Is it possible that in multiple regression, the regression model is non-significant but it still results in a significant main or interaction effect?

And if it is possible, is it still possible to interpret the main or interaction effect, if one is not interested in predicting the criterion variable with all predictors but particularly...

25 March 2013 4,284 7 View