Hello Altruists,

My panel data consists of n=30 and t=27 where I am using two dummy variables. In my variables list, i have Human Capital Index (HCI). If i use log of Human Capital Index (HCI), it seems to produce better estimation in all aspects. In panel unit root tests, most variables are I(0) and some are I(1).

My questions are:

1. Do i need to run unit root test for dummy variables?
2. Can i take log of Human Capital Index (HCI) variable?
3. If i want to use FE model, should i take first difference or just run regression on level value?

It will be of immense help for me, if you respond. Thanks in advance.