Run a whole normal-mixture simulation without looping over replicates — vectorised_normal_mixture

Shared implementation behind the vectorised fast path. Given the prior mixture for every replicate, this reproduces exactly what the replicate loop would have produced: posterior moments, medians, credible intervals, test decisions and effective sample sizes.

Usage

vectorised_normal_mixture_simulation(
  weights,
  means,
  sds,
  samples,
  target_data,
  to_return,
  critical_value,
  theta_0,
  confidence_level,
  null_space,
  posterior_parameters = NULL,
  posterior = NULL,
  mcmc = FALSE
)

Arguments

weights: Prior component weights, shared across replicates (a vector) or one row per replicate (a matrix).
means: Prior component means, shaped like weights.
sds: Prior component standard deviations, shaped like weights.
samples: Data frame of generated replicates, with columns treatment_effect_estimate, treatment_effect_standard_error and standard_deviation.
target_data: Target data object, used for its sample size per arm.
to_return: Character vector of requested outputs.
critical_value: Critical value for the test decision.
theta_0: Null hypothesis value.
confidence_level: Credible interval level.
null_space: Either "left" or "right".
posterior_parameters: Optional data frame of per-replicate posterior parameters to report.
posterior: Optional output from normal_mixture_posterior() when the caller already needed it for method-specific parameter summaries.
mcmc: Whether the calling model samples when it is not on this fast path, which decides what the MCMC diagnostics report.

Value

A list shaped like the return value of Model$simulation_for_given_treatment_effect().