Make sure you're regression and correlations are using the same cases.
Depending upon procedures used, missing data are treated differently.
-----Original Message-----
From: SPSSX(r) Discussion [mailto:
[hidden email]] On Behalf Of
Stephan
Sent: Wednesday, August 19, 2009 4:28 AM
To:
[hidden email]
Subject: Partial correlations
Hi, I ran a multiple regression yesterday on a data set with 500 cases
using 5 independent variables. I checked my MR assumptions and these have
not been broken. On examination of my output, I saw that my partial
correlations were higher than my zero order correlations. I've not come
across this before. On researching this a little I found that this can
happen if there are negative correlations between the variables and one is
acting as a suppressor. However all of mine are positive. Anybody else
come across this or know of a reason/explanation for this result? Thanks
=====================
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
Locked
