Google Groups no longer supports new Usenet posts or subscriptions. Historical content remains viewable.
Dismiss

Making SPSS to display empty rows/columns - ?

572 views
Skip to first unread message

Alexandre Bougakov

unread,
Apr 17, 2002, 10:04:01 AM4/17/02
to
Hello, All,

By default SPSS hides rows/columns with zeroes or missing values in
tables in the output. Zeroes display can be enabled in the TableLook
properties by unmarking "Hide empty rows.." option. But the missing
values are still not displayed!

This is a big problem for me since the tables from the output tables
are copied in Word and the VB Script inside Word document expects,
that all my tables of the same type will have the same number of
columns.

For example, it passes well first table:

question1 > Yes > No > Don't know
Male > 11 > 12 > 13
Female > 14 > 15 > 16

question2 > Yes > Don't know
Male > 21 > 23
Female > 24 > 26

And fails, when it meets next one, simply because the people, who
answered "no" to the question 2 were filtered out or they didn't
answered it for any other reason and there are SYSMIS values in the
corresponding data cells:

"SPSS Tables" need an additional option in the TableLook properties -
Hide/Unhide rows/cols with missing values - maybe it will be added
there in the future, but does anyone have any solution to solve this
problem using scripts or something else now? Or there is an setup
option that I am missing?

Thanks in advance, Sanja

Arthur J. Kendall

unread,
Apr 17, 2002, 11:01:26 AM4/17/02
to

Alexandre Bougakov

unread,
Apr 17, 2002, 1:34:42 PM4/17/02
to
"Arthur J. Kendall" <Arthur....@verizon.net> wrote in message news:<3CBD8E44...@verizon.net>...

Thanks a lot! It looks like a kind of voodoo, but it works...

Sanja.

Bruce Weaver

unread,
Apr 17, 2002, 5:09:40 PM4/17/02
to

On Wed, 17 Apr 2002, Arthur J. Kendall wrote:


Does anyone else find this work-around cumbersome? Why can't the good
folks at SPSS make display of empty rows and columns in pivot tables
optional? It doesn't seem like it should be that much of a programming
challenge.

--
Bruce Weaver
E-mail: wea...@mcmaster.ca
Homepage: http://www.angelfire.com/wv/bwhomedir/

Nick Nelson

unread,
Apr 19, 2002, 2:51:31 AM4/19/02
to

Bruce Weaver wrote:

> Does anyone else find this work-around cumbersome? Why can't the good
> folks at SPSS make display of empty rows and columns in pivot tables
> optional? It doesn't seem like it should be that much of a programming
> challenge.

I imagine it would be trickier than you might think.
Bear in mind that you may know that no one in a tabulation
chose '10' even though it would have ben a valid option,
but how is SPSS to know that. One would need to tell
SPSS all the valid responses even though they might
not actually occur in the data.

Actually I thought you could do this using 'integer
mode' in syntax, but when I just tried (using v10.1)
it didn't seem to work the way I expected.


Nick

Arthur J. Kendall

unread,
Apr 19, 2002, 4:18:06 PM4/19/02
to
Perhaps one approach would be to have an option where all valid values with
explicit labels would be displayed.
When this option was chosen, the s/w would count the valid values and
tabulate using those. It even could be further generalized so the values did
not have to be sequential.
value labels decade 40 '1940s' 50 '1950s' 60 '1960s' . . .
0 new messages