1. What have the authors contributed in "Running head: missing data in experiments 1 missing data in experiments: challenges and solutions" ?
The authors describe different classes of missing data typically encountered in experimental datasets, and they discuss how each of them impacts researchers ’ causal inferences.. In this tutorial, the authors provide concrete guidelines for handling each class of missingness, focusing on two methods that make realistic assumptions: i ) Inverse Probability Weighting ( IPW ) for mild instances of missingness, and ii ) Double Sampling and Bounds for severe instances of missingness.. After reviewing the reasons why these methods increase the accuracy of researchers ’ estimates of effect sizes, the authors provide lines of R code that researchers may use in their own analyses.
read more
2. What is the way to handle missingness in experimental data?
After addressing all missingness in covariates included in their analysis using mean substitution, researchers may correct for missingness in the dependent variable with inverse probability weighting or double sampling and bounds.
read more
3. What is the purpose of the procedure for estimating the effect of the treatment on the DV?
In the experimental framework, including covariates in the analysis serves one purpose: increasing the precision of your estimate of the effect of the treatment on the DV (the ATE).
read more
4. How do the authors estimate the ATE in a sample of 2,000 employees?
To estimate the ATE in the population of 20,000 employees based on this sample of 2,000 employees, the authors use one of the following two linear regression models:Model 1:lm(DV ~ treatment , data = dat)Model 2:lm(DV ~ treatment + race + gender + education + pretest ,data = dat)
read more
