已收录 273081 条政策
 政策提纲
  • 暂无提纲
Methods to Control for Overt and Hidden Biases in Comparative Studies
[摘要] When the goal of a comparative study is to ascertain the effect of some treatment condition, problems arise when it is not randomly assigned to units.In the absence of random assignment, units compared cannot be expected to be similar in terms of pretreatment covariates, yet the validity of resulting causal inferences relies on this equivalence.This thesis develops techniques that build upon existing methods to analyze comparative studies, lifting certain of their limitations.These methods focus on reducing bias due to nonequivalence of covariates across groups and can be easily combined with techniques that aim to reduce other biases, such as those that arise from a mismatch in the sample and target population.To reduce bias in estimates from comparative studies, the best analysis ensures the likeness of the distributions of measured confounders across comparison groups.Methods such as matching or post-stratification on the measured covariates group similar units, and analysis is performed within subgroups.We apply this bias-reducing idea to the Peters-Belson method, which assesses the existence of a disparity with regression models, to restrict comparisons to groups of units with similar covariate distributions.Propensity scores are a common way to organize units into groups.In practice, the propensity score is estimated by a parametric model, and the literature is divided regarding the selection of the best model.Consistent with one thread of the literature, we develop a method that improves the propensity score model by focusing it on covariates most relevant to an outcome of interest with the creation of a multidimensional prognostic score.By improving the propensity score model, units compared are more similar, and resulting analyses have greater validity.While adjusting for measured confounders can sometimes suffice in the analysis of comparative studies, additional methods -- broadly known as methods of sensitivity analysis -- aim to quantify the potential impact of unmeasured confounders on the effect estimate.We introduce a method of sensitivity analysis for a linear regression model that is unique in its simplicity and ability to assess the impact of unmeasured confounders on the entire confidence interval, rather than only the point estimate.
[发布日期]  [发布机构] University of Michigan
[效力级别] Propensity Scores [学科分类] 
[关键词] Comparative Studies;Propensity Scores;Prognostic Scores;Statistics and Numeric Data;Science;Statistics [时效性] 
   浏览次数:39      统一登录查看全文      激活码登录查看全文