ABSTRACT
The correlation matrix in Table 4 reveals low levels of pair-wise correlation values among the variables. Technology variables have significant positive correlation with export intensity but of low order. Other firm characteristics like age, advertisement intensity and leverage and CI have statistical significant negative correlation coefficient with export intensity. However, it is generally found that cross-sectional studies suffer from the problem of heteroscedasticity (Gujarati and Sangeetha 2007). The Breusch-Pagan/Cook-Weisberg test reveals that the sample suffers from the problem of heteroscedasticity. Therefore, to get rid of this problem we have estimated the models using the robust4 option available in the STATA 11 statistical package.
