Re: Multiple single-sample t-tests
- From: Ray Koopman <koopman@xxxxxx>
- Date: Fri, 12 Dec 2008 14:55:23 -0800 (PST)
On Dec 12, 4:32 am, "Et. Gaudrain" <egaudr...@xxxxxxxxx> wrote:
Here are epsilon values:
- A: GG=0.557, HF=0.798, the p-value is <0.001 with and without
correction
- B: GG=0.894, HF=0.987, p=0.270, pGG=0.270, pHH=0.270
- A*B: GG=0.392, HF=0.670, p=0.366, pGG=0.377, pHF=0.374
n = 19 ?
Yes, indeed, A is actually a continuous factor. The pattern of X as a
function of A is a kind of psychometric function with a ceiling at .5
for low values of A and a floor a 0, for high values of A. We're
interested in some point just before the function drops from .5 to 0.
Here (well we did not exactly new where it would be), we expect a
hump>.5, and that's what we want to test: Is there really a hump >.5
or not? So we also considered looking at the values >.5 and testing
if their are significantly different from .5. So we would only test
3 or 4 values... and then not use Bonferroni correction...? not too
convincing...
Right. You would have to start the stepwise Bonferroni with a
correction factor of 10.
What are X and A? What kind of function do you think the regression
of X on A might be? The goal would be to construct a Scheffe-style
confidence region for the regression.
.
- Follow-Ups:
- Re: Multiple single-sample t-tests
- From: Et. Gaudrain
- Re: Multiple single-sample t-tests
- References:
- Multiple single-sample t-tests
- From: Etienne Gaudrain
- Re: Multiple single-sample t-tests
- From: RichUlrich
- Re: Multiple single-sample t-tests
- From: Et. Gaudrain
- Re: Multiple single-sample t-tests
- From: Ray Koopman
- Re: Multiple single-sample t-tests
- From: Et. Gaudrain
- Re: Multiple single-sample t-tests
- From: Ray Koopman
- Re: Multiple single-sample t-tests
- From: Et. Gaudrain
- Multiple single-sample t-tests
- Prev by Date: Re: Dealing with correlated predictors
- Next by Date: Re: Multiple single-sample t-tests
- Previous by thread: Re: Multiple single-sample t-tests
- Next by thread: Re: Multiple single-sample t-tests
- Index(es):
Relevant Pages
|