search

rsaggio87 / LeaveOutTwoWay

Bias corrected estimates of variance components in two fixed effects models as described in Kline, Saggio and Sølvsten (2020)
24 stars 11 forks source link

Confused #16

Open thomasszuber opened 3 years ago

thomasszuber commented 3 years ago

Hi, Thank you very much for all this work! We are just confused about the explanation for epsilon in leave_out_COMPLETE.m: "Smaller values of epsilon implies less accuracy but higher computation time." High values of epsilon seem like a good bargain? Best, Thomas Zuber

rsaggio87 commented 3 years ago

Hi Thomas,

Thanks for your question and apologies for the late reply, I missed the notification on this.

It's a typo! It should read "Smaller values of epsilon implies more accuracy but higher computation time."

Thanks for noticing this! Raffa.