Home / Accounting Information

Kmo And Bartlett's Test For The Measurement Of Independent Variables Kmo And Bartlett's Test

Table 4.3: KMO and Bartlett's Test for the scale of independent variables KMO and Bartlett's Test

KMO coefficient

0.710


Bartlett test model

Chi-Square Value

1193.272

Degrees of freedom

190

Sig (p – value)

.000

Maybe you are interested!

(Source: Model testing results)

Test of variance extracted factors

The research results shown through the extracted variance table show that the change of independent variables in the study is explained 67.620% by observed variables. Thus, the EFA factor analysis model is appropriate and the scale is accepted.

Table 4.4: Extracted variance table

Total Variance Explained



Factor

Eigenvalues

Only

number after deduction

Only

number after rotation


Total


Direction

wrong

Accumulate method

error (%)


Total


Direction

incorrect quote

Cumulative Variance Extraction (%)


Total


Direction

incorrect quote

Cumulative Variance Extraction (%)

1

4.125

20,627

20,627

4.125

20,627

20,627

2,547

12,736

12,736

2

3,037

15,183

35,810

3,037

15,183

35,810

2,512

12,560

25,296

3

2,497

12,484

48,294

2,497

12,484

48,294

2,451

12,257

37,553

4

1,453

7,263

55,557

1,453

7,263

55,557

2,120

10,600

48,153

5

1,240

6,200

61,757

1,240

6,200

61,757

2,034

10,172

58,325

6

1,173

5,863

67,620

1,173

5,863

67,620

1,859

9,295

67,620

7

0.822

4.110

71,729







Extraction Method: Principal Component Analysis.

(Source: Model testing results)

Load factor verification

The author used observed variables that achieved the reliability of 7 independent variables to conduct EFA factor analysis, the research results are shown in the table below:

Table 4.5: Rotation matrix


Rotated Component Matrix a


Component


1

2

3

4

5

6

HLPL3

0.817






HLPL4

0.814






HLPL1

0.768






HLPL2

0.747






KSNB4


0.829





KSNB1


0.801





KSNB3


0.725





KSNB2


0.686





KNPM2



0.865




KNPM1



0.757




KNPM3



0.731




TDQL1




0.848



TDQL2




0.744



TDQL3




0.670



DTBD3





0.818


DTBD2





0.776


DTBD1





0.671


CLDL3






0.774

CLDL2






0.736

CLDL1






0.666

(Source: Model testing results)

The results of EFA factor analysis for the independent variables of the rotated factor matrix show that: the factor loading coefficients of the observed variables all satisfy the conditions when analyzing the factor.

factor ( 0.5) and the number of factors generated by factor analysis is 6 factors. This result is consistent with the initial hypothesis about the corresponding measurement variables for each independent factor.

4.2.3.2 Exploratory analysis EFA for the dependent variable "Organization of accounting information system of state budget spending units in Bac Tan Uyen district, Binh Duong"

The research results in the table below show that the KMO coefficient = 0.656 (> 0.5) and Bartlett's test is statistically significant with Sig. = 0.000 (< 0.05). Thus, using the EFA model to evaluate the value of the dependent variable scale is appropriate.

Table 4.6: KMO and Bartlett test for dependent variable scale


KMO coefficient

.656


Bartlett test model

Chi-Square Value

71,736

Degrees of freedom

3

Sig (p – value)

.000

(Source: Model testing results) Testing the extracted variance of factors

The results of the analysis of variance extracted for the dependent variable scale showed that 60.426% of the change in the dependent factor was explained by the observed variables of this variable. Thus, it was concluded that the EFA factor analysis model was appropriate and the scale was accepted.

Table 4.7: Extracted variance



Factor

Eigenvalues

Index after deduction


Total


Variance

extract

Accumulated variance extraction


Total


Variance

extract

Accumulated variance extraction

1

1,813

60,426

60,426

1,813

60,426

60,426

2

0.647

21,579

82,005




3

0.540

17,995

100,000




Extraction Method: Principal Component Analysis.

(Source: Model testing results)

4.2.4 Regression analysis

4.2.4.1 General regression model

To examine the relationship between independent variables (System of legal documents on accounting; Management level and understanding of accounting of managers; Responsiveness of accounting software and applications; Quality of accounting data; Training and fostering of accounting staff and Internal control procedures) and dependent variable (Organization of accounting information system of units using state budget in Bac Tan Uyen district, Binh Duong), the study uses the following multivariate regression model:

Accounting system = β 0 + β 1 System of payment + β 2 TDQL + β 3 KNPM + β 4 CLDL + β 5 DTBD + β 6

KSNB + ε

In there:

- Dependent variable:

HTTTKT: Organizing the HTTTKT of units using the State budget in Bac Tan Uyen district, Binh Duong.

- Independent variable:

HTPL: System of legal documents on accounting

TDQL : Management level and accounting knowledge of the manager

KNPM: Responsiveness of accounting software and applications

CLDL: Accounting data quality

DTBD: Training and fostering technical staff;

KSNB: KSNB procedures;

ε: Noise factor.

β: Regression coefficient.

