.
Briefly, I am conducting research dealing with different ways of
presenting medical data to clinicians. One aspect I am interested
in is the time it takes for clinicians to reach decisions when
using these displays, and whether different ways of presenting the
same data take longer or shorter times than others.
My experimental design consists of a series of cases (i.e., data
taken from individual patients), each shown in different display
formats (i.e., modalities), to a variety of clinician readers, who
then decide on what to do with that patient. *Every* case is shown
in *every* modality to *every* reader. If I'm using the correct
terminology, this is a complete, balanced design, with no nesting,
blocking, etc. For each reader, the order of cases/modalities is
randomized. Different data are collected from each combination of
case, modality, and reader, including the time it takes for
decisions to be rendered in each case/modality/reader combination.
If I understand,
Case: fixed within subject effect
Display_format: fixed within subject effect
Reader: random subject effect
See the Stata FAQ for this design:
http://www.stata.com/support/faqs/stat/anova2.html#half713
2) If I wanted to analyze this as a mixed model, what is an
appropriate way to use xtmixed? I consider case and reader to be
random effects, as both are drawn from larger populations of
possible cases and readers. Likewise, I consider modality to be a
fixed factor, as there are only N modalities of interest in any
particular phase of my study. As such, I came up with the following
as a rough guess of how to express it:
There seems a discrepancy between the randomized block factorial
design anova solution given at UCLA and that given in the Stata FAQ.
Unless the designs are different. I can't tell from the data table.
http://www.stata.com/support/faqs/stat/anova2.html#half713
--------------------------------------------------------------------
| 10 minute time periods and dial
| ------- 1 ------ ------- 2 ------ ------- 3
------
subject | 1 2 3 1 2 3 1
2 3
----------
+---------------------------------------------------------
1 | 45 53 60 40 52 57 28
37 46
2 | 35 41 50 30 37 47 25
32 41
3 | 60 65 75 58 54 70 40
47 50
--------------------------------------------------------------------
. anova score subject period / subject*period dial / subject*dial
period*dial, repeated(period dial)
or
. anova y s a / s*a b / s*b a*b, repeated(a b)
versus
http://www.ats.ucla.edu/stat/stata/faq/xtmixed.htm
----------+---------------------------------------------------------
| a and b
| ------- 1 ------ ------- 2 ------ ------- 3 ------
s | 1 2 3 1 2 3 1 2 3
----------+---------------------------------------------------------
1 | 37 43 48 39 35 46 31 41 64
2 | 42 44 47 30 40 36 21 50 52
3 | 33 36 29 34 31 45 20 39 53
4 | 29 27 38 26 22 27 18 36 42
5 | 24 25 28 21 27 26 10 34 49
----------+---------------------------------------------------------
/
************************************************************************
******************/
/
*
*/
/* randomized block factorial
design */
/
*
*/
/* from kirk -- see http://www.ats.ucla.edu/stat/stata/examples/kirk/
kirkstata10.htm */
/
*
*/
/
************************************************************************
******************/
use http://www.ats.ucla.edu/stat/stata/examples/kirk/rbf33, clear
/* anova code: */ anova y s a b a*b, repeated(a b)
xi3 e.a*e.b
xtmixed y _Ia_2 _Ia_3 _Ib_2 _Ib_3 _Ia2Xb2 _Ia2Xb3 _Ia3Xb2 _Ia3Xb3 || s:
test _Ia_2 _Ia_3
/* scale chi^2 as F-ratio to compare with anova */
display r(chi2)/r(df)
test _Ib_2 _Ib_3
display r(chi2)/r(df)
test _Ia2Xb2 _Ia2Xb3 _Ia3Xb2 _Ia3Xb3
display r(chi2)/r(df)
*
* For searches and help try:
* http://www.stata.com/support/faqs/res/findit.html
* http://www.stata.com/support/statalist/faq
* http://www.ats.ucla.edu/stat/stata/
