Date:         Fri, 7 Feb 2003 11:21:06 -0500
Reply-To:     Art@DrKendall.org
Sender:       "SPSSX(r) Discussion" <SPSSX-L@LISTSERV.UGA.EDU>
From:         "Arthur J. Kendall" <Art@DrKendall.org>
Organization: Social Research Consultants
Subject:      Re: Question about missing value
Comments: To: Nan Li <nanl@nlchi.nf.ca>
Content-Type: text/plain; charset=us-ascii; format=flowed

wrt alternatives, of course a lot depends on the nature of the original
string variable and what you want to find its influence on.

start with something like this assuming the missing values are blank.
this will tell you the percent of all cases and the percent of non
missing cases.

if the string is A8 or smaller you could:
missing values old_str ("").
frequencies vars= old_str.

or

autorecode vars= old_str /into new_num (a) /print.
* you might want to recode new_num into fewer categories if diiferent
values really mean the same thing. e.g., "vs", "VS" "vs." "VS." .
missing values new_num (1).
frequencies vars = new_num / statistics= all.

Hope this helps.

Art
Art@DrKendall.org
Social Research Consultants
University Park, MD USA
(301) 864-5570




Nan Li wrote:
> Dear Lister,
>
> I have a String type variable with 7.5% missing data. I am wondering whether
> I could delete these missing data and perform some descriptive analysis on
> this variable? If I would like to test the distribution of this variable
> before and after deleting the missing data or to test the influence of these
> missing data, how would I do it?  I'd greatly appreciated for any input on
> this matter.
>
> Thanks ...
>
> nan
>