Normality Test Multicollinearity Test

30 Stock Exchange which is included in Indonesian Capital Market Directory ICMD from year 2010 to 2013. Hence, the data of exchange rate, interest rate, and inflation are obtained from Indonesian Financial Statistics issued by Indonesia Bank and issuance of publications such as: Financial Statements of Bank Indonesia, Central Bureau of Statistics, and other sources.

G. DATA PROCESSING METHOD

1. Normality Test

Normality test is performed to determine whether the data are normally distributed population. Normality test for each variable is done by looking at the distribution of data points on the graph QQ plot. Data - Data from a variable can be said to be normal, if the distribution of the data are spread on a straight line plot point. According Sarjono and Julianita 2011:64 states that the normality test if the researcher has the respondent 50, then Sig. Kolmogorov-Smirnov compared with alpha, whereas if the researcher has the following respondents 50, then Sig. Shapiro-Wilk compared with Alpha to test the normality of the data obtained by the researcher. Because the respondent in this research are more than 50, then Sig. Kolmogorov-Smirnov compared with alpha will be used. Basis for a decision on the normality test is as follows:  If the numeric significance Kolmogorov- Smirnov Test Sig ≥ 0, 05 then the data are normally distributed. 31  If the numeric significance Kolmogorov-Smirnov Test Sig 0, 05 then the data distribution is not normal. The value of Sig. or significance can be obtained by calculating a test of normality or plot through SPSS tools with confidence level of 95 or 5 error rate. Also in the figure of Q-Q plot is straight line from left to right. The line was derived from the value of Z. if the data is normal distribution, then the data will be scattered around the line.

2. Multicollinearity Test

Multicolinierity test aims to test whether the regression model found a correlation between the independent variables independent Ghozali, 2006. Results are expected in testing is not the correlation between the independent variables. There are several ways to test whether or not multicoloniarity in the regression model. In this test, the researchers used the analysis of the correlation matrix between the independent variables by looking at the value of Tolerance and Variance Inflation Factor VIF. If the tolerance value is greater than or equal to 0.10 VIF value of less than 10, it means not occur multicoloniarity in the regression model.

3. Heteroscedastity Test