Hi everyone,
I am working with a very small sample, 29 years, of data. I am trying to check the nightlight elasticity for Bangladesh for the 29 years in question. I wanted to know given that I have such a small sample, would including a time-trend make sense? I am worried that it would negate the full extent of the relationship from the regression given the small sample. Is there any literature I can refer to for de-trending with a small sample, or what to do in such a situation?
Thank you!!
I am working with a very small sample, 29 years, of data. I am trying to check the nightlight elasticity for Bangladesh for the 29 years in question. I wanted to know given that I have such a small sample, would including a time-trend make sense? I am worried that it would negate the full extent of the relationship from the regression given the small sample. Is there any literature I can refer to for de-trending with a small sample, or what to do in such a situation?
Thank you!!
