Looking for a reference

[Salbod]

Dear Friends,

I am in need of a reference.

The type of design I am using is an "Untreated Control Group Design with
Dependent Pretest and Posttest Samples" also known as the nonequivalent
comparison group design (see Shadish, Cook,& Campbell, 2002, p 136-137.).

I am using the BDI to measure severity of depressive symptoms. It is
hypothesized that there will be greater reduction of depressive symptoms in
the treatment group than in the comparison group.

I planned to analyze the data using a 2 X 2(Treatment/Comparison X
Pretest/Posttest) ANOVA repeated.  However, someone I cannot ignore has
suggested that t-tests be run between the two groups on pretest BDIs and
then a t-test be run between the two groups on posttest BDIs.

Can anyone point me to a good reference that discusses the analysis of this
design?

I thank you.

Stephen Salbod, Pace University, NYC

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

[Burleson,Joseph A.]

Stephen:

The suggestion has merit in that tx and comparison groups may lack
equivalence at BL. Rather than run t-tests, however, I would run an
ANOVA with whatever variables you have in efforts to demonstrate
equivalence at BL (i.e., on gender, the dependent measure, everything).

Running a t-test at post-test is, however, not necessary, in that the
repeated-measure analysis will show differential change. In fact, I am
not sure what you would do with the results of this t-test if you did
run it.

One has to live with the pre-test differences, or try to show how if the
groups are not equivalent, that they are still within range of each
other, such that one or the other is not experiencing a floor or ceiling
effect that would cause them to change less.

More importantly is to show that the distribution of the difference
scores in the continuous dependent measures is distributed symmetrically
(i.e., not skewed). The BL and the PT themselves, however, do not have
to meet this criterion, only the difference scores.

Other assumptions of Rm anove do have to be met, nonetheless.

Joe Burleson

-----Original Message-----
From: SPSSX(r) Discussion [mailto:[hidden email]] On Behalf Of
Stephen Salbod
Sent: Saturday, March 15, 2008 4:20 PM
To: [hidden email]
Subject: Looking for a reference

Dear Friends,

I am in need of a reference.

The type of design I am using is an "Untreated Control Group Design with
Dependent Pretest and Posttest Samples" also known as the nonequivalent
comparison group design (see Shadish, Cook,& Campbell, 2002, p
136-137.).

I am using the BDI to measure severity of depressive symptoms. It is
hypothesized that there will be greater reduction of depressive symptoms
in
the treatment group than in the comparison group.

I planned to analyze the data using a 2 X 2(Treatment/Comparison X
Pretest/Posttest) ANOVA repeated.  However, someone I cannot ignore has
suggested that t-tests be run between the two groups on pretest BDIs and
then a t-test be run between the two groups on posttest BDIs.

Can anyone point me to a good reference that discusses the analysis of
this
design?

I thank you.

Stephen Salbod, Pace University, NYC

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

[Martin P. Holt-2]

Stephen, there's this:

Cook, Thomas D. and Donald T. Campbell, Quasi-experimentation: Design and
analysis issues for field settings. Boston: Houghton-Mifflin, 1979. A
leading classic.

HTH,
Martin Holt
----- Original Message -----
From: "Stephen Salbod" <[hidden email]>
To: <[hidden email]>
Sent: Saturday, March 15, 2008 8:20 PM
Subject: Looking for a reference

> Dear Friends,
>
> I am in need of a reference.
>
> The type of design I am using is an "Untreated Control Group Design with
> Dependent Pretest and Posttest Samples" also known as the nonequivalent
> comparison group design (see Shadish, Cook,& Campbell, 2002, p 136-137.).
>
> I am using the BDI to measure severity of depressive symptoms. It is
> hypothesized that there will be greater reduction of depressive symptoms
> in
> the treatment group than in the comparison group.
>
> I planned to analyze the data using a 2 X 2(Treatment/Comparison X
> Pretest/Posttest) ANOVA repeated.  However, someone I cannot ignore has
> suggested that t-tests be run between the two groups on pretest BDIs and
> then a t-test be run between the two groups on posttest BDIs.
>
> Can anyone point me to a good reference that discusses the analysis of
> this
> design?
>
> I thank you.
>
> Stephen Salbod, Pace University, NYC
>
> =====================
> 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