Step-downs analysis: aggregated data and analytical code

This file contains analytical code and aggregate data for evaluating the impact of step-downs (reductions in workers' comp payments after several months in the system) on scheme exit, plus the R project file. I have included the cleaning file but no case-level data.To use:Download all files into a single folder and open the project file. You should be able to run most RMarkdown files from there. The meta-analysis file depends on outputs from all other analytical files.Note on v5:This update reflect analytical changes to reflect peer review and the correction of an existing error. Analytical changes include an additional sensitivity analysis investigating effects on claims unaffected by step-downs, which would indicate confounding. The existing error was exclusion criteria had not been applied correctly in some cases. Claims affected by step-downs must consider both maximum/minimum caps and compensation rates. Otherwise, claims may have minimal effects of step-downs. E.g., if the cap was $2000 per week with an initial rate of 95%, injured workers could only be included if they made $2000 / 95% or less.本文件包含用于评估降级待遇（step-downs，即系统运行数月后职工工伤赔偿支付额削减）对职工退出工伤补偿计划之影响的分析代码与汇总数据，同时附带R项目文件。本次已包含数据清洗文件，但未提供个案层面数据。使用方法：将所有文件下载至同一文件夹后打开项目文件，即可运行其中绝大多数RMarkdown文档。本元分析文件依赖其余所有分析文件的输出结果。关于v5版本的说明：本次更新纳入适配同行评审的分析调整内容，并修正了一处现存错误。本次分析调整新增一项敏感性分析，用于探究降级待遇对未受其影响的理赔案件的影响，以此排查混杂偏倚。此前存在的错误为：部分场景下排除标准未被正确应用。受降级待遇影响的理赔案件，需同时考量赔付上限/下限与赔偿费率。若未考量上述要素，则降级待遇对该类理赔案件的影响可能微乎其微。例如，若赔付上限为每周2000美元，初始费率为95%，则仅当受伤职工周收入不高于2000美元/95%时，方可纳入该理赔案例。