Stata The Stata listserver
[Date Prev][Date Next][Thread Prev][Thread Next][Date index][Thread index]

Re: st: SP/RM[1,2] EMS


From   David Airey <david.airey@vanderbilt.edu>
To   statalist@hsphsun2.harvard.edu
Subject   Re: st: SP/RM[1,2] EMS
Date   Sun, 24 Aug 2003 01:16:24 -0500

Here I provide ANOVA F tests for various split plot repeated measures
ANOVAs I was considering, where there are 7 potential factors.

A = animal (animal is random, nested in strain or strain*sex)

between-animal fixed effects(levels):
S = strain(20), F = sex(2)

within-animal fixed effects(levels):
D = drug(3), T = trial set(4), P = prepulse(2), I = isi(2)
D, T, P, and I are crossed within subject
P and I are stimulus characteristics aiding response
D order is randomized by animal to deal with carry over problems

The study is a survey of mouse strain (genetic) differences in response to
drugs, thus A(?), S(20), and D(3) are necessary for any model. Because
strain has 20 levels, it is hard to make more complex designs
work within the matsize limit of 11,000. I worked out the required matsize
assuming either 4 or 8 animals per strain*sex combination (160 or 320 animals),
suggesting only a SPRM[1,2] design with a matsize of 2-4K is possible.
I could just manipulate P rather than P and I. It's also possible
to average over the 4 5-minute trial sets, dropping out T. It's also possible
to do separate ANOVAs for each P*I combination. Also, existing
literature does not suggest strong sex effects. That leaves S, A, and D, and P.
I could assume higher order interactions are not significant too. These
are not the most desirable choices. I just did not realize complicated
ANOVA designs are so computationally intensive in large data sets. How annoying.

Anybody have design suggestions to reduce problems with matsize? Note,
the within subject design was chosen because of individual differences and an
inability to get more than 20 animals per strain.

SP/RM[1,1] SP/RM[1,2] SP/RM[1,3] SP/RM[1,4]
A, S, D A, S, D, T A, S, D, T, P A, S, D, T, P, I
- 20 3 - 20 3 4 - 20 3 4 2 - 20 3 4 2 2
------------------------------------------------------------------------ -----
S / A|S S / A|S S / A|S S / A|S

D / DA|S D / DA|S D / DA|S D / DA|S
SD / DA|S SD / DA|S SD / DA|S
SD / E
T / TA|S T / TA|S T / TA|S
ST / TA|S ST / TA|S ST / TA|S

DT / DTA|S P / PA|S P / PA|S
SP / PA|S SP / PA|S
SDT / E
DT / DTA|S I / IA|S
SDT / DTA|S SI / IA|S

DP / DPA|S DT / DTA|S
SDP / DPA|S SDT / DTA|S

TP / TPA|S DP / DPA|S
STP / TPA|S SDP / DPA|S

DTP / DTPA|S DI / DIA|S
SDI / DIA|S
SDTP / E
TP / TPA|S
STP / TPA|S

TI / TIA|S
STI / TIA|S

PI / PIA|S
SPI / PIA|S

DTP / DTPA|S
SDTP / DTPA|S

DTI / DTIA|S
SDTI / DTIA|S

DPI / DPIA|S
SDPI / DPIA|S

TPI / TPIA|S
STPI / TPIA|S

DTPI / DTPIA|S

SDTPI / E


SD/RM[2,1] SD/RM[2,2] SD/RM[2,3] SD/RM[2,4]
A, S, F, D A, S, F, D, T A, S, F, D, T, P A, S, F, D, T, P, I
- 20 2 3 - 20 2 3 4 - 20 2 3 4 2 - 20 2 3 4 2 2
------------------------------------------------------------------------ -----
S / A|SF S / A|SF S / A|SF S / A|SF
F / A|SF F / A|SF F / A|SF F / A|SF
SF / A|SF SF / A|SF SF / A|SF SF / A|SF

D / DA|SF D / DA|SF D / DA|SF D / DA|SF
SD / DA|SF SD / DA|SF SD / DA|SF SD / DA|SF
FD / DA|SF FD / DA|SF FD / DA|SF FD / DA|SF
SFD / DA|SF SFD / DA|SF SFD / DA|SF SFD / DA|SF

T / TA|SF T / TA|SF T / TA|SF
ST / TA|SF ST / TA|SF ST / TA|SF
FT / TA|SF FT / TA|SF FT / TA|SF
SFT / TA|SF SFT / TA|SF SFT / TA|SF

DT / DTA|SF P / PA|SF P / PA|SF
SDT / DTA|SF SP / PA|SF SP / PA|SF
FDT / DTA|SF FP / PA|SF FP / PA|SF
SFP / PA|SF SFP / PA|SF
SFDT / E
DT / DTA|SF I / IA|SF
SDT / DTA|SF SI / IA|SF
FDT / DTA|SF FI / IA|SF
SFDT / DTA|SF SFI / IA|SF

DP / DPA|SF DT / DTA|SF
SDP / DPA|SF SDT / DTA|SF
FDP / DPA|SF FDT / DTA|SF
SFDP / DPA|SF SFDT / DTA|SF

TP / TPA|SF DP / DPA|SF
STP / TPA|SF SDP / DPA|SF
FTP / TPA|SF FDP / DPA|SF
SFTP / TPA|SF SFDP / DPA|SF

DTP / DTPA|SF DI / DIA|SF
SDTP / DTPA|SF SDI / DIA|SF
FDTP / DTPA|SF FDI / DIA|SF
SFDI / DIA|SF
SFDTP / E
TP / TPA|SF
STP / TPA|SF
FTP / TPA|SF
SFTP / TPA|SF

TI / TIA|SF
STI / TIA|SF
FTI / TIA|SF
SFTI / TIA|SF

PI / PIA|SF
SPI / PIA|SF
FPI / PIA|SF
SFPI / PIA|SF

DTP / DTPA|SF
SDTP / DTPA|SF
FDTP / DTPA|SF
SFDTP / DTPA|SF

DTI / DTIA|SF
SDTI / DTIA|SF
FDTI / DTIA|SF
SFDTI / DTIA|SF

DPI / DPIA|SF
SDPI / DPIA|SF
FDPI / DPIA|SF
SFDPI / DPIA|SF

TPI / TPIA|SF
STPI / TPIA|SF
FTPI / TPIA|SF
SFTPI / TPIA|SF

DTPI / DTPIA|SF
SDTPI / DTPIA|SF
FDTPI / DTPIA|SF

SFDTPI / E

Matsize requirements [between, within effects]:

A = 4
[1,1] 403
[1,2] 2019
[1,3] 6059
[1,4] > 11K

[2,1] 891
[2,2] 4459
[2,3] > 11K
[2,4] > 11K

A = 8
[1,1] 723
[1,2] 3619
[1,3] 10859
[1,4] > 11K

[2,1] 1531
[2,2] 7659
[2,3] > 11K
[2,4] > 11K

-Dave

*
* 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/




© Copyright 1996–2014 StataCorp LP   |   Terms of use   |   Privacy   |   Contact us   |   What's new   |   Site index