parameters (version 0.4.0)

ci_wald: Wald-test approximation for CIs and p-values

Description

The Wald-test approximation treats t-values as Wald z. Since the t distribution converges to the z distribution as degrees of freedom increase, this is like assuming infinite degrees of freedom. While this is unambiguously anti-conservative, this approximation appears as reasonable for reasonable sample sizes (Barr et al., 2013). That is, if we take the p-value to measure the probability of a false positive, this approximation produces a higher false positive rate than the nominal 5% at p = 0.05.

Usage

ci_wald(
  model,
  ci = 0.95,
  dof = NULL,
  effects = c("fixed", "random", "all"),
  component = c("all", "conditional", "zi", "zero_inflated", "precision", "scale",
    "smooth_terms"),
  robust = FALSE,
  ...
)
p_value_wald(model, ...)
# S3 method for merMod
p_value_wald(model, dof = Inf, ...)

Arguments

model

A statistical model.

ci

Confidence Interval (CI) level. Default to 0.95 (95%).

dof

Degrees of Freedom. If not specified, for ci_wald(), defaults to model's residual degrees of freedom (i.e. n-k, where n is the number of observations and k is the number of parameters). For p_value_wald(), defaults to Inf.

effects

Should standard errors for fixed effects or random effects be returned? Only applies to mixed models. May be abbreviated. When standard errors for random effects are requested, for each grouping factor a list of standard errors (per group level) for random intercepts and slopes is returned.

component

Should all parameters, parameters for the conditional model, or for the zero-inflated part of the model be returned? Applies to models with zero-inflated component. component may be one of "conditional", "zi", "zero-inflated" or "all" (default). May be abbreviated.

robust

Logical, if TRUE, robust standard errors are computed by calling standard_error_robust(). standard_error_robust(), in turn, calls one of the vcov*()-functions from the sandwich-package for robust covariance matrix estimators.

...

Arguments passed down to standard_error_robust() when confidence intervals or p-values based on robust standard errors should be computed.

Value

The p-values.

References

Barr, D. J. (2013). Random effects structure for testing interactions in linear mixed-effects models. Frontiers in psychology, 4, 328.

Examples

# NOT RUN {
library(lme4)
model <- lmer(Petal.Length ~ Sepal.Length + (1 | Species), data = iris)
p_value_wald(model)
ci_wald(model, ci = c(0.90, 0.95))
# }
# NOT RUN {
# }