You cannot select more than 25 topics
Topics must start with a letter or number, can include dashes ('-') and can be up to 35 characters long.
87 lines
3.2 KiB
Plaintext
87 lines
3.2 KiB
Plaintext
9 months ago
|
.-
|
||
|
|
help for ^wsanova^ (STB-47: sg103)
|
||
|
|
.-
|
||
|
|
|
||
|
|
within subjects ANOVA, with zero or more between subjects factors
|
||
|
|
-----------------------------------------------------------------
|
||
|
|
|
||
|
|
^wsanova^ yvar wfact [weight] [^if^ exp] [^in^ range] , ^id(^svar^)^
|
||
|
|
[^bet^ween^(^beffects^) wo^nly^(^weffects^) eps^ilon ^nom^atr ]
|
||
|
|
|
||
|
|
^aweights^ and ^fweights^ are allowed; see help @weights@.
|
||
|
|
|
||
|
|
|
||
|
|
Description
|
||
|
|
-----------
|
||
|
|
|
||
|
|
^wsanova^ performs a within subjects (repeated measures) analysis of variance for
|
||
|
|
the response variable yvar classified by the within subjects factor wfact. The
|
||
|
|
subjects are identified by the variable svar, and may be classified by one or
|
||
|
|
more between subjects factors. Within subjects F tests can be adjusted for lack
|
||
|
|
of sphericity using the Greenhouse-Geisser or Huynh-Feldt correction factor.
|
||
|
|
|
||
|
|
|
||
|
|
Options
|
||
|
|
-------
|
||
|
|
|
||
|
|
^id(svar)^ declares that the variable svar uniquely identifies each subject to be
|
||
|
|
used in the analysis. [Not optional.]
|
||
|
|
|
||
|
|
^between(beffects)^ supplies a list of between subjects factors that classify the
|
||
|
|
subjects, along with zero or more of their interactions. Up to 7 such fac-
|
||
|
|
tors can be used; their interactions must be explicitly requested.
|
||
|
|
|
||
|
|
^wonly(weffects)^ selects within subjects effects that should be included in the
|
||
|
|
analysis. By default, wfact and all of its interactions with the elements
|
||
|
|
of the ^between^ option are included.
|
||
|
|
|
||
|
|
^epsilon^ requests that p-values for within subjects F-tests be adjusted for lack
|
||
|
|
of sphericity using the Greenhouse-Geisser and Huynh-Feldt adjustment fac-
|
||
|
|
tors. [Requires that yvar be non-missing for each subject.]
|
||
|
|
|
||
|
|
^nomatr^ discards covariance and cell mean matrices at exit. The ^epsilon^ option
|
||
|
|
creates a (pooled) covariance matrix WSAoV_ and a matrix WSAov_ of marginal
|
||
|
|
means of yvar across levels of wfact. In addition, when there are between
|
||
|
|
subjects factors epsilon creates a cell means matrix and a within group co-
|
||
|
|
variance matrix for each distinct group of subjects; these matrices will be
|
||
|
|
named WSAov1 WSAoV1, WSAov2 WSAoV2, ... . By default, all these matrices
|
||
|
|
are left in memory; ^nomatr^ erases each of them before wsanova exits.
|
||
|
|
|
||
|
|
|
||
|
|
Examples
|
||
|
|
--------
|
||
|
|
|
||
|
|
. ^wsanova lhist time, id(dog)^
|
||
|
|
. ^predict resid, res^
|
||
|
|
(One-way repeated measures ANOVA; save residuals in resid)
|
||
|
|
|
||
|
|
. ^wsanova rtime trial, id(subj) between(age) epsilon^
|
||
|
|
(Split-plot ANOVA, with subjects grouped by levels of age; adjust trial and
|
||
|
|
age*trial F-tests for lack of sphericity)
|
||
|
|
|
||
|
|
. ^wsanova rtime trial, id(subj) bet(age sex age*sex)^
|
||
|
|
(Traditional analysis of a "two between, one within" design: age, sex, and
|
||
|
|
age*sex as between effects; trial, trial*age, trial*sex, trial*age*sex as
|
||
|
|
within effects; assume sphericity)
|
||
|
|
|
||
|
|
. ^wsanova rtime trial, id(subj) bet(age sex) wonly(trial) eps nomatr^
|
||
|
|
(Main effects only version of the last example; adjust the trial F-test for
|
||
|
|
lack of sphericity, and discard the matrices used)
|
||
|
|
|
||
|
|
|
||
|
|
Author
|
||
|
|
------
|
||
|
|
|
||
|
|
John R. Gleason
|
||
|
|
Syracuse University
|
||
|
|
Syracuse NY, USA
|
||
|
|
loesljrg@@ican.net
|
||
|
|
|
||
|
|
|
||
|
|
Also see
|
||
|
|
--------
|
||
|
|
|
||
|
|
STB: STB-47 sg103
|
||
|
|
Manual: [R] anova; [R] predict
|
||
|
|
On-line: ^help^ for @anova@, @predict@
|