Data for: Using the negative control design with a prenatal exposure subject to assortative mating
Description
The following files simulate the data and run all models. Each do file takes arguments "nobs" and "effsize" to define the number of observations and effectsize - neg_contout_datsim (continuous outcome with no paternal effect) - neg_contout_datsim2 (continuous outcome with paternal effect) - neg_contr_datsim.do (binary outcome) Each of these uses "progs.do" to run the bootstrapping of the confidence interval of the difference in maternal and paternal effect estimates The separate datasets for each sample size and effect size are compiled into a single dataset using "compile_results.do". Statistics across simulations are calculated using the following files: - sim_cont_ready.do (continuous outcome with no paternal effect) - sim_cont_ready2.do (continuous outcome with paternal effect) - sim_bin_ready.do (binary outcome) The final datasets are included as the following .csv files: - neg_cont_simout (bias and other statistics for continuous outcome) - neg_cont_difs (mean and bootstrapped 95% CI of difference in estimates for continuous outcome) - neg_cont_simout (bias and other statistics for continuous outcome with paternal effect) - neg_cont_difs (mean and bootstrapped 95% CI of difference in estimates for continuous outcome with paternal effect) - neg_bin_simout (mean and bootstrapped 95% CI of difference in estimates for binary outcome) - neg_bin_difs (mean and bootstrapped 95% CI of difference in estimates for binary outcome)