repeated measures

[Jim Lalley]

I have data that are based on the comparison of two comparable
learning treatments: computer based or traditional.

Students completed their respective learning session and then immediately
completed a 15 item test on its content.  Students were also given a
parallel version of that test after a one-week delay.  The same instruments
were used with both groups

Means and standard deviations for the groups are as follows:

Computer Treatment (n=99): Posttest=12.06, SD=2.128; Delayed Posttest=9.68,
SD=2.601

Traditional Treatment (n=85): Posttest=10.18, SD=2.734; Delayed
Posttest=9.05, SD=2.563

I am less interested in the change over time (from test to test) than if
there were differences between the two groups.

My question concerns analysis of the data. I am using SPSS and have chosen
to do repeated measures found under the general linear models menu. I first
add a factor called testchange which has two levels: the immediate test and
the delayed test. Then I define the two levels as the immediate test and
delayed test, in that order, as the within-subjects variable. I also define
treatment type as my between-subjects variable.

The Mauchly's test of sphericity is not significant, which I believe gives
me the green light to use the repeated measures analysis.

When I run the analysis I get an F=72.44, p=.000 for the within-subject
effect of testchange. I believe this indicates that overall, for both
groups, there was a difference found between the immediate and delayed tests.

I also get an F=17.72, p=.000 for the within-subject interaction of
testchange*treatment. I believe that this indicates that the difference that
occurred between the immediate and delayed tests was found for both groups.
 The same output appears under within-subject contrasts.

Under between-subject effects I get an F=26.56, p=.000 for the variable
treatment. I believe this indicates that there were differences in test
scores between the two treatment groups. What I am not sure of is: does this
mean that there was a difference between groups on the immediate test, the
delayed test, or perhaps both?

I have tried changing settings under options, contrasts and post hoc but
have not had success in getting an answer for these questions. I have also
searched the net and spss help but have had no luck finding an example with
a design like mine.

Sorry for the lengthy description.

Thanks, Jim Lalley

D'Youville College

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

[Swank, Paul R]

The between subjects effect in the repeated measures design is a sum or average of the measures. Usually, with a repeated measures design, you also get the univariate results, that is the difference at each time. This will inform you if the groups differed at pretest and if they differed at post.

Dr. Paul R. Swank,
Professor and Director of Research
Children's Learning Institute
University of Texas Health Science Center-Houston


-----Original Message-----
From: SPSSX(r) Discussion [mailto:[hidden email]] On Behalf Of Jim Lalley
Sent: Wednesday, April 15, 2009 1:38 PM
To: [hidden email]
Subject: repeated measures

I have data that are based on the comparison of two comparable
learning treatments: computer based or traditional.

Students completed their respective learning session and then immediately
completed a 15 item test on its content.  Students were also given a
parallel version of that test after a one-week delay.  The same instruments
were used with both groups

Means and standard deviations for the groups are as follows:

Computer Treatment (n=99): Posttest=12.06, SD=2.128; Delayed Posttest=9.68,
SD=2.601

Traditional Treatment (n=85): Posttest=10.18, SD=2.734; Delayed
Posttest=9.05, SD=2.563

I am less interested in the change over time (from test to test) than if
there were differences between the two groups.

My question concerns analysis of the data. I am using SPSS and have chosen
to do repeated measures found under the general linear models menu. I first
add a factor called testchange which has two levels: the immediate test and
the delayed test. Then I define the two levels as the immediate test and
delayed test, in that order, as the within-subjects variable. I also define
treatment type as my between-subjects variable.

The Mauchly's test of sphericity is not significant, which I believe gives
me the green light to use the repeated measures analysis.

When I run the analysis I get an F=72.44, p=.000 for the within-subject
effect of testchange. I believe this indicates that overall, for both
groups, there was a difference found between the immediate and delayed tests.

I also get an F=17.72, p=.000 for the within-subject interaction of
testchange*treatment. I believe that this indicates that the difference that
occurred between the immediate and delayed tests was found for both groups.
 The same output appears under within-subject contrasts.

Under between-subject effects I get an F=26.56, p=.000 for the variable
treatment. I believe this indicates that there were differences in test
scores between the two treatment groups. What I am not sure of is: does this
mean that there was a difference between groups on the immediate test, the
delayed test, or perhaps both?

I have tried changing settings under options, contrasts and post hoc but
have not had success in getting an answer for these questions. I have also
searched the net and spss help but have had no luck finding an example with
a design like mine.

Sorry for the lengthy description.

Thanks, Jim Lalley

D'Youville College

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