search

kkawato / rdlearn

Safe Policy Learning under Regression Discontinuity Designs with Multiple Cutoffs
Other
0 stars 0 forks source link

Feature request: Summary function to summarize the output. #8

Open sou412 opened 3 months ago

sou412 commented 3 months ago

Currently, the only way to interact the output from rdlearn() is to either plot or directly look at objects inside of the output. It would be nice to have a summary() function to summarize some of the results.

kkawato commented 3 months ago

That's a good idea. But actually the results in the paper (6.2. Empirical Findings) are basically the plots of the sensitivity analysis that the function plot provides.

Of course, it's better than nothing to show some numerical summaries(e.g. numerical summaries of the difference from base cutoff), but I'm not sure how statistically meaningful this is. @carolyyzz Do you have any idea on the summary function?

carolyyzz commented 3 months ago

I think it would be beneficial to have a numerical output of the cutoff changes compared to the baseline (per sensitivity/cost parameter). Additionally, including some standard RD estimates, such as the local ATE at the original cutoff, would be helpful.

kkawato commented 3 months ago

I apologize for the delay. I have been focused on my Master's course application since last week. I plan to start working on coding on July 2nd.