help with combining/collapsing/reducing data

[The D Man]

Hello, and thank you in advance for your help.

I have searched through the web, and through this forum and have not
found the information I am looking for. I think it may be because I am
not sure the exact terminology to describe the problem.

I have data from a test that was performed in 2 different sets of
conditions - quiet v noise, and forward v backward. I coded the data
using dummy variables in order to see what interactions and patterns
there may be. I found that backward was worse than forward, but no
differences existed with quiet v. noise. My understanding is that I
can collapse the quiet and noise results together in order to get more
power for the subsequent statistics. I don't know how to do that. I
thought about creating a new variable that was the average of the
quiet and noise score for each subject, but that doesn't give me any
additional power. Is there another function in SPSS that allows the
software to run the quiet and noise results together?

[Bruce Weaver]

Tell us more about the design. Were the two factors manipulated
between or within subjects? Or was there one of each (i.e., a mixed
design)? What is the layout of the data file?

--
Bruce Weaver
bwe...@lakeheadu.ca
http://sites.google.com/a/lakeheadu.ca/bweaver/Home
"When all else fails, RTFM."

[Rich Ulrich]

On Wed, 2 Dec 2009 18:41:37 -0800 (PST), The D Man
<dere...@gmail.com> wrote:

>Hello, and thank you in advance for your help.
>
>I have searched through the web, and through this forum and have not
>found the information I am looking for. I think it may be because I am
>not sure the exact terminology to describe the problem.
>
>I have data from a test that was performed in 2 different sets of
>conditions - quiet v noise, and forward v backward. I coded the data
>using dummy variables in order to see what interactions and patterns
>there may be. I found that backward was worse than forward, but no
>differences existed with quiet v. noise. My understanding is that I
>can collapse the quiet and noise results together in order to get more
>power for the subsequent statistics. I don't know how to do that. I
>thought about creating a new variable that was the average of the
>quiet and noise score for each subject, but that doesn't give me any
>additional power.

Well, that is how you do it. You merely average the two
scores for each condition. The outcome of "increased power" is
not guaranteed, and the increase might not be huge.

The increase of power comes from the use of a "more
reliable" version of the variable - It is the same idea as
using a Total score instead of separate Item scores.

Of course, if there is no "real effect" in what you are
trying to test, then even a huge increase of power will
merely show you - with smaller confidence intervals -
that you continue to fail to reject the null hypothesis.

> Is there another function in SPSS that allows the
>software to run the quiet and noise results together?

--
Rich Ulrich