pdbailey0
commented
2 years ago @niguarino those are for example purposes only, to get a model to run. I would not use those weights. I'll update the code to make that clearer.
What weights to use is an open question with ongoing debate in the literature. Here is a recent paper on the topic.
If you have questions about weighting a particular dataset, please ask NCES.
Can you describe what you're doing when you "built the weights" in post #3? I am not sure how to incorporate the weights in order to run a properly weighted student-level fixed effects model. Also, I am not sure why you subset the initial data set. Please address this if it influenced how you built the weights. Thank you!