Home / Training Quality

Regression Equation of Learner Rating Scale


2. Factor 2 (named Lecturer - Teacher) includes the following variables: NL1 Lecturer has solid professional knowledge NL3 Lecturer has good teaching skills

NL2 The lecturer has good teaching methods HH4 The lecturer's demeanor is very standard

DU4 Students' requests are always responded to promptly by lecturers.


Maybe you are interested!

3. Factor 3 (named School - NT) includes the following variables:


Regression Equation of Learner Rating Scale

TC2 Necessary information is always delivered to students accurately TC4 Staff always accurately recognizes student requests TC3 Necessary information is always delivered to students promptly TC5 Staff completes work on time

TC1 The school fulfills all its commitments to students.


4. Factor 4 (named Facilities - VC) includes the following variables: HH2 Practical classrooms have modern equipment

HH1 Spacious theory classroom


HH3 The practice room is fully equipped for students to practice. The scale model is reconstructed as Diagram 2-3:

Figure 2-3: Model of learner evaluation scale after factor analysis


HL1

Satisfaction (HL)

HL2

HL3

Attention (QT)

DU2, DC2, DC1, DC4, DU1, DC3

Lecturer (GV) NL1, NL3, NL2, HH4, DU4

School (NT) TC2, TC4, TC3, TC5, TC1

Facilities (VC) HH2, HH1, HH3


Conducting Cronbach's Alpha analysis on the new scale to assess reliability, the results show that the variables in the scale all have a total item correlation ≥ 0.3 and Cronbach's Alpha of the scale = 0.815 ≥ 0.7, showing that the scale has usable reliability.

2.3.5.4. Regression equation of learner evaluation scale

Correlation test between pairs of variables shows that the pairs of variables are strongly correlated with each other so regression can be performed (Table PL-B-5, page 92).

Regression model (estimation) of learner rating scale:

𝐻𝐿 = 𝑏 0 + 𝑏 1 (𝑄𝑇) + 𝑏 2 (𝐺𝑉) + 𝑏 3 (𝑁𝑇) + 𝑏 4 (𝑉𝐶) (Equation ̀ nh 2-1)

In which: HL = Learner Satisfaction; QT = Care; GV = Teacher; NT = School; VC = Facilities.

Regression results are as follows (Table PL-B-6, page 93): R 2 adjusted = 0.465 shows that the four variables QT, GV, NT, VC explain 46.5% of the variance in student satisfaction. Sig. = 0.000 shows that the model fits the population with very high reliability.

With the regression coefficient (Table PL-B-7, page 93), all 4 variables have VIF < 10, so it is confirmed that there is no multicollinearity phenomenon. The three variables Interest (QT), Teacher (GV) and School (NT) have the largest Sig. of 0.017, so it is acceptable with a significance level of 0.017.

– ie 98.3% confidence level. The Facility (VC) variable has Sig. = 0.183 which is quite large (reliability only reaches 81.7%). However, according to Hoang Trong & Chu Nguyen Mong Ngoc (2008), this only shows that with the current sample data set and scale, there is no evidence that the VC variable is unrelated to satisfaction.

To check whether VC should be removed or not, we performed a test of removing VC in the regression model. The regression result after removing VC showed that R 2 adjusted = 0.464, lower than the model with VC (Table PL-B-8, page 93). In addition, in terms of content, the VC variable (facilities) is a meaningful variable in practice, so the original regression model with 4 variables VC, GV, NT and QT is still kept intact (using standardized regression coefficients):

Satisfaction (NH) = 0.309(𝑄𝑇) + 0.340(𝐺𝑉) + 0.107(𝑁𝑇) + 0.056(𝑉𝐶) (Equation ̀ nh 2-2)


Check the assumptions required for linear regression


1. Linear relationship between dependent variable and independent variables:


The scatter plot of the standardized residuals is randomly scattered in a region around the y-axis rather than following a rule so the linearity assumption is satisfied (Chart PL-B-1, page 94).

2. Constant variance:


Conduct Spearman rank correlation test to test the hypothesis H 0 is the Rank Correlation Coefficient of the population (TT) and the dependent variable (HL) = 0 (i.e. Constant Variance). With the significance level = 0.05, the result Sig. = 0.91 > 0.05 shows that the variance is constant (Table PL-B-9, page 95).

