glm repeated question

[Maguin, Eugene]

I’m seeing something unexpected in the glm repeated ouput. Consider this statement:

Glm y1 y2 with x1 x2/wsfactor=yname 2 polynominal.

 

The tests output table shows rows for yname, yaname*x1 yname*x2.

The design note for those tables shows intercept + x1 + x2 and wsdesign: yname.

 

So suppose the tests output table shows yname*x2 to be significant. What’s confusing to me is this: ‘*’ is used in design/wsdesign statements to define interactions and (I may be misremembering) that symbol would appear in the tests table had an interaction term been specified. Here, however, no interaction has been specified. What is the purpose of the ‘*’ in this context?

 

Thanks, Gene Maguin

 

 

===================== 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

[Kirill Orlov]

Interactions between within-subject factors and between-subject factors (including continuous covariates here too) are considered part of within-subject univariate and multivariate effects. SPSS always includes the interactions between within-subject factors and between-subject factors in the analysis irrespective whether you omit them in syntax or specify them explicitly as

Glm y1 y2 with x1 x2
 /wsfactor=yname 2 polynominal.
 /WSDESIGN=yname yname*x1 yname*x2.

If you want to exclude all or partly the interaction between the within and between-subjects factors restructure your data into the "long format" (from your current "wide format") and do GLM as univariate, "split-plot" approach (this is what is equivalent to repeated measures) or use MIXED.

This is what I've learned from experimenting, and it is likely to be so.

25.02.2019 19:53, Maguin, Eugene пишет:

I’m seeing something unexpected in the glm repeated ouput. Consider this statement:

Glm y1 y2 with x1 x2/wsfactor=yname 2 polynominal.

 

The tests output table shows rows for yname, yaname*x1 yname*x2.

The design note for those tables shows intercept + x1 + x2 and wsdesign: yname.

 

So suppose the tests output table shows yname*x2 to be significant. What’s confusing to me is this: ‘*’ is used in design/wsdesign statements to define interactions and (I may be misremembering) that symbol would appear in the tests table had an interaction term been specified. Here, however, no interaction has been specified. What is the purpose of the ‘*’ in this context?

 

Thanks, Gene Maguin

 

 

===================== 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

[Bruce Weaver]

Kirill's post prompted me to go find one of my own old syntax files from a
dozen years ago.  What I found pretty much agrees with what Kirill says
below.  But note that when you restructure to LONG data and use UNIANOVA, if
there is more than one error term, you will have to specify some of the
error terms explicitly (all but the last one) to get what you're expecting.
Here are some comments from my old 2007 syntax that explain what I mean in a
bit more detail.  HTH.


* ----------------------------- .
* ERROR TERMS MUST BE SPECIFIED .
* ----------------------------- .

* Using the LONG file format and UNIANOVA is more flexible than
* GLM REPEATED MEASURES, but that flexibility comes at a price.
* The price is that you have to specify ALL of the error terms to
* be included in the model--actually, all but the last error term,
* which will be the residual.

* If you do not specify all of the error terms (except the last),
* SPSS will use a pooled error term, which is most likely not
* what you want to do.

* For example, the following syntax is for a mixed-design ANOVA
* with A as a between-Ss factor, and B and C within-Ss.  Each
* subject had 6 rows of data, one for each of the 6 B*C combinations.
* The dependent variable was Y.

*UNIANOVA
  y  BY subj a b c
  /RANDOM = subj
  /METHOD = SSTYPE(3)
  /INTERCEPT = INCLUDE
  /CRITERIA = ALPHA(.05)
  /print = etasq
  /plot = resid
  /DESIGN = a subj(a)
                b b*a b*subj(a)
                c c*a c*subj(a)
                b*c b*c*a .

* The notation "subj(a)" means "subjects nested within A".
* The final error term, which I've omitted is: b*c*subj(a) .
* This will show up as the residual, and will be used appropriately.




Kirill Orlov wrote

> Interactions between within-subject factors and between-subject factors
> (including continuous covariates here too) are considered part of
> within-subject univariate and multivariate effects. SPSS always includes
> the interactions between within-subject factors and between-subject
> factors in the analysis irrespective whether you omit them in syntax or
> specify them explicitly as
>
> Glm y1 y2 with x1 x2
>   /wsfactor=yname 2 polynominal.
>   /WSDESIGN=yname yname*x1 yname*x2.
>
> If you want to exclude all or partly the interaction /between /the
> within and between-subjects factors restructure your data into the "long
> format" (from your current "wide format") and do GLM as univariate,
> "split-plot" approach (this is what is equivalent to repeated measures)
> or use MIXED.
>
> This is what I've learned from experimenting, and it is likely to be so.

-----
--
Bruce Weaver
[hidden email]
http://sites.google.com/a/lakeheadu.ca/bweaver/

"When all else fails, RTFM."

NOTE: My Hotmail account is not monitored regularly.
To send me an e-mail, please use the address shown above.

--
Sent from: http://spssx-discussion.1045642.n5.nabble.com/

=====================
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