Autocorrelation Multicollinearity Test Heteroskesdastisity

65 Total mean of influence of consumer decision is 3.74. Following to the likert scale above, it shows that correspondents of this research are attending to “neutral”. The indicators show the almost of respondents are agree. Its explained Products offered through the internet has a competitive price is 3.77, Cheaper product prices when shopping over the Internet from the ordinary the store is 3.65, Products on the Internet often give a discount is 3.79, Products on the Internet has variable values is 3.66, The products on the internet have a significant price is 3.64, I can compare prices on each different type of goods in each store online is 3.82, By shopping through the Internet, I often get a product with a lower price is 3.86.

C. Classical Assumption Test

1. Autocorrelation

Table 4.11 Durbin Watson Durbin Watson Conclusion Less than 1,10 Autocorrelation available 1,10 and 1,54 Without conclusion 1,55 and 2,46 No autocorrelation available 1,46 and 2,90 Without conclusion More than 2,91 Autocorrelation available Source: Muhammad firdaus 2004:101 66 Autocorrelation test in certain model is aimed to know the availability of correlation between disturbance variable e1 in the previous period et-1. Autocorrelation test can be made by using Durbin Watson Test. Table 4.12 Autocorrelation Test Model Summary b a. Predictors: Constant: PP, SE, CS b. Dependent Variable: The influence of consumer decision on online shopping Source: Processed data by SPSS To know whether there is autocorrelation or not, it can be seen through a certainty in the table 4.12 Durbin Watson. The result of Durbin Watson test used SPSS is 2.114 with the standard of significant 0.05, according to the table 4.12, there is no autocorrelation in this linear regression models, and this model is suitable to be used. Model Durbin-watson 1 2.114 67

2. Multicollinearity Test

Table 4.13 Multikolinearity Test a.Dependent Variable: The influence of consumer decision on online shopping Based on the table 4.13 that the value of tolerance is not less than 0.1 and the value of Variant Inflation Factor not more than 10, meaning that this analysis does not indicate there is a tendency towards variable multikolinearitas research, so that it can be concluded that this research is supported by classical theory and reasonable use in testing. Collinearity Statistics Model Tolerance VIF Constant Product Perception .852 1.174 Shopping Experience .847 1.180 1 Customer Service .963 1.038 68

3. Heteroskesdastisity

Graphic 4.1 __ To find is there any or no Heteroskesdastisity is using scatterplot that referred to graphic of plot between prediction value dependent ZPRED and the residual ZRESID, when Y has predicted and X is residual. In the output SPSS 16 for windows scatter plot shows the dots that described is spread above and below or around 0 number and did not formed specific pattern. Hence could be concluding as the regression model did not have any serious problem. In the other words that residual variance model from one observation to other observation is constant or homokedastisitas. This finding shows that multiple regression model is suitable to be used in this research. 69

4. Normality Data