Date:         Tue, 29 Jul 2003 11:59:25 -0400
Reply-To:     SPSSX-L@LISTSERV.UGA.EDU
Sender:       "SPSSX(r) Discussion" <SPSSX-L@LISTSERV.UGA.EDU>
From:         Bob Walker <spss@surveys-forecasts.com>
Subject:      Re: statistical question - association between ordinal and
              continuous variables?
Comments: To: "Ohlott, Patty" <ohlott@leaders.ccl.org>
In-Reply-To:  <49B2F333F1CAD411B80100B0D068172204A514CB@gsoms01.ccl.org>
Content-type: text/plain; charset=us-ascii

Patty,

Based on your measures, I don't see why bivariate correlations and multiple
regression can't be used, but first standardize your variables (run
"Descriptive Statistics=>Descriptives" and check the "Save standardized
values as variables" box), since each skill (independents) may have its own
distribution, and your performance measure (dependent)is a three point
scale. I do this when analyzing 360 evaluations. Regards,

Bob Walker
Surveys & Forecasts, LLC
www.surveys-forecasts.com


|-----Original Message-----
|From: SPSSX(r) Discussion [mailto:SPSSX-L@LISTSERV.UGA.EDU] On Behalf Of
|Ohlott, Patty
|Sent: Tuesday, July 29, 2003 11:30 AM
|To: SPSSX-L@LISTSERV.UGA.EDU
|Subject: statistical question - association between ordinal and continuous
|variables?
|
|A colleague is looking at the relationship of specific managerial skills to
|overall performance.  The skills consist of 21 continuous scales (range 1
|to
|5).  The overall performance measure is ordinal - high, medium, low.  The
|sample size is >1000.  We want to see which of the skills are most related
|to performance.  She has tried SPSS's PLUM/ordinal regression procedure but
|is having some trouble interpreting the results - from the parameter
|estimates given the software seems to be treating the continuous variables
|as ordinal as well.  Is this the proper procedure to use?  If so, any
|advice/insight on interpreting the results? If not, are there others you
|would recommend?   Recommendations of any reference materials that might be
|helpful would be appreciated also.  She is on a tight deadline so if anyone
|can help, the sooner the better.
|
|Many thanks in advance!
|
|Patty
|
|Patricia J. Ohlott
|Senior Associate
|Center for Creative Leadership
|P.O. Box 26300
|Greensboro NC 27438-6300
|(336)286-4423