Logistic Regression with Interactions Categorical DV

[Arakashi]

I have a 2 (group: patients, controls) between subject X 2 (Task 1, Task 2) within subjects X 3 (Trial type 1, Trial type2 and Trial type3) within subjects.  My DV is Accuracy (binary, yes or no).

My data is not normally distributed because accuracy is at ceiling, so I was going to use logistic regression.  
However, I cannot for the life of me figure out what to do with the Accuracy data since it's categorical/binary.  I need to be able to look at the interactions…

For the reaction time data, I used the following syntax to explore the main effects and interactions (RT was for correct trials only):
MIXED Probe.RT WITH Condition Trial Group
/FIXED=Condition Trial Group Condition*Group Trial*Group Condition*Trial Condition*Trial*Group | SSTYPE(3)
/METHOD=reml
/PRINT = SOLUTION TESTCOV
/RANDOM=INTERCEPT | SUBJECT(Subject)

If anyone could help me with the syntax for the Accuracy data, I'd be so so grateful.

[Ryan]

You could employ a random effects logistic regression via the procedure GENLINMIXED or a GEE logistic regression model via the procedure GENLIN. The syntax is fairly straightforward, and there are many examples if you search the archives:

http://listserv.uga.edu/cgi-bin/wa?S1=spssx-l

or read the manual.

Ryan

On Tue, May 6, 2014 at 1:30 PM, Arakashi <[hidden email]> wrote:

I have a 2 (group: patients, controls) between subject X 2 (Task 1, Task 2)
within subjects X 3 (Trial type 1, Trial type2 and Trial type3) within
subjects.  My DV is Accuracy (binary, yes or no).

My data is not normally distributed because accuracy is at ceiling, so I was
going to use logistic regression.
However, I cannot for the life of me figure out what to do with the Accuracy
data since it's categorical/binary.  I need to be able to look at the
interactions…

For the reaction time data, I used the following syntax to explore the main
effects and interactions (RT was for correct trials only):
MIXED Probe.RT WITH Condition Trial Group
/FIXED=Condition Trial Group Condition*Group Trial*Group Condition*Trial
Condition*Trial*Group | SSTYPE(3)
/METHOD=reml
/PRINT = SOLUTION TESTCOV
/RANDOM=INTERCEPT | SUBJECT(Subject)

If anyone could help me with the syntax for the Accuracy data, I'd be so so
grateful.



--
View this message in context: http://spssx-discussion.1045642.n5.nabble.com/Logistic-Regression-with-Interactions-Categorical-DV-tp5725838.html
Sent from the SPSSX Discussion mailing list archive at 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

[Bruce Weaver]

Here's one old thread that may be useful to the OP:

http://spssx-discussion.1045642.n5.nabble.com/GenLinMixed-GEE-or-something-else-for-unbalanced-repeated-measures-data-with-dichotomous-DV-tp5725040.html

Ryan Black wrote

You could employ a random effects logistic regression via the procedure
GENLINMIXED or a GEE logistic regression model via the procedure GENLIN.
The syntax is fairly straightforward, and there are many examples if you
search the archives:

http://listserv.uga.edu/cgi-bin/wa?S1=spssx-l

or read the manual.

Ryan


On Tue, May 6, 2014 at 1:30 PM, Arakashi <[hidden email]> wrote:

> I have a 2 (group: patients, controls) between subject X 2 (Task 1, Task 2)
> within subjects X 3 (Trial type 1, Trial type2 and Trial type3) within
> subjects.  My DV is Accuracy (binary, yes or no).
>
> My data is not normally distributed because accuracy is at ceiling, so I
> was
> going to use logistic regression.
> However, I cannot for the life of me figure out what to do with the
> Accuracy
> data since it's categorical/binary.  I need to be able to look at the
> interactions…
>
> For the reaction time data, I used the following syntax to explore the main
> effects and interactions (RT was for correct trials only):
> MIXED Probe.RT WITH Condition Trial Group
> /FIXED=Condition Trial Group Condition*Group Trial*Group Condition*Trial
> Condition*Trial*Group | SSTYPE(3)
> /METHOD=reml
> /PRINT = SOLUTION TESTCOV
> /RANDOM=INTERCEPT | SUBJECT(Subject)
>
> If anyone could help me with the syntax for the Accuracy data, I'd be so so
> grateful.
>
>
>
> --
> View this message in context:
> http://spssx-discussion.1045642.n5.nabble.com/Logistic-Regression-with-Interactions-Categorical-DV-tp5725838.html
> Sent from the SPSSX Discussion mailing list archive at 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
>

[Ryan]

It is beyond me why the SPSS developers have yet to provide a TEST statement for GENLIN and GENLINMIXED. It really inhibits one's ability to construct specific contrasts of interest. I hope they are working on it. 

Ryan

On Tue, May 6, 2014 at 3:04 PM, Bruce Weaver <[hidden email]> wrote:

Here's one old thread that may be useful to the OP:

http://spssx-discussion.1045642.n5.nabble.com/GenLinMixed-GEE-or-something-else-for-unbalanced-repeated-measures-data-with-dichotomous-DV-tp5725040.html



Ryan Black wrote

> You could employ a random effects logistic regression via the procedure
> GENLINMIXED or a GEE logistic regression model via the procedure GENLIN.
> The syntax is fairly straightforward, and there are many examples if you
> search the archives:
>
> http://listserv.uga.edu/cgi-bin/wa?S1=spssx-l
>
> or read the manual.
>
> Ryan
>
>
> On Tue, May 6, 2014 at 1:30 PM, Arakashi <

> arakashi@

> > wrote:
>
>> I have a 2 (group: patients, controls) between subject X 2 (Task 1, Task
>> 2)
>> within subjects X 3 (Trial type 1, Trial type2 and Trial type3) within
>> subjects.  My DV is Accuracy (binary, yes or no).
>>
>> My data is not normally distributed because accuracy is at ceiling, so I
>> was
>> going to use logistic regression.
>> However, I cannot for the life of me figure out what to do with the
>> Accuracy
>> data since it's categorical/binary.  I need to be able to look at the
>> interactions…
>>
>> For the reaction time data, I used the following syntax to explore the
>> main
>> effects and interactions (RT was for correct trials only):
>> MIXED Probe.RT WITH Condition Trial Group
>> /FIXED=Condition Trial Group Condition*Group Trial*Group Condition*Trial
>> Condition*Trial*Group | SSTYPE(3)
>> /METHOD=reml
>> /PRINT = SOLUTION TESTCOV
>> /RANDOM=INTERCEPT | SUBJECT(Subject)
>>
>> If anyone could help me with the syntax for the Accuracy data, I'd be so
>> so
>> grateful.
>>
>>
>>
>> --
>> View this message in context:
>> http://spssx-discussion.1045642.n5.nabble.com/Logistic-Regression-with-Interactions-Categorical-DV-tp5725838.html
>> Sent from the SPSSX Discussion mailing list archive at Nabble.com.
>>
>> =====================
>> To manage your subscription to SPSSX-L, send a message to
>>

> LISTSERV@.UGA

>  (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
[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.

--
View this message in context: http://spssx-discussion.1045642.n5.nabble.com/Logistic-Regression-with-Interactions-Categorical-DV-tp5725838p5725840.html

Sent from the SPSSX Discussion mailing list archive at 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