NEWEY-WEST adjustment and sample size requirement

[Fredrick orwa]

Dear Team,

I would like to ask if anyone knows the minimum required sample size to use Newey-West standard errors. I conducted an analysis with 20 data points (12 pre-intervention and 8 post-intervention) using Negative Binomial Regression and included Newey-West standard errors. However, a statistician rejected the analysis, stating that the sample size is too small for Newey-West adjustment.

I haven’t come across any published literature that specifies a minimum required sample size for this method. Could you kindly advise?

Regards,
Fred

[Fredrick orwa]

Hi team still checking if someone has an idea on this.regards fred

[Carlo Lazzaro]

Fredrick:
you may want to take a look at r - Small sample (20-25 observations) - Robust standard errors (Newey-West) do not change coefficients/standard errors. Is this normal? - Stack Overflow and related threads.

[Fredrick orwa]

Thanks so much Carlo Lazzaro