Home / Service

Research Model of the Relationship Between Technical Quality and Functional Quality with Ground Service Quality


This model consists of four latent variables, of which two latent variables are identified as the causes affecting the overall satisfaction level of ground service quality. These two causes are check-in quality (hlci) and boarding process quality. The latent variable or hlci factor is composed of two factors: check-in organization quality (tcci) and boarding process quality (ttmb). The outcome variable is the overall satisfaction level HLTT_1. The objective of the study is to provide the best measurement model, and then provide the best structural model of the factors affecting the overall satisfaction level of ground service.

3.3.1.2.2 Testing the validity of the measurement model

1

1

z1

hlxh

e11 e10 e9

e8

1

1

hclci

1

1

CAU67C_1

1

CAU67D_1

1

tdci

1

d2

ttmb

1

CAU67A_1

CAU67B_1

HLTTHE_1

Before calculating the model validity indices, the relationships between variables are represented as two-way arrows as follows:

e12

1

CAU64_1

1 e13

CAU66_1



CAU7A_1


1

Maybe you are interested!

Research Model of the Relationship Between Technical Quality and Functional Quality with Ground Service Quality

CAU7B_1


1

CAU7C_1


1

CAU7D_1


1

CAU7E_1


1


e1

e2

e3

e4

e5

Figure 3.2. Measurement model


The two-way arrows represent correlations between variables. In this model, causal relationships are not yet established.

The parameters measuring the validity of the model are as follows:


Table 3.17. RMR and GFI coefficients


Model

GFI

AGFI

PGFI

Default model

.926

.885

.594


Table 3.18. Baseline Comparisons


NFI

Delta1

RFI

rho1

IFI

Delta2

TLI

rho2


CFI

.967

.956

.968

.957

.968


Table 3.19. RMSEA coefficient



RMSEA

LO 90

HI 90

PCLOSE

.091

.087

.095

.000


Table 3.20. HOELTER coefficient



HOELTER

.05

HOELTER

.01

158

179


The values ​​of GFI and TLI are both approximately 1, indicating that the model has a very good fit. However, the RMSEA and Holter values ​​are not really good (greater than 0.8 and less than 200), so it is necessary to adjust the model to have higher fit indices. This is done by looking at the MI table. See details of the MI values ​​in table 3.21 in Appendix 2.


1

1

z1

hlxh

e11 e10 e9

e8

1

1

hclci

CAU67A_1

1

1

CAU67C_1

1

CAU67D_1

1

tdci

1

d2

ttmb

CAU67B_1

CLTTMD

The MI table shows that the residual e4 appears many times with a fairly large MI value. The residual e1 corresponding to variable 7e_1 also has a fairly high MI value. Therefore, removing variables 7b_1 and 7e_1 corresponding to these residuals will make the model better. The new model is as follows:

e12

1

CAU64_1

1 e13

CAU66_1



1

e3

CAU7C_1

CAU7A_1


1

CAU7D_1


1


e2

e5

Figure 3.3. Model A_1.2

The correlation coefficients between the three latent variables hlci; ttmb and hlltt are as follows:


Table 3.22. Correlations





Estimate

CLCI

<-->

CLTTMD

.784

ttmb

<-->

CLTTMD

.793

ttmb

<-->

Clci

.794


This new model has the following RMSEA and Holter:


Table 3.23. RMSEA coefficient


RMSEA

LO 90

HI 90

PCLOSE

.069

.064

.074

.000


Table 3.24. HOELTER coefficient


HOELTER

.05

HOELTER

.01

286

332

e12 1

CAU64_1


1

1

z1

hlxh

e11 CAU67A_1

1

hclci

1

e10

1

CAU67B_1

1

e9 CAU67C_1

1

tdci

1

e8

1

CAU67D_1

d2

ttmb

CLTTMD

RMSEA is below 0.7 and Holter is above 200. Thus the second model is much better than the first model. Therefore, a structural model can be built from the second model.


1 e13

CAU66_1



1

e3

CAU7C_1

CAU7A_1


1

CAU7D_1


1


e2

e5

Figure 3.4. A-1.3 relational structure model

This model represents a causal relationship, in which CLTTMD is the result of two causes, Clci and ttmb. The results of measuring the validity of this structural model are as follows:


Table 3.25. RMSEA coefficient

RMSEA

LO 90

HI 90

PCLOSE

.157

.152

.162

.000

Table 3.26. HOELTER coefficient

HOELTER

.05

HOELTER

.01

59

68

The two coefficients RMSEA and Holter show that the model does not have a high fit. Therefore, although the measurement model has a high fit, this structural model does not represent the relationship between the latent variables correctly. The model needs to be adjusted through the MI table. Details of the MI table are in table 3.27, appendix 2.

e12 1

CAU64_1


1

1

z1

CLXH

e11 CAU67A_1

e10 CAU67B_1

1

Clci

1

1

1

e9

1

tdci

1

d3

e8

1

d2

ttmb

CLTTMD

Observing the MI table shows that the two variables ttmb and Clci have very large MI. This allows us to think of the hypothesis that these two variables have a very close correlation with each other, and therefore it is necessary to redraw the model. It is possible that drawing the correlation between the two variables will make the model more suitable or that one of the two variables must be removed from the model.


