On Wed, 7 Mar 2012, Data Analytics Corp. wrote: > Thanks. This will work for my current problem. But what if I have a very > large data set and the econ observations are scattered throughout the data > set? This won't work.
Put your string-valued categorical variable somewhere other than the first column, export the data as CSV, and open the CSV in gretl. Then gretl won't interpret that variable as an observation label, but will code it numerically. Allin Cottrell > > On 3/7/2012 3:54 PM, Allin Cottrell wrote: >> On Wed, 7 Mar 2012, Data Analytics Corp. wrote: >> >>> I have car data that I imported from excel. The first column of the >>> excel worksheet has a car designation label: econ or prestige. The >>> first 40 observations are econ cars and the next 40 are prestige cars. >>> In Gretl, the designation appears as a row label. I'd like to subset >>> the data to model the econ cars. I tried to create a dummy variable, >>> seg, as >>> >>> seg = (obs == "econ") >> >> Try seg = (obs < 41) >> >> Allin Cottrell >> >> >> > > -- Allin Cottrell Department of Economics Wake Forest University, NC
