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.
69 lines
2.4 KiB
Plaintext
69 lines
2.4 KiB
Plaintext
7 months ago
|
.-
|
||
|
|
help for ^iclassr^, ^l1way^
|
||
|
|
.-
|
||
|
|
|
||
|
|
One-way ANOVA, random effects, and intra-class correlation [STB-35 sg65]
|
||
|
|
----------------------------------------------------------
|
||
|
|
|
||
|
|
^iclassr^ response_var group_var [weight] [^if^ exp] [^in^ range]
|
||
|
|
[^, c^enter^(^Fpos^) e^ms ^noi^sily]
|
||
|
|
|
||
|
|
^l1way^ response_var group_var [weight] [^if^ exp] [^in^ range]
|
||
|
|
[^, c^enter^(^Fpos^) e^ms]
|
||
|
|
|
||
|
|
|
||
|
|
^aweight^s are allowed; see help @weights@.
|
||
|
|
|
||
|
|
|
||
|
|
Description
|
||
|
|
-----------
|
||
|
|
|
||
|
|
^l1way^ and ^iclassr^ estimate one-way analysis of variance (ANOVA) models and use
|
||
|
|
the random effects ANOVA model to estimate the intra-class (within group) cor-
|
||
|
|
relation, and the reliability of group averages from the Spearman-Brown proph-
|
||
|
|
ecy formula.
|
||
|
|
|
||
|
|
^l1way^ and ^iclassr^ are replacements for @loneway@; ^iclassr^ is faster than ^l1way^
|
||
|
|
but limited to fewer than 376 values of group_var. Both are substantially fast-
|
||
|
|
er than ^loneway^, especially ^iclassr^, and both offer more accurate estimation of
|
||
|
|
the intra-class correlation in unbalanced designs or in weighted analyses; see
|
||
|
|
the option ^ems^, below.
|
||
|
|
|
||
|
|
|
||
|
|
Options
|
||
|
|
-------
|
||
|
|
|
||
|
|
^center^ chooses a reference point in the F distribution to center the observed
|
||
|
|
F statistic in estimating the intra-class correlation. The argument Fpos
|
||
|
|
may be 1 (the default), or "med" or "mean" to choose the median or the mean
|
||
|
|
of the appropriate F distribution. This option has little effect except in
|
||
|
|
small samples.
|
||
|
|
|
||
|
|
^ems^ uses the coefficient of the group variance in the expected value of the be-
|
||
|
|
tween group mean square to scale the estimated intra-class correlation. The
|
||
|
|
default, for compatibility with ^loneway^, is to use the average group size
|
||
|
|
instead. This option is ^very^ strongly recommended unless the design is
|
||
|
|
nearly balanced.
|
||
|
|
|
||
|
|
^noisily^ causes ^iclassr^ to display the ANOVA summary table.
|
||
|
|
|
||
|
|
|
||
|
|
Examples
|
||
|
|
--------
|
||
|
|
|
||
|
|
. ^iclassr rating stimulus^ (intra-stimulus r, < 376 stimuli, equal n's)
|
||
|
|
. ^l1way rating stimulus^ (intra-stimulus r, ANOVA table, equal n's)
|
||
|
|
|
||
|
|
. ^iclassr trait family, ems^ (intra-family r, < 376 families, unequal n's)
|
||
|
|
. ^l1way trait family, e c(med)^ (intra-family r, unequal n's, center the ob-
|
||
|
|
served F on the median of F)
|
||
|
|
|
||
|
|
|
||
|
|
Also see
|
||
|
|
--------
|
||
|
|
|
||
|
|
STB: sg65 (STB-35)
|
||
|
|
Manual: ^[R] loneway, oneway^
|
||
|
|
On-line: help for @loneway@, @oneway@
|
||
|
|
|