median_rlmc: Median RLMC

Description Usage Arguments Details Value Note References See Also

View source: R/median_rlmc.R

Description

Computes the median RLMC by the Monte Carlo simulation according to Ott et al. (2019)

Usage

1
median_rlmc(df, r.tau.prior, MM = 1e+06, seed.value = 12567)

Arguments

df

data frame containing a column sigma

r.tau.prior

randomisation function for the prior

MM

number of Monte Carlo samples

seed.value

random seed for random number generator

Details

Given the MC sample {RLMC^{(j)}, j = 1, …, J}, the median RLMC is

MRLMC = Median({RLMC^{(j)}, j = 1, …, J)}.

See more details in Ott et al. (2019).

Value

median_rlmc

MRLMC computed given individual study-specific sigmai values

median_rlmc_ref

MC MRLMC computed given a geometric mean (sigma_ref) of study-specific sigmai values

Note

Large discrepancy between median_rlmc and median_rlmc_ref can indicate outliers in study-specific sigmai values.

References

Ott, M., Hunanyan, S., Held, L., Roos, M. (2019). The relative latent model complexity adjustment for heterogeneity prior specification in Bayesian meta-analysis. Research Synthesis Methods (under revision).

See Also

sigma_ref


si4bayesmeta documentation built on Dec. 16, 2019, 3:39 p.m.