question

[joost]

Hi,

my SPSS-file consists of rows and columns (obviously). every row is a
subject and every column is a variable. my subjects can be divided in 2
groups. i have 3 variables that belong to a "region" such that my table
looks like:

   region1_var1 region1_var2 region1_var3 region2_var1 region2_var2 ...etc.
subject1
subject2
subject3
.
.
.
etc.

i thus have a between-subjects factor: group and (i think) 1 within-subjects
factor: region.

i would like to do a test that tests whether there is an effect of region
for my 2 groups. this test -if significant- may allow me to
'post-hoc'-assess the regions separately. can this be done using the
Repeated Measures analysis under the General Linear Model?

thanks,
-joost

=====================
To manage your subscription to SPSSX-L, send a message to
[hidden email] (not to SPSSX-L), with no body text except the
command. To leave the list, send the command
SIGNOFF SPSSX-L
For a list of commands to manage subscriptions, send the command
INFO REFCARD

[Maguin, Eugene]

Joost,

It may be just how you named your variables and my assumption that 'subjects' are persons but I'm having difficulty understanding how you have a repeated measures design. I don't understand what is repeated. Is var1, var2 and var3 the repeated element within a region or is region1, region2, etc the repeated element for var1. Perhaps somebody else understands this better than I do.

Gene Maguin



-----Original Message-----
From: SPSSX(r) Discussion [mailto:[hidden email]] On Behalf Of joost
Sent: Tuesday, March 27, 2012 11:29 AM
To: [hidden email]
Subject: question

Hi,

my SPSS-file consists of rows and columns (obviously). every row is a
subject and every column is a variable. my subjects can be divided in 2
groups. i have 3 variables that belong to a "region" such that my table
looks like:

   region1_var1 region1_var2 region1_var3 region2_var1 region2_var2 ...etc.
subject1
subject2
subject3
.
.
.
etc.

i thus have a between-subjects factor: group and (i think) 1 within-subjects
factor: region.

i would like to do a test that tests whether there is an effect of region
for my 2 groups. this test -if significant- may allow me to
'post-hoc'-assess the regions separately. can this be done using the
Repeated Measures analysis under the General Linear Model?

thanks,
-joost

=====================
To manage your subscription to SPSSX-L, send a message to
[hidden email] (not to SPSSX-L), with no body text except the
command. To leave the list, send the command
SIGNOFF SPSSX-L
For a list of commands to manage subscriptions, send the command
INFO REFCARD

=====================
To manage your subscription to SPSSX-L, send a message to
[hidden email] (not to SPSSX-L), with no body text except the
command. To leave the list, send the command
SIGNOFF SPSSX-L
For a list of commands to manage subscriptions, send the command
INFO REFCARD

[David Marso]

Typically my brain processes region as living location.  Region as a within subject variable therefore is an alien concept.   I contend that Joost needs to better explain what his/her data are if anyone is likely to be of help aside from making ESPeculations!

Maguin, Eugene wrote

Joost,

It may be just how you named your variables and my assumption that 'subjects' are persons but I'm having difficulty understanding how you have a repeated measures design. I don't understand what is repeated. Is var1, var2 and var3 the repeated element within a region or is region1, region2, etc the repeated element for var1. Perhaps somebody else understands this better than I do.

Gene Maguin



-----Original Message-----
From: SPSSX(r) Discussion [mailto:[hidden email]] On Behalf Of joost
Sent: Tuesday, March 27, 2012 11:29 AM
To: [hidden email]
Subject: question

Hi,

my SPSS-file consists of rows and columns (obviously). every row is a
subject and every column is a variable. my subjects can be divided in 2
groups. i have 3 variables that belong to a "region" such that my table
looks like:

   region1_var1 region1_var2 region1_var3 region2_var1 region2_var2 ...etc.
subject1
subject2
subject3
.
.
.
etc.

i thus have a between-subjects factor: group and (i think) 1 within-subjects
factor: region.

i would like to do a test that tests whether there is an effect of region
for my 2 groups. this test -if significant- may allow me to
'post-hoc'-assess the regions separately. can this be done using the
Repeated Measures analysis under the General Linear Model?

thanks,
-joost

=====================
To manage your subscription to SPSSX-L, send a message to
[hidden email] (not to SPSSX-L), with no body text except the
command. To leave the list, send the command
SIGNOFF SPSSX-L
For a list of commands to manage subscriptions, send the command
INFO REFCARD

=====================
To manage your subscription to SPSSX-L, send a message to
[hidden email] (not to SPSSX-L), with no body text except the
command. To leave the list, send the command
SIGNOFF SPSSX-L
For a list of commands to manage subscriptions, send the command
INFO REFCARD

[Rich Ulrich]

As I read it, potentially, you have two within factors. 

According to your example, for each Region (I think of,
Brain-region), you have 3 variables.  The number of Regions
is not stated.  You can have a simple repeated measures for
one factor, Regions, if you analyze each variable separately.

Post-hoc tests for repeated measures are best performed as
paired t-tests; and you make your own adjustments after the
test.  Certain stat-packages have offered analogs to HST,
etc., for post-hocs, but they have tough assumptions.  If you
know enough to justify using them, you won't be asking here.

If the 3 variables are "the same variable in three locations"
then you have a doubly-repeated measures, if you want to
analyze them together.  If the 3 variables are various, then
you have MANOVA design.  That one will be, most likely,
extremely difficult to make good sense of, and it is probably
something to avoid.

--
Rich Ulrich

> Date: Tue, 27 Mar 2012 11:29:15 -0400

> From: [hidden email]
> Subject: question
> To: [hidden email]
>
> Hi,
>
> my SPSS-file consists of rows and columns (obviously). every row is a
> subject and every column is a variable. my subjects can be divided in 2
> groups. i have 3 variables that belong to a "region" such that my table
> looks like:
>
> region1_var1 region1_var2 region1_var3 region2_var1 region2_var2 ...etc.
> subject1
> subject2
> subject3
> .
> .
> .
> etc.
>
> i thus have a between-subjects factor: group and (i think) 1 within-subjects
> factor: region.
>
> i would like to do a test that tests whether there is an effect of region
> for my 2 groups. this test -if significant- may allow me to
> 'post-hoc'-assess the regions separately. can this be done using the
> Repeated Measures analysis under the General Linear Model?
>

[...]