3. Autocorrelation:


The Durbin – Watson statistic of the model d = 1.836 (Table PL-B-6, page 93). With the number of independent variables being 4, the number of observations being 475 and the significance level = 5%, looking up the Durbin – Watson value table, we have d L = 1.8685 and d U = 1.83249, thus: d U = 1.83249 d = 1.836 4d U . According to Table 2-16: Durbin – Watson test empirical rules (page 46), we determine that there is no autocorrelation.

4. The residuals have a normal distribution:


Through the PP diagram of the standardized residuals (Chart PL-B-2, page 95), the standardized residuals are quite close to the expected normal distribution, thus confirming that the residuals are normally distributed.

Thus, the regression model of the Learner Assessment scale is accepted (with standardized regression coefficients) as follows (Equation 2-1):

Satisfaction (NH) = 0.309(𝑄𝑇) + 0.340(𝐺𝑉) + 0.107(𝑁𝑇) + 0.056(𝑉𝐶)


2.3.5.5. Identify strengths and weaknesses

The overall average estimate of the 4 factors affecting Learner Satisfaction (Table PL-B-10, page 95) shows that the average order of factors from low to high is QT, NT, VC, GV.

In the QT factor, the two variables DC1 “The school is very concerned about your living and studying conditions” and DU2 “Staff are always willing to help students” were the lowest although they were still rated above the neutral level (Table PL-B-11, page 96).

In the teacher factor, the two variables NL1 "Lecturers have solid professional knowledge" and NL2 "Lecturers have good teaching methods" are the highest and are both close to the "agree" level (Table PL-B-12, page 97).

Learner satisfaction is rated at a fairly high level (Mean=3.8028)


2.3.5.6. Identify differences in learner assessments

1. Differences between industry and occupation groups

Levene's test was performed to check the conditions for applying analysis of variance, with a significance level of 5%, the GV and HL components satisfied the conditions for applying ANOVA (Table PL-B-13, page 97). The ANOVA analysis results for the GV component had Sig. = 0.018 and HL had Sig. = 0.002, both were smaller than = 5%, so there was a difference and was analyzed in depth using Bonferroni statistics, the remaining components were analyzed using Tamhane's T2 statistics.

The results of the in-depth analysis are as follows:


QT component (Interest): there is a difference between the Receptionist profession and the remaining 5 professions (Restaurant Management, Hotel Management, Kitchen, Guide, Hotel and Tourism Management) and all are differences < 0.

GV (Teacher) composition: there is a difference (>0) between the profession of Guide and DLNHKS Management.

NT component (School): no difference between occupations.


VC (Materials) component: there is a difference (>0) between Restaurant Management and Reception, Restaurant Management and Hotel Management.

HL (Satisfaction): there is a difference (>0) between Guide and Receptionist, Guide and Hotel Management.

2. Differences between school years

Test the hypothesis by applying ANOVA analysis (Table PL-B-14, page 98): with a significance level of 5%, the QT and VC components satisfy the conditions for applying ANOVA. The ANOVA analysis result for the VC component has Sig. = 0.422 > significance level = 5%, therefore, there is no difference in the VC component. The QT component has Sig. = 0.000 less than = 5%, therefore, there is a difference and is analyzed in depth using Bonferroni statistics, the remaining components are analyzed using Tamhane's T2 statistics.

The results of the in-depth analysis are as follows:


Components of QT (Interest), GV (Teacher), NT (School), HL (Satisfaction): there is a difference >0 between year 1 and year 2, year 1 and year 3.

VC (Materials) component: no difference between years.


3. Difference between students who work part-time and those who do not

Testing the hypothesis by applying ANOVA analysis (Table PL-B-15, page 98): with a significance level of 5%, the components NT, VC and HL satisfy the conditions for applying ANOVA. The results of ANOVA analysis for all 3 components have Sig. > significance level = 5%, so there is no difference. The component QT has Sig. = 0.000 less than = 5%, so there is a difference and is analyzed in depth using Bonferroni statistics, the remaining components are analyzed using Tamhane's T2 statistics.

The results of the in-depth analysis are as follows:


QT component (Interest): there is a difference >0 between not working extra and working extra in the same profession as studying.


