Additional file 2: of A comparison of analytic approaches for individual patient data meta-analyses with binary outcomes Doneal Thomas Robert Platt Andrea Benedetti 10.6084/m9.figshare.c.3692995_D5.v1 https://springernature.figshare.com/articles/journal_contribution/Additional_file_2_of_A_comparison_of_analytic_approaches_for_individual_patient_data_meta-analyses_with_binary_outcomes/4658323 Median (Interquartile range (IQR)) (%) root mean square error for treatment effect, β1 for different approach, by number of studies, total average sample size, mixture of studies sizes and degree of random effects variances - data generated from random study- and treatment effect: Eq. 1 with 5% outcome rate. (DOC 70 kb) 2017-02-16 05:00:00 Individual patient data meta-analyses One- and two-stage models Generalized linear mixed models Penalized quasi-likelihood Adaptive gauss-hermite quadrature Fixed and random study-effects