1 e13

CAU66_1




CAU67C_1


CAU67D_1


1


e3

CAU7C_1

CAU7A_1


1

CAU7D_1


1


e2

e5

Figure 3.5. Model A_1.4

In this model, hlci influences the perception of the quality of the boarding procedures and the boarding procedures in turn influence the overall satisfaction with the ground services. The results of the validity test of the model are as follows:


Table 3.28. RMSEA coefficient

RMSEA

LO 90

HI 90

PCLOSE

.086

.081

.091

.000


Table 3.29. HOELTER coefficient

HOELTER

.05

HOELTER

.01

188

218

These two indexes have clearly improved in this model. In addition to the RMSEA coefficient, the GFI; TLI and CFI indexes are all very high, corresponding to 0.95; 0.963 and 0.974, respectively. It can be said that the model has a good fit and can be used to represent the relationship structure of the variables clci; ttmb and CLTTMD.

The regression coefficients between the variables are as follows:

Table 3.30. Regression coefficients





Estimate

Error (SE)

P

clxh

<---

Clci

.861

.027

***

tdci

<---

Clci

1,000



ttmb

<---

Clci

.871

.022

***

ttmb

<---

z15

.651

.017

***

CAU67D_

<---

Tdci

1,000



CAU67C_1

<---

Tdci

1,044

.013

***

CAU67A_

<---

Tdci

1,024

.011

***

CAU64_1

<---

Clxh

1,000



CAU66_1

<---

Clxh

1,039

.025

***

CLTTMD

<---

ttmb

.933

.016

***

CLTTMD

<---

z10

.747

.011

***

CAU67B_1

<---

e10

.397

.008

***

CAU67B_1

<---

Tdci

1,069

.011

***

CAU7D_1

<---

ttmb

1,000



CAU7A_1

<---

ttmb

.950

.017

***

CAU7C_1

<---

ttmb

1,009

.015

***

Ttmb

<---

Clci

.871

.022

***

CLTTMD

<---

ttmb

.933

.016

***


Thus, the factor that directly affects the overall quality is the quality of the boarding procedures, however, the perceived quality of the boarding procedures is affected by the perceived quality of the check-in procedures. The two factors that measure the quality of check-in are the attitude of the check-in staff (tdci) and the quality of the check-in organization (tcci), which includes waiting time and order in the queue area.

Thus, the structural model shows that the variable clci or the quality of the check-in process is the starting point for VNA in its efforts to improve the overall satisfaction with the quality of ground services. A 1-point increase in the quality of check-in will increase the quality of the boarding procedures by 0.871 points. However, because the correlation coefficient between clci and ttmb is 0.794, equivalent to a coefficient of determination of 0.63; meaning that the quality of check-in determines 63% of the variation in boarding procedures. The correlation coefficient between ttmb and CLTTMD is 0.793, also corresponding to a coefficient of determination of 63%. Therefore, in order to increase the overall satisfaction with ground services, VNA must also improve the quality of boarding procedures. For boarding procedures, two important metrics are the convenience of getting from the lounge to the plane and the enthusiasm of the staff in guiding the boarding procedures.

3.3.1.3 Research model of the relationship between technical quality and functional quality with ground service quality

The set of indicators measuring elements of ground quality can be combined in a different way to measure technical and functional quality. Specifically:

The technical quality latent variable is measured by three indicators i) question 64 Waiting time for check-in procedures; ii) question 66 Orderliness at the check-in area; iii) question 7d Convenience of going from the waiting room to the airport.

The skill quality variable includes the indicators i) question 67a; ii) question 67b; iii) question 67c; iv) question 67d; v) question 7a; vi) question 7b; vii) question 7c and viii) question 68.

The process of building a comprehensive measurement model is carried out through the following steps: Step 1: Check the validity of each set of indicators measuring each latent variable Check the validity of the set of technical quality variable measurements


Table 3.31. Variance explained by latent variables (Communalities)



Initial

After extraction

(Extraction)

question64_1

.452

.564

cau66_1

.493

.786

cau7d_1

.229

.275


Table 3.32. Total Variance Explained



Factor

Initial Eigenvalues

The sum of the squared load factors follows

Extraction Sums of Squared Loadings


Total

%

variance (% of

Variance)

Cumulative %

%)


Total

%

variance (% of

Variance)

Cumulative %

%)

1

2,025

67,510

67,510

1,624

54,141

54,141

2

.647

21,581

89,091




3

.327

10,909

100,000





Table 3.33. Factor Matrix



Factor

1

question64_1

.751

cau66_1

.886

cau7d_1

.524

The three tables above show that cau 7d has a very low communality coefficient, however, there is still a factor created with the loading coefficient of all three indicators greater than 0.5, although cau7d still has the lowest loading coefficient. Therefore, the author still keeps cau7d in the technical quality measurement model.

Table 3.34. Reliability Statistics



Cronbach's Alpha

Number of variables (N of

Items)

.756

3

Comment


Agree Privacy Policy *