Repeated Measures ANOVA

[ms]

I have a question regarding which statistic I should use for my
design. I have 2 groups (men and women) and am testing a single trial
of wrist circumference on their right arm and on their left arm. So I
have two independent variables (GENDER and SIDE). I believe I would
use ANOVA with Repeated Measures for SIDE (right and left) being the
Within-Subject Factor and GENDER as the Between-Subject Factor.
Currently, I have GENDER listed as a column of data (1s and 2s coded
for men and women) and then the Right Wrist Circumference as another
column and Left Wrist Circumference as its own column (so a total of 3
columns). Please let me know if you think I am correct, or if there
is anything else I would do differently. Thanks!

[Ryan]

I'm not really sure what you're trying to test here, but if you're
running a mixed ANOVA with one b/n Ss variable (i.e. Gender) and one w/
n Ss variable (i.e. Side), then the following data set structure
should be used:

ID Gender Left_Side Right_Side
1 1 23 22
2 2 45 22
3 1 16 15
.
.
.
N

If you have data that are missing at random, then you should probably
consider running a linear mixed model. In such a case you would need
to set up your data set such that ID repeats for each level of your
within-subjects variable:

ID Gender Side Y
1 1 1 23
1 1 2 22
2 2 1 45
2 2 2 22
3 1 1 16
3 1 2 15
.
.
.
N

HTH,

Ryan

[Michelle Ames]

You have two independent variables, but one is an independent measure
(Gender) while the other is a repeated measure (Side). In this case
you need to use a Mixed ANOVA. However, Side is only a repeated
measure if every participant was tested on both their right arm and
their left arm. If some people were tested on just the left arm and
others on just the right arm then you would need to run a two-way
independent ANOVA.

[John Reece]

Your data setup is correct, provided every person has measures taken of both
their right and left wrists. This design can be described in a number of
ways, such as a 2 x 2 split-plot factorial ANOVA. Some use the term "mixed"
factorial ANOVA to describe a design with a mix of between- and
within-subjects factors, but others don't like this, because the word
"mixed" has several interpretations.

John

"ms" <silve...@gmail.com> wrote in message
news:80edcf59-3e3d-4b64...@c37g2000prb.googlegroups.com...