Microsoft Word 2007 ichrie conference Proceedings Final-Final 06-06-07. doc

 
2007 Annual International CHRIE Conference & Exposition 
55
considered the best that the outliers be removed from the sample. After removing the outliers, the regression model 
was reestimated and the studentized residuals versus the independent variables were regraphed. The predicted values 
showed no evidence of unequal variance. A normal probability plot of the residuals exhibited the characteristics of 
short tailed data. However, short tailed residuals do not necessarily have huge effects on the estimated regression 
coefficients. Since the main objective in this portion is to evaluate the regression coefficient of the recession 
variable, the possible short tailed nature of the residuals is not a huge concern. Multicollinearity was not an issue as 
none of the variance inflation factors approached unacceptable heights. Most of the VIFS were under two.
Test Results 
Several notable observations were identified with respect to mean total assets and sample observation 
years. As shown on Table 2, the mean total assets for the current accrual approach show considerably larger 
amounts for every industry when compared to those from the total accruals approach. The authors also observe the 
same phenomenon regarding the sample observation years. The reason behind the significant difference between 
the two approaches might have resulted from the accounting variable requirements for each approach. Depending 
on the firms that are included as final samples, the mean total assets are affected and resulted in great differences in 
amounts.