[R] Question : simtest result

[Chihiro Kuroki]

Dear Mr.Torsten:

At Wed, 16 Jun 2004 08:04:08 +0200 (CEST),
Torsten Hothorn wrote:
> > Call:
> > simtest.formula(formula = y4 ~ f2, data = dat2, type = "Dunnett")
> >
> > 	 Dunnett contrasts for factor f2
> >
> > Contrast matrix:
> >           f21 f22 f23 f24 f25
> > f22-f21 0  -1   1   0   0   0
> > f23-f21 0  -1   0   1   0   0
> > f24-f21 0  -1   0   0   1   0
> > f25-f21 0  -1   0   0   0   1
> >
> >
> > Absolute Error Tolerance:  0.001
> >
> > Coefficients:
> >         Estimate t value Std.Err. p raw p Bonf p adj
> > f25-f21    5.167  -4.644    1.022 0.000  0.000 0.000
> > f23-f21    2.875  -2.813    1.022 0.008  0.024 0.022
> > f24-f21    2.625  -2.569    1.022 0.015  0.029 0.028
> > f22-f21    2.125  -2.079    1.113 0.045  0.045 0.045
> > ---------------------------------
> 
> Chihiro,
> 
> Frank and I used your data to check the program and example with an
> independent
> algorithm and implementation (Westfall-Young stepdown resampling
> procedure). Theory suggests that the
> results should be similar to (but not necessarily the same as) those
> obtained with multcomp in this special case. These are the adjusted
> p-values obtained with the Westfall-Young approach for 100,000
> replications:
> 
> which fit nicely with the ones obtained from multcomp.
> 
> Hope this helps & sorry for the delay,

Thank you.

At Mon, 07 Jun 2004 12:14:26 +0900,
myself-oak wrote:
> dunnett(dat2,1,2)
> rho=0.426
> group:5 t=4.644 p=0.000
> group:3 t=2.813 p=0.028
> group:4 t=2.569 p=0.051 --- (B)
> group:2 t=2.079 p=0.145
> (sorted in order of p values.)
> 
> p values are different although t values are equal.
> 
> > > I got the following inequality from the appended chart of a
> > > book.
> > >
> > 
> > hm, without knowing what
> > 
> > > 2.558 < d(5, 35, 0.4263464, 0.05) < 2.598
> > 
> > means it is hard to tell what the problem is. Could you please explain it
> > further?
> 
> The alternative hypothesis is "two sided".
> 
> When significant level is equal to 0.05 , number of groups=5,
> df of error=35 and rho=0.426, I think that absolute t-value
> should be between 2.558 and 2.598.
> 
> So, (B) is easy to understand for me than (A).

You said "adj p" values are ...

> 0.0437
> 0.0260
> 0.0204
> 0.0001

I used SPSS ver.10 and got the following result.

Dunnett t (two sided) 

 | --------------- | ---------- | -------- | -------- | ------------------ | 
 |                 | mean diff. | s.e.     | adj p    | 95% C.I.           | 
 | ------ | ------ | (I-J)      |          |          | ---------- | ----- | 
 | (I) V3 | (J) V3 |            |          |          |            |       | 
 | ------ | ------ | ---------- | -------- | -------- | ---------- | ----- | 
 | 2      | 1      | 2.125      | 1.022    | .145     | -.507      | 4.757 | 
 | ------ | ------ | ---------- | -------- | -------- | ---------- | ----- | 
 | 3      | 1      | 2.875(*)   | 1.022    | .028     | .243       | 5.507 | 
 | ------ | ------ | ---------- | -------- | -------- | ---------- | ----- | 
 | 4      | 1      | 2.625      | 1.022    | .051     | -7.325E-03 | 5.257 | 
 | ------ | ------ | ---------- | -------- | -------- | ---------- | ----- | 
 | 5      | 1      | 5.167(*)   | 1.113    | .000     | 2.301      | 8.032 | 
 | ------ | ------ | ---------- | -------- | -------- | ---------- | ----- | 

I might make a mistake in the way to use the simtest()...How should I think? 

Best regards,
-- 
kuroki
GnuPG fingerprint = 90FD FE79 905F 26F9 29C4  096F 8AA2 2C42 5130 1469