As indicated in Table 4, all the VIFs are within acceptable decision criterion, that is, within 5, which suggest that the predicting variables are free of multicollinearity.
4.3.2 Serial (Auto) Correlation
A number that tests for serial or autocorrelation in the residuals from a statistical regression analysis is called the Durbin-Watson statistics, always between 0 and 4. A value of approaching 2 means that there is no autocorrelation in the sample, values approaching 0 indicate positive autocorrelation and values toward 4 indicate negative autocorrelation. The results of test for serial correlation are reported in Table 5.

As indicated in Table 5, the Durbin-Watson statistics for the three models with moderation of firm size (natural logarithm of total assets) and general economic condition (interest rate) were 1.846, 2.626 and 2.133 respectively, which show that there is no serial or auto correlation problem in the data (Durbin & Watson, 1951).
4.3.3 Heteroskedasticity
The null hypothesis was that there is no heteroskedasticity in all the models with or without moderating variables. The results of heteroskedasticity test for the three models are reported in Table 6.

As indicated in Table 6, for the regression model with NPM as the response variable, the test yielded a chi-square value of 12.41 with a p-value of 0.0004 with moderation and a chi2 value of 3.26 with a p-value of 0.0712 without moderation. The chi-square value with moderating variables was statistically significant at 5% significance level and hence the null hypothesis was rejected to signify the existence of heteroskedasticity. However, the chi-square value without moderating variables was not statistically significant at 5% significance level and hence the null hypothesis is accepted to signify the non-existence of heteroskedasticity.