Can anyone tell me why the ncv.test output and the gvlma output would be
contradictory on the question of heteroscedasticity? Below, the ncv.test
output reveals a problem with heteroscedasticity, but the gvlma output says
that the assumptions are acceptable. How is this reconciled?
> ncv.test(defmodA)
Non-constant Variance Score Test
Variance formula: ~ fitted.values
Chisquare = 7.360374 Df = 1 p = 0.00666769
> gvlma(defmodA)
Call:
lm(formula = DefPunWmn1 ~ DefPersBenef, data = Data)
Coefficients:
(Intercept) DefPersBenef
1.2579 0.1572
ASSESSMENT OF THE LINEAR MODEL ASSUMPTIONS
USING THE GLOBAL TEST ON 4 DEGREES-OF-FREEDOM:
Level of Significance = 0.05
Call:
gvlma(x = defmodA)
Value p-value Decision
Global Stat 37.3746 1.508e-07 Assumptions NOT satisfied!
Skewness 32.8916 9.744e-09 Assumptions NOT satisfied!
Kurtosis 2.6248 1.052e-01 Assumptions acceptable.
Link Function 0.3684 5.439e-01 Assumptions acceptable.
Heteroscedasticity 1.4899 2.222e-01 Assumptions acceptable.
[[alternative HTML version deleted]]
______________________________________________
R-help@r-project.org mailing list
https://stat.ethz.ch/mailman/listinfo/r-help
PLEASE do read the posting guide http://www.R-project.org/posting-guide.html
and provide commented, minimal, self-contained, reproducible code.
