded a strong positive correlation of 1.000 between
IDV1 and IDV2 can be unheeded because the regression model of IDV1 and IDV2 has shown
that there is some unique variance that distinctly separates the 2 sets.

Case 3 - DV and IDV3

Model Summaryb
Adjusted R Std. Error of the
Model R R Square Square Estimate
1 .622a .387 .346 .430
a. Predictors: (Constant), IDV3G., IDV3F. , IDV3D., IDV3E., IDV3C. ,
IDV3B. , IDV3A.
b. Dependent Variable: DV

ANOVAa
Model Sum of Squares df Mean Square F Sig.
1 Regression 12.146 7 1.735 9.376 .000b
Residual 19.247 104 .185
Total 31.393 111
a. Dependent Variable: DV
b. Predictors: (Constant), IDV3G., IDV3F. , IDV3D., IDV3E., IDV3C. , IDV3B. , IDV3A.

Observation: F (7,104 = 9.378), p < 0.05, r2 = 0.39

Interpretation: Based on the obtained results, the overall regression model of the IDV3
predicting that the outcome for the DV accounts for 39% of the variance. Since the r2 > 0 and the
significant, p < 0.05 then this model is a well fitted for this research.

Coefficientsa
Standardized
Unstandardized Coefficients Coefficients
Model B Std. Error Beta t Sig.
1 (Constant) 1.798 .282 6.384 .000
IDV3A. -.029 .055 -.060 -.525 .601
IDV3B. .017 .063 .029 .264 .792
IDV3C. .078 .054 .147 1.431 .156
IDV3D. -.042 .063 -.072 -.666 .507
IDV3E. .145 .053 .234 2.728 .007
IDV3F. .220 .058 .338 3.795 .000
IDV3G. .131 .052 .233 2.506 .014
a. Dependent Variable: DV

Based on the coefficient table above, it can be observed that, IDV3E, IDV3F, and IDV3G are
unique as they are significant, p < 0.05. So, these IDV’s has the potential to determine the
outcome of the DV is stronger way compared to other IDV’s in this set.