search

rasmusab / bayesian_first_aid

Inside every classical test there is a Bayesian model trying to get out.
326 stars 40 forks source link

cred.mass parameter not passed to mcmc_stats when computing diff_stats in bayes.prop.test #7

Closed benjaminMears closed 9 years ago

benjaminMears commented 9 years ago

On line 182 in bayes_prop_test.R, I believe the parameter cred_mass = cred.mass should be passed to mcmc_stats so that the specified HDI is computed for the theta differences.

Example current output:

bayes.prop.test(x = c(34, 54), n = c(400, 500), cred.mass = .9)$stats mean sd HDI% comp HDIlo HDIup theta[1] 0.08717148 0.01400703 90 0.5 0.06398359 0.10948452 theta[2] 0.10958137 0.01388894 90 0.5 0.08678420 0.13215677 x_pred[1] 34.89780000 7.88862708 90 0.5 22.00000000 47.00000000 x_pred[2] 54.80966667 9.98124649 90 0.5 38.00000000 70.00000000 theta_diff[1,2] -0.02240990 0.01966577 95 0.0 -0.06066122 0.01670726

rasmusab commented 9 years ago

Thanks for reporting! Fixed now.