In my research related to social sciences (human resources) I have 5 independent variables which have a highly significant positive correlation with the dependent variable. But on running the multiple regression, the value of constant is appearing with a negative sign though all the Independent variables have positive beta and relationship. The Adj R square value is also considerably good (greater than .60)
I have checked the autocorrelation (durbin-watson) and multicollinearity (VIF and tolerance) and the statistical tests suggests that there are no such issues.
What can be the possible reason of such negative sign, when all are indicators are signifying a good strong effect of the independent variables on dependent.
If this signifies a problem in data or in the tools of analysis applied then what can be the best possible measure to rectify that.