lm_robust clustered errors crashes

DeclareDesign / estimatr

estimatr: Fast Estimators for Design-Based Inference

https://declaredesign.org/r/estimatr

Other

131 stars 20 forks source link

lm_robust clustered errors crashes #318

Open jpcoblentz opened 5 years ago

jpcoblentz commented 5 years ago

I have a dataset with 1046167 rows, and lm_robust crashes when I specify a cluster. Is this a known bug with large datasets?

nfultz commented 5 years ago

Can you please post the error message if available?

jpcoblentz commented 5 years ago

it doesn't throw an error, it just hangs and crashes my R session

acoppock commented 5 years ago

Thank you for posting this -- we'll look into large datasets.

nfultz commented 5 years ago

This is likely related / more extreme case of #312 - the large number of rows is likely causing loss of precision somewhere, causing it to loop.

lukesonnet commented 5 years ago

@Jacoblentz415, I recommend setting se_type = "stata" until we can resolve the above. (note: these will be different SEs and dfs as per our docs, but the problem is likely in the CR2 estimator)

lukesonnet commented 5 years ago

@nfultz, I actually think this is closer to #276