How to specify the nested variables for repeated measures in SPSS
epistat
Assume that i have two repeated measured variables, A with three
levels (A1,A2,A3) and B with two levels. (B1,B2).
How to correctly specify the two repeated measured variables in
SPSS?
(analyze->general linear model->repeated measures)
Q1: Variable B is nested under variable A.
That is to say, A1(B1,B2), A2(B1,B2), A3(B1,B2). I have tried many
methods, but fail to do that.
Q2: Variable B is not nested under variable A.
That is to say, (A1,A2,A3) (B1,B2).
I specify A(3) and B(2) in the first interface of "repeated
measures", and then found it shows that
there are 6 variables in the "Within-subjects variables (A,B)". But i
think that there should be only 5 variables, am i wrong?
Is this a bug? Or i have some misunderstanding on this method in
SPSS.
Any suggestions or help are greatly appreicated.
Bruce Weaver
> Dear all,
> Assume that i have two repeated measured variables, A with three
> levels (A1,A2,A3) and B with two levels. (B1,B2).
> How to correctly specify the two repeated measured variables in
> SPSS?
> (analyze->general linear model->repeated measures)
>
> Q1: Variable B is nested under variable A.
> That is to say, A1(B1,B2), A2(B1,B2), A3(B1,B2). I have tried many
> methods, but fail to do that.
If B is nested within A, I would write that as A1(B1,B2), A2(B3,B4), A3
(B5,B6). This makes it clear that the two levels of B within each
level of A are unique to that level of A.
> Q2: Variable B is not nested under variable A.
> That is to say, (A1,A2,A3) (B1,B2).
> I specify A(3) and B(2) in the first interface of "repeated
> measures", and then found it shows that
> there are 6 variables in the "Within-subjects variables (A,B)". But i
> think that there should be only 5 variables, am i wrong?
> Is this a bug? Or i have some misunderstanding on this method in
> SPSS.
> Any suggestions or help are greatly appreicated.
You won't be able to generate the syntax you need via the GUI. You
can generate a first draft via the GUI, but will then have to do some
editing. In SPSS syntax, "B nested within A" is written as B(A).
I've never run a model like this, so if I was in your shoes, I might
also try it via the MIXED procedure to see if I get the same (or at
least similar) results. MIXED requires a long file format (i.e., one
row per observation rather than one row per subject), so you'd have to
restructure the data first.
--
Bruce Weaver
bwe...@lakeheadu.ca
http://sites.google.com/a/lakeheadu.ca/bweaver/
"When all else fails, RTFM."
RichUlrich
wrote:
I don't think that you know what "nesting" indicates.
For repeated measures, with two factors A(3) and B(2),
there *will* be 6 variables.
Else, you seem to be thinking of two Repeated analyses,
one with 3 variables and another with 2. (What tells
you to combine them?)
--
Rich Ulrich