The components of GV (Teacher), NT (School), VC (Materials), HL (Satisfaction) are all the same.

4. Differences between training systems

Testing the hypothesis by applying ANOVA analysis (Table PL-B-16, page 98): with a significance level of 5%, the components GV, NT, VC, HL satisfy the conditions for applying ANOVA. The ANOVA analysis results for the VC and HL components have Sig. > significance level = 5%, so there is no difference. The GV and NT components have Sig. less than = 5%, so there is a difference and are analyzed in depth using Bonferroni statistics, the QT component is analyzed using Tamhane's T2 statistics.

The results of the in-depth analysis are as follows:


QT (Care), VC (Facilities) and HL (Satisfaction) components: no difference.

Teacher composition: there is a difference <0 between University and Secondary School, University and College.

NT component (School): there is a difference <0 between University and Secondary School.

2.3.6. Analysis of teacher evaluation

2.3.6.1. Satisfaction scale analysis

Conducting a reliability assessment of the scale shows that the scale has good reliability with all variables in the scale having a Variable-Total correlation of ≥ 0.3 and Cronbach's Alpha of the scale ≥ 0.7.

The results of the exploratory factor analysis of the Satisfaction scale showed that there was only one factor with all three initial observed variables (Table PL-B-17, Table PL-B-18, page 99).

2.3.6.2. Evaluation of the reliability of components

Evaluate the reliability of 5 components: Trust, Tangibility, Empathy, Responsiveness, Competence through Cronbach's Alpha analysis on SPSS software.


All five components have Cronbach's Alpha coefficient ≥ 0.6, indicating that the scale is reliable. The variables TC1 of the Reliability scale, HH1, HH2, HH5 of the Tangibility scale have a Variable - Total correlation < 0.3, so they were eliminated.

2.3.6.3. Exploratory factor analysis of teacher evaluation scale

Performing KMO & Bartlett's Test shows that the set of variables can be factor analyzed: KMO = 0.803 > 0.5 and Bartlett's Test has Sig. = 0.000 < 0.005 (Table PL-B-19, page 99).

Conduct factor analysis using Principal Component analysis combined with Varimax rotation. The result at eigenvalue = 1.049 extracted 5 factors with a total extracted variance of 74.354% (Table PL-B-20 page 99, Table PL-B-21 page 100). The learner's assessment scale now includes 5 new factors with a total of 16 independent variables as follows:

1. Factor 1 (Management organization = TCQL)


HH3 Convenient use of classrooms and equipment NL5 Reasonable content of subject programs

HH4 School grounds and classrooms are well kept NL1 Reasonable teaching plan

TC5 Staff completed work on time


DU2 Staff quickly carried out teacher requests


2. Factor 2 (Training management = Training management)


NL4 All subjects have clear programs


DC2 The school is flexible in arranging timetables NL2 The school's discipline is good

3. Factor 3 (Professional = CN)


TC2 Necessary information is always accurately transmitted to teachers


TC3 Necessary information is always delivered to teachers in a timely manner TC4 Staff always accurately recognize teachers' requests

4. Factor 4 (Interest = QT)


DC1 Convenient break room for teachers


DC3 The school pays salaries and bonuses to teachers on time.


5. Factor 5 (Employees = NV)


NL3 Staff with solid expertise


DU1 Staff are always willing to respond to teachers' requests. The scale model is reconstructed as follows:

Figure 2-4: Teacher evaluation scale model after factor analysis


Satisfaction

(HL)

HL1

HL2

Management organization (TCQL) HH3, NL5, HH4, NL1, TC5, DU2

Training Management (TM)

NL4, DC2, NL2

Professional (CN) TC2, TC3, TC4

Attention (QT) DC1, DC2

Staff (NV) NL3, DU1


Cronbach's Alpha analysis was performed on the new scale to assess reliability. The results showed that all variables in the scale had a total item correlation of ≥ 0.3 and Cronbach's Alpha of the scale of ≥ 0.7, indicating that the scale has usable reliability.

2.3.6.4. Regression equation of teacher evaluation scale

Correlation test between pairs of variables shows that the pairs of variables are strongly correlated with each other so regression can be performed (Table PL-B-22, page 100).

Comment


Agree Privacy Policy *