27
C. Data Collection Method
In obtaining the data in this study, researchers used data that has been available. The data will be used in this research is secondary data taken from the
annual financial statements. The data used in this study comes from an external source, companies’s Financial Report listed on the Stock Exchange in the period
2011 - 2014 were obtained from the companys annual financial report obtained from the website Indonesian Stock Exchange IDX is
www.idx.co.id , Indonesian
Capital Market Directory ICMD.
D. Analysis Method
The analytical tool used in this research is multiple linear regression using SPSS, where the regression equation contains elements of interaction
Multiplication of two or more independent variables. This interaction test used to determine the extent of interaction variables can independent commissioner,
managerial ownership, growth opportunities and sales growth influencing Accounting Conservatism.
1. Desriptive Analysis
Descriptive Analysis were used to describe the variables in this study, the level of Accounting conservatism, good corporate governance, growth
opportunities and sales growth on companies listed on the Stock Exchange. Descriptive statistics will provide an overview of each of the variables. The
analytical tool used is the average value mean, distribution frequency, minimum and maximum values and standard deviation Ghozali, 2011
28
2. Classic Assumption Test
Classic assumption test aims to obtain regression results can be accounted for and have results that are not biased or Best Linear Unbiased
Estimator BLUE. Assumptions that must be met, of the test are normality test, autocorrelation test, multicollinearity test and heterocedasticity test.
a. Normality Test Normality test aims to test whether the regression model, the
disturber variables or residuals have a normal distribution. There are two ways to detect whether or not residual normal distribution, namely by
looking at the analysis graph normal probability plot and statistical tests. The regression model that has a data distribution is normal or near-normal
regression model is said to be good Ghozali, 2011. In principle normality can be detected by looking at the spread of the data points on the
diagonal axis of the graph or to view the histogram of the residual. The basis for a decision as follows:
a. If the data is spread around the diagonal line and follow the direction of the diagonal line or histogram graph showing a normal distribution pattern,
then the regression model to meet the assumption of normality. b. If the data are spread far from the diagonal and does not follow the
direction of the diagonal line or histogram chart, present a normal distribution, then the regression model did not meet the assumption of
normality. b. Multicollinearity Test
Multicoloniarity test aims to test whether the regression model found a correlation between independent variables . Multicoloniarity occur
29
in logistic regression analysis when the independent variables are correlated between. Mutikolonieritas can be seen from:
• The value of tolerance and the opponent • Variance Inflation Factor VIF
The results of these tests can be seen from VIF VIF using the equation = 1tolerance. If VIF 10 then there is no multicollinearity.
Ghozali, 2011 c. Heterocedasticity Test
Heterocedasticity test aims to test whether the regression model occurred inequalities residual variance from one observation to another
observation. If the variance of the residuals of the observations to other observations remain, then called Homoskedastisitas, and if different called
Heteroskidastity. This test can be done by looking at the graph plot between the predicted value of the variable ZPRED with residual value
SREID. A good regression model if the residual variance from one observation to another permanent, so that there is no identifiable
heteroskedastisitas Ghozali, 2011. d. Autocorrelation Test
Autocorrelation test whether a correlation exists between the regression model error bullies in the current period t with an error in the
previous period t-1. A good regression model is a regression that is free from autocorrelation. If there is a correlation, then there is a problem
called autocorrelation.
Autocorrelation arise
because successive
30
observation at all times in relation to each other. This test uses a model Durbin Watson DW-Test. The hypothesis to be tested are:
Ho = no autocorrelation r = 0, Ha = no autocorrelation r ≠ 0
If the value of DW is greater than the upper limit or the upper bound du and less than 4- du means no autocorrelation Ghozali, 2011.
3. Hypothesis Testing