robust_mixed: Cluster robust standard errors with degrees of freedom adjustments for lmerMod/lme objects

Description

Function to compute the CR2/CR0 cluster robust standard errors (SE) with Bell and McCaffrey (2002) degrees of freedom (dof) adjustments. Suitable even with a low number of clusters. The model based (mb) and cluster robust standard errors are shown for comparison purposes.

Usage

robust_mixed(m1, digits = 3, type = "CR2", satt = TRUE, Gname = NULL)

Value

A data frame (results) with the cluster robust adjustments with p-values.

Estimate: The regression coefficient.
mb.se: The model-based (regular, unadjusted) SE.
cr.se: The cluster robust standard error.
df: degrees of freedom: Satterthwaite or between-within.
p.val: p-value using CR0/CR2 standard error.
stars: stars showing statistical significance.

Arguments

m1: The lmerMod or lme model object.
digits: Number of decimal places to display.
type: Type of cluster robust standard error to use ("CR2" or "CR0").
satt: If Satterthwaite degrees of freedom are to be computed (if not, between-within df are used).
Gname: Group/cluster name if more than two levels of clustering (does not work with lme).

Author

Francis Huang, huangf@missouri.edu

Bixi Zhang, bixizhang@missouri.edu

References

Bell, R., & McCaffrey, D. (2002). Bias reduction in standard errors for linear regression with multi-stage samples. Survey Methodology, 28, 169-182. (link)

Liang, K.Y., & Zeger, S. L. (1986). Longitudinal data analysis using generalized linear models. Biometrika, 73(1), 13-22. (tools:::Rd_expr_doi("10.1093/biomet/73.1.13"))

Examples

Run this code

require(lme4)
data(sch29, package = 'MLMusingR')
robust_mixed(lmer(math ~ male + minority + mses + mhmwk + (1|schid), data = sch29))

Run the code above in your browser using DataLab