ir_clustBoot function

Updates the covariance using cluster bootstrap

Adjusts error estimates for repeated measures data by use of the cluster bootstrap.

ir_clustBoot(fit, ID, bs_samples = 1000)

Arguments

• fit: Either an ic_par or ic_sp model
• ID: Subject identifier
• bs_samples: Number of bootstrap samples

Details

Standard models in icenReg assume independence between each observation. This assumption is broken if we can have multiple observations from a single subject, which can lead to an underestimation of the standard errors. ir_clustBoot

addresses this by using a cluster bootstrap to fix up the standard errors.

Note that this requires refitting the model bs_samples, which means this can be fairly time consuming.

Examples

# Simulating repeated measures data 
simdata = simIC_cluster(nIDs = 10, nPerID = 4)

# Fitting with basic model
fit = ic_par(cbind(l,u) ~ x1 + x2, data = simdata)
fit

# Updating covariance
ir_clustBoot(fit, ID = simdata$ID, bs_samples = 10)
# (Low number of bootstrap samples used for quick testing by CRAN, 
# never use this few!!)

# Note that the SE's have changed from above
fit

References

Sherman, Michael, and Saskia le Cessie. "A comparison between bootstrap methods and generalized estimating equations for correlated outcomes in generalized linear models." Communications in Statistics-Simulation and Computation 26.3 (1997): 901-925.