Date:         Thu, 24 Jun 1999 15:41:38 -0700
Reply-To:     glaser@pacific-science.com
Sender:       "SPSSX(r) Discussion" <SPSSX-L@LISTSERV.UGA.EDU>
From:         Dale Glaser <glaser@PACIFIC-SCIENCE.COM>
Subject:      Re: SPSS 9.0----Missing values
Comments: To: LEGBAS@AOL.COM
In-Reply-To:  <119ae040.24a40219@aol.com>
Content-Type: text/plain; charset="iso-8859-1"

Melissa..........by default, the reliability procedure employs a listwise
type of deletion; so as you found out, when missing data exists that
respondent is eliminated from the analysis, and there is no way around it.
Possibly others have experimented with missing data techniques (e.g.,
maximum likelihood) in the context of reliability, so they may be able to
guide you here, but I don't recall seeing much literature on
imputation/missing data techniques in the context of reliability assessment;
obviously one option is to run the full complement of the data without the
two items that had n/a as an option...............dale glaser

-----Original Message-----
From:   SPSSX(r) Discussion [mailto:SPSSX-L@LISTSERV.UGA.EDU] On Behalf Of
LEGBAS@AOL.COM
Sent:   Thursday, June 24, 1999 2:50 PM
To:     SPSSX-L@LISTSERV.UGA.EDU
Subject:        SPSS 9.0----Missing values

Hi.  I am using SPSS 9.0 for Graduate Students to anaylze data for a study
in
the field of psychology.  My problem concerns missing values.  I
administered
a quesionnaire that has several subscales.  I would like to run reliability
analyses (alpha) for the scales.  Respondents were permitted to leave two
quesitons blank if they did not apply.  Now when I try to run reliability
analyses for scales that include these items, SPSS completely leaves out any
respondents who left these questions blank when it computes alpha.  I don't
know how to include them, and include them properly, in finding alpha.
Help!
 Thanks very much.
Melissa