定义

attcic是一款实现 Ghanem 等人（2024b）提出的 CiC 样本损耗修正方法的Stata社区命令，适用于连续型与二元型结果变量，同时包含若干对照比较方法。本命令适用于单一处理干预的情境，且每次仅针对一个随访时间点计算样本损耗修正。使用前，需将面板数据集整理为长格式，并安装 esttab 与 leebounds 两个外部命令包。

依赖

ssc install esttab,replace
ssc install leebounds,replace

安装

ssc install attcic, replace

示例

Example 1. Computing attrition corrections for a continuous outcome of a cluster randomized  experiment:

attcic val_prod_animals treat hhid post, rct(yes) ytype(continuous) clustervar(comuid2)  yname(Value of production animals)

 If you want to save the report for the quantile treatment effects, you should specify it in the option qreport(string):

attcic val_prod_animals treat hhid post, rct(yes) ytype(continuous) clustervar(comuid2) yname(Value of production animals) qreport(yes)

In the examples above, attcic uses the default of 200 bootstrap replications to calculate standard errors. If you want to use more bootstrap replications, you should edit the option breps(integer) accordingly.

Example 2. Computing attrition corrections for a binary outcome of a cluster randomized experiment:

attcic own_prod_animals treat hhid post, rct(yes) ytype(binary) clustervar(comuid2) yname(Ownership of production animals)

作者

    Karen Ortiz-Becerra,  University of San Diego, kortizbecerra@sandiego.edu

    Dalia Ghanem, University of California, Davis, dghanem@ucdavis.edu

    Sarojini Hirshleifer, University of California, Riverside,  sarojini.hirshleifer@ucr.edu

    Désiré Kédagni, University of North Carolina, Chapel Hill,  dkedagni@unc.edu