Is it a Mixed Model or t test?

[Salbod]

Good Evening, Dear Friends,



I've got data on preschoolers after 9-11, teachers rated the kids on

the BASC (a behavior checklist) and I wanted to compare the BASC on

those kids who were exposed (close) to WTC events to those not exposed (not
close).

As I sat down to start doing this I realize that a single teacher could

have rated anywhere from 1 to 15 kids--i.e., I've got 17 teachers who

rated about 70 kids. A preliminary ICC analysis revealed that Teacher
Differences explained 28% of the BASC variability.



Am I looking at a Mixed Model design here or a simple t-test?

I also looking for a reference of ICC cutoffs when you in this context.



All suggestions and references will be welcomed. Thank you.



Stephen Salbod, Pace University, NYC

[Richard Ristow]

At 05:30 PM 9/12/2007, Stephen Salbod wrote:

>I've got data on preschoolers after 9-11, teachers rated the kids on
>the BASC (a behavior checklist) and I wanted to compare the BASC on
>those kids who were exposed (close) to WTC events to those not exposed
>(not
>close).
>
>I realize that a single teacher could have rated anywhere from 1 to 15
>kids--i.e., I've got 17 teachers who rated about 70 kids. A
>preliminary ICC analysis revealed that Teacher Differences explained
>28% of the BASC variability.
>
>Am I looking at a Mixed Model design here or a simple t-test?

On the face of it, this is clearly a mixed model, with student nested
within teacher. I'm not up enough on the SPSS procedure to suggest
syntax, though.

[Salbod]

Richard,
        Thanks for giving me the exciting news: Mixed Design. Now, I've got
to crack the books and figure out the syntax.

        This might be a blessing in disguise.

Steve

-----Original Message-----
From: SPSSX(r) Discussion [mailto:[hidden email]] On Behalf Of
Richard Ristow
Sent: Friday, September 14, 2007 11:57 AM
To: [hidden email]
Subject: Re: Is it a Mixed Model or t test?

At 05:30 PM 9/12/2007, Stephen Salbod wrote:

>I've got data on preschoolers after 9-11, teachers rated the kids on
>the BASC (a behavior checklist) and I wanted to compare the BASC on
>those kids who were exposed (close) to WTC events to those not exposed
>(not
>close).
>
>I realize that a single teacher could have rated anywhere from 1 to 15
>kids--i.e., I've got 17 teachers who rated about 70 kids. A
>preliminary ICC analysis revealed that Teacher Differences explained
>28% of the BASC variability.
>
>Am I looking at a Mixed Model design here or a simple t-test?

On the face of it, this is clearly a mixed model, with student nested
within teacher. I'm not up enough on the SPSS procedure to suggest
syntax, though.