Sensitivity to bootstrap iteration count in qgcompmulti fits

Re-fits qgcomp.glm.multi() across multiple B values while preserving the rest of the analysis specification.

Usage

b_sensitivity(
  f,
  data,
  mix1,
  mix2,
  B_values,
  interaction = TRUE,
  family = gaussian(),
  q = 4,
  centering = "none",
  id = NULL,
  MCsize = nrow(data),
  seed = NULL,
  keep_fits = TRUE
)

Arguments

f, data, mix1, mix2, interaction, family, q, centering, id, MCsize, seed

Arguments passed through to qgcomp.glm.multi().

B_values

Integer vector of bootstrap iteration counts to compare.

keep_fits

Logical; if TRUE, retain the full fitted objects.

Value

An object of class "qgcompmulti_b_sensitivity".

Details

This helper is intended to answer a practical inferential question: are the reported bootstrap-based standard errors and related summaries reasonably stable as the number of bootstrap replications changes?

B sensitivity is implemented as a repeated-fit workflow in which all settings other than B are held fixed.

When seed is supplied, this helper treats it as a master seed and deterministically derives one distinct fit-specific seed for each requested value of B. This preserves reproducibility of the overall sensitivity workflow while avoiding reuse of the same bootstrap resamples across the different refits.

The resulting object is designed for stability assessment rather than automatic tuning. Users should focus especially on whether bootstrap-based standard errors, bootstrap retention counts, and broad qualitative conclusions are reasonably consistent across the requested B values.

Examples

if (FALSE) { # \dontrun{
dat <- sim_mixture_data(
  n = 400,
  pA = 3,
  pB = 3,
  rho_within_A = 0.3,
  rho_within_B = 0.3,
  rho_between = 0.2,
  psi1 = 0.5,
  psi2 = 0.3,
  psi12 = 0.2,
  seed = 123
)

b_sensitivity(
  f = Y ~ X1 + X2 + X3 + W1 + W2 + W3 + C,
  data = dat,
  mix1 = c("X1", "X2", "X3"),
  mix2 = c("W1", "W2", "W3"),
  B_values = c(50, 100, 200),
  q = 4,
  MCsize = nrow(dat),
  seed = 13
)
} # }