Re: [Gretl-devel] difftest
Am 29.11.2017 um 14:24 schrieb Allin Cottrell:
On Wed, 29 Nov 2017, Sven Schreiber wrote:
>>> 3) Could the parametric t-test of differing means in
(paired)
>>> samples be subsumed under difftest? (Currently there is no direct
>>> scripting way of performing it, I think.)
The paired two-means t-test is trivial in hansl:
d = x - y
ols d const
Yes of course, what I wrote was misleading. Indeed this approach would
have the straightforward extension with --robust for example.
It just seems to me that the 'difftest' label by its inherent semantics
very naturally would encompass the standard two-sample t-test.
Also, anything which isn't a one-liner isn't "direct" enough for me ;-)
After all, that's why it exists in the test calculator in the first
place, instead of telling people to run an auxiliary regression. But I
admit that that would be defensible as a GUI-only thing.
cheers,
sven