4.2.4.2 Testing the suitability of the research model

The results show that the model has a satisfactory fit (R Square = 0.614). The adjusted R 2 coefficient in this model is 0.599, which means that the linear regression model built fits the data set to 59.9%, in addition, the F test is statistically significant with Sig. < 0.05 (ANOVA table), from which it can be concluded that the research model is suitable, the independent variables explain 59.9% of the change in the dependent variable.

dependent variable, while 40.1% of the variation in the dependent variable is explained by other factors.

were not considered in this study.

Table 4.8: Regression model summary table


Model


R factor

R2 coefficient

R2 coefficient - correction

Standard error of estimate

Durbin-Watson

1

.784a​

0.614

0.599

0.35239

2,120

(Source: Model testing results)

Table 4.9: ANOVA analysis table ANOVA a


Model

Total average

direction


Degrees of freedom

Mean square


F


Sig.


1

Regression

30,272

6

5,045

40,631

.000 b

Remainder

18,999

153

0.124



Total

49,272

159




(Source: Model testing results)

4.2.4.3 Regression weight testing

Through the research results shown in the regression weight table, it can be seen that the Sig values ​​of the variables HTPL, TDQL, KNPM, CLDL, DTBD, and KSNB are all less than 0.05, so the author concludes that the variables HTPL, TDQL, KNPM, CLDL, DTBD, and KSNB are correlated and significant with the variable CL.

Table 4.10: Regression results table



Model

Unstandardized coefficient

Standardization factor


t


Sig.


Multicollinearity statistics

B

Standard error

Beta

Tolerance factor

VIF coefficient


1

(Constant)

-0.565

0.300


-1.885

0.061



HTPL

0.114

0.043

0.150

2,632

0.009

0.774

1,293

TDQL

0.303

0.049

0.358

6.210

0.000

0.760

1,315

KNPM

0.287

0.051

0.317

5,582

0.000

0.781

1,281

CLDL

0.227

0.045

0.256

5.003

0.000

0.963

1,038

DTBD

0.159

0.044

0.188

3,643

0.000

0.946

1,057

KSNB

0.127

0.045

0.161

2,829

0.005

0.776

1,289

a. Dependent Variable: HTTTKT

(Source: Model testing results)

Based on the research results shown in the regression weight table, the author determined the regression equation for the research on factors affecting the organization of the financial management system of the units using the state budget in Bac Tan Uyen district, Binh Duong as follows:

Technical Information System = 0.358.TDQL + 0.317.KNPM + 0.256.CLDL + 0.188.DTBD + 0.161.KSNB3 + 0.150.HTPL

To compare the level of influence of each independent factor on the Organization of Information Technology

of the units using the State budget in Bac Tan Uyen district, Binh Duong, we base on the standardized Beta coefficient. Accordingly, the factor with the larger standardized Beta weight means that the factor has a stronger influence on the dependent variable. We see that, in the regression equation, of the 6 factors affecting the organization of the accounting information system of the units using the State budget in Bac Tan Uyen district, Binh Duong, the factor of Management level and understanding of accounting of managers has the strongest influence with Beta = 0.358; the factor of Responsiveness of software and accounting applications has the second strongest influence with Beta coefficient = 0.317; the factor of Accounting data quality has the third strongest influence with Beta coefficient = 0.256; the next factor is the factor of Training and

Training of auditors ranked fourth with Beta coefficient = 0.188, followed by Internal Control Procedures ranked fifth with Beta coefficient = 0.161 and the accounting legal system ranked lowest with Beta coefficient = 0.150.

4.2.5 Testing research hypotheses

4.2.5.1 Testing for multicollinearity

According to Nguyen Dinh Tho (2011), multicollinearity is a phenomenon in which independent variables including TDQL, KNPM, CLDL, DTBD, KSNB and HTPL are completely correlated with each other. To check this phenomenon, people use the variance inflation factor VIF. The results shown in the regression weight table show that the VIF coefficients of the variables TDQL, KNPM, CLDL, DTBD, KSNB and HTPL are all less than 2, therefore, the author concludes that the research model on factors affecting CL does not have multicollinearity.

4.2.5.2. Testing for autocorrelation of residuals

According to Nguyen Dinh Tho (2011), when random errors are correlated, autocorrelation may occur. To test this phenomenon, we use the Durbin-Watson coefficient. If the error parts do not have a first-order serial correlation with each other, the Durbin-Watson coefficient value will be close to 2. Based on the research results, d = 2.120 (d 2), so it is concluded that there is no autocorrelation between the residuals in the research model, the research model is meaningful.

4.2.5.3 Testing for normal distribution of residuals

Histogram and PP Plot are used to test the normal distribution of the residuals. Based on the research results, the Histogram shows a normal distribution curve superimposed on the frequency graph, the graph also shows the standard deviation Std.Dev is 0.981 and Mean 0, therefore, the author concludes that the normal distribution of the residuals is not violated.


Figure 4.1: Histogram of residuals – normalized

(Source: Model testing results)

Regarding the PP Plot of standardized residuals, it can be seen that the observation points are not scattered far apart but concentrated near the expected diagonal, so the author concludes that the normal distribution of the residuals is not violated.

Figure 4.2: PP Plot of residuals – standardized

(Source: Model testing results)

Comment


Agree Privacy Policy *