r value differences

[Lori]

In a small sample (N=13!) what is the best method to test the difference
between correlations (r values) of related measures? So, parents reported
on child behavior at 2 time points; the correlation for one composite over
time = .634 (p=.02) and for the second composite = .338 (p=.258). What’s
the best way to determine whether these correlations (for two composites
from the same rating scale) are significantly different?

It’s my understanding that the Fisher r to z transformation would not be
appropriate because these are not two independent samples. . . but I’m not
sure whether there is a better or more appropriate test.

Thank you!

=====================
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]

You could use James Steiger's Multicorr program, available here:

   http://www.statpower.net/Software.html

Lori wrote

In a small sample (N=13!) what is the best method to test the difference
between correlations (r values) of related measures? So, parents reported
on child behavior at 2 time points; the correlation for one composite over
time = .634 (p=.02) and for the second composite = .338 (p=.258). What’s
the best way to determine whether these correlations (for two composites
from the same rating scale) are significantly different?

It’s my understanding that the Fisher r to z transformation would not be
appropriate because these are not two independent samples. . . but I’m not
sure whether there is a better or more appropriate test.

Thank you!

=====================
To manage your subscription to SPSSX-L, send a message to
[e-mail] (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