Re: [Gretl-users] dummy variables
On Mon, 16 Nov 2009, Data Analytics Corp. wrote:
I have two question on how to create dummy variables. Suppose...
1. I have a time series of annual data, say from 1970 to 2000. How
do I create a dummy for the period 1975-1980 (1 for this period, 0
otherwise)?
Ignacio answered that one.
2. I have cross-sectional data by states. How do I create a
dummy
for the southern region where I know the states in that region?
Basically the same answer, except that the Southern states might
not be contiguous in your dataset, in which case you'd need
something like
series south = (obs==1 || obs==2 || obs==5 ...)
Or, if the dataset contains state labels for the observations,
series south = (obs=="AL" || obs=="SC" || ...)
Allin Cottrell