Re: [Gretl-users] Sample selection issues
On Wed, 16 Nov 2016, George Matysiak wrote:
I'm using Version 2016d. Specifically, I have a dataset of 63
observations,
consisting of of 3 different categories each with 21 members. I have
defined a dummy for each sub-set i.e 3 dummies. So, I first select a sample
based on one of the dummies/Use dummy variable/ which gives me a sample of
21. I then select/Random sub-sample/ and receive the message Number of
observations to select (max 62) - it list 10 as the default. If, for
example, I enter 22 I get the message Invalid number of cases 22. If I
click ok again it reports dropped 41 observations. Clearly not appropriate.
Perhaps a few if conditions greater than/less than are required in the code
in order to captures the breaches in the conditions. Thanks.
Ah, after simulating exactly what you describe, I'm seeing the effect
you mention. Strange! But since I can replicate the problem the fix
is probably not far away.
Allin Cottrell