IT IS AN EXPERIMENTAL FUNCTION BE CAREFUL! Performs Wald tests of the significance for the dispersion components by response variables. This function is useful for joint hypothesis tests of dispersion coefficients associated with categorical covariates with more than two levels. It is not designed for model comparison.
mc_anova_disp(object, idx_list, names_list, ...)
| object | an object of class |
|---|---|
| idx_list | list with indexes for parameter tests. |
| names_list | list of names to appear in the anova table. |
| ... | additional arguments affecting the summary produced. Note that there is no extra options for mcglm object class. |
A data.frame with Chi-square statistic to test the
null hypothesis of a parameter, or a set of parameters, be
zero. Degree of freedom (Df) and p-values.
The Wald test based on the observed covariance matrix of
the parameters is used.
x1 <- seq(0, 1, l = 100) x2 <- gl(5, 20) beta <- c(5, 0, -2, -1, 1, 2) X <- model.matrix(~ x1 + x2) set.seed(123) y <- rnorm(100, mean = 10, sd = X%*%beta) data = data.frame("y" = y, "x1" = x1, "x2" = x2, "id" = 1) fit.anova <- mcglm(c(y ~ 1), list(mc_dglm(~ x1 + x2, id = "id", data)), control_algorithm = list(tuning = 0.9), data = data)#> Automatic initial values selected.X <- model.matrix(~ x1 + x2, data = data) idx <- attr(X, "assign") idx_list <- list("idx" = idx, "idx" = idx) names_list <- list(colnames(X), colnames(X)) mc_anova_disp(object = fit.anova, idx = idx_list, names_list = names_list)#> Wald test for fixed effects #> #> Covariate Chi.Square Df p.value #> 1 x1 4.9642 1 0.0259 #> 2 x22 45.3509 4 0.0000 #>