EValue - Sensitivity Analyses for Unmeasured Confounding and Other Biases in Observational Studies and Meta-Analyses

Conducts sensitivity analyses for unmeasured confounding, selection bias, and measurement error (individually or in combination; VanderWeele & Ding (2017) <doi:10.7326/M16-2607>; Smith & VanderWeele (2019) <doi:10.1097/EDE.0000000000001032>; VanderWeele & Li (2019) <doi:10.1093/aje/kwz133>; Smith, Mathur, & VanderWeele (2021) <doi:10.1097/EDE.0000000000001380>). Also conducts sensitivity analyses for unmeasured confounding in meta-analyses (Mathur & VanderWeele (2020a) <doi:10.1080/01621459.2018.1529598>; Mathur & VanderWeele (2020b) <doi:10.1097/EDE.0000000000001180>) and for additive measures of effect modification (Mathur et al., <doi:10.1093/ije/dyac073>).

Last updated

6.89 score 3 stars 1 dependents 155 scripts 2.2k downloads

MetaUtility - Utility Functions for Conducting and Interpreting Meta-Analyses

Contains functions to estimate the proportion of effects stronger than a threshold of scientific importance (function prop_stronger), to nonparametrically characterize the distribution of effects in a meta-analysis (calib_ests, pct_pval), to make effect size conversions (r_to_d, r_to_z, z_to_r, d_to_logRR), to compute and format inference in a meta-analysis (format_CI, format_stat, tau_CI), to scrape results from existing meta-analyses for re-analysis (scrape_meta, parse_CI_string, ci_to_var).

Last updated

3.76 score 2 dependents 26 scripts 1.9k downloads

Replicate - Statistical Metrics for Multisite Replication Studies

For a multisite replication project, computes the consistency metric P_orig, which is the probability that the original study would observe an estimated effect size as extreme or more extreme than it actually did, if in fact the original study were statistically consistent with the replications. Other recommended metrics are: (1) the probability of a true effect of scientifically meaningful size in the same direction as the estimate the original study; and (2) the probability of a true effect of meaningful size in the direction opposite the original study's estimate. These two can be computed using the package \code{MetaUtility::prop_stronger}. Additionally computes older metrics used in replication projects (namely expected agreement in "statistical significance" between an original study and replication studies as well as prediction intervals for the replication estimates). See Mathur and VanderWeele (under review; <https://osf.io/apnjk/>) for details.

Last updated

1.58 score 19 scripts 133 downloads

NRejections - Metrics for Multiple Testing with Correlated Outcomes

Implements methods in Mathur and VanderWeele (in preparation) to characterize global evidence strength across W correlated ordinary least squares (OLS) hypothesis tests. Specifically, uses resampling to estimate a null interval for the total number of rejections in, for example, 95% of samples generated with no associations (the global null), the excess hits (the difference between the observed number of rejections and the upper limit of the null interval), and a test of the global null based on the number of rejections.

Last updated

1.15 score 14 scripts 212 downloads

SimTimeVar - Simulate Longitudinal Dataset with Time-Varying Correlated Covariates

Flexibly simulates a dataset with time-varying covariates with user-specified exchangeable correlation structures across and within clusters. Covariates can be normal or binary and can be static within a cluster or time-varying. Time-varying normal variables can optionally have linear trajectories within each cluster. See ?make_one_dataset for the main wrapper function. See Montez-Rath et al. <arXiv:1709.10074> for methodological details.

Last updated

1.11 score 13 scripts 170 downloads