Testing the Information Matrix Equality with Robust Estimators

Figure 11: Regression: RP-Power curves: 20% Vertical outliers
(a) at Y = 5 (n = 50) (b) at Y = 5 and Y = -5 (n = 100)

Figure 12: Regression: RP-Power curves: Bad leverage points:
(a) 5 points at X = (1 6),Y = 6 (n = 50); (b) 5 points at X =
(1 6), Y = 6 (n = 100)

7 Conclusion

We have studied the behaviour of the IM test when robust estimators re-
place the ML estimator in the construction of the test. Particular attention
has been given to the simplest of models, the normal location-scale model
without covariates, where the IM test with ML estimator reduces to the

24

<<   23   24   25   26   27   28   29   >>

More intriguing information