Skip to contents

Simulate a research literature with a gender-composition moderator

Source: R/make_studies.R
make_studies.Rd

Generates multiple simulated research literatures in which the true ATE in each study equals 0.1 * proportion_women, where proportion_women ranges uniformly from -0.5 to 0.5 across N_studies studies. Each "literature" is a random ordering of those studies, allowing the cumulative meta-regression estimate to be traced as evidence accumulates. Used in Chapter 11 to illustrate how the path of knowledge accumulation affects how quickly a moderating relationship is recovered.

Usage

make_studies(N, N_studies, N_literatures)

Arguments

N

Integer. Sample size for each individual study.

N_studies

Integer. Number of studies in the literature.

N_literatures

Integer. Number of random orderings (literature histories) to generate.

Value

A tibble of simulation results from DeclareDesign::simulate_designs(), with an added study column giving each study's position in its literature's random ordering.

Details

Requires the DeclareDesign package.