How to manually provide data for benchmark covariates?

[sharifamlani]

My model requires me to adjust for cluster standard errors. I use the coeftest function in R to do this. Unfortunately, the object created by the coeftest function is not compatible with the model argument in sensemakr. Thankfully, the estimate, se, and dof can be inputted manually into the sensemakr function.

However, I do not see a way to input the benchmark covariates manually? How can I input the benchmark covariates manually into the sensemakr function?

[carloscinelli]

Hi Sharif, I answered by e-mail but will also include the answer here for future reference.

For the benchmark covariate Xj, you will need to provide the partial R2 of Xj both with the treatment D and the outcome Y.

That is, there are two arguments:

• For the argument r2dxj.x you will provide the partial R2 of the benchmark covariate Xj with the treatment D. You can compute that with the function partial_r2 using a regression of the treatment on the covariates (or using directly the t-value and degrees of freedom).

• For the argument r2yxj.dx you will provide the partial R2 of the benchmark covariate Xj with the outcome Y. You can compute that with the function partial_r2 using a regression of the outcome on the treatment and the covariates (or using directly the t-value and degrees of freedom).

You can see an example of how to do that in the software paper, page 18.