How to do listwise deletion in SPSS or Excel?

[zencaroline]

Hi,

I asked similar question some time ago. However, I still can't
completely understand the feedbacks kindly provided. So, I try to ask
the question more specificly once again.

Imaging in the SPSS or Excel data set, the horizon from left to
right are variable 1 to variable 16; the vertical from top to down are
invidual reponses 1000 cases. Some individuals don't answer all the
16 questions (variables). I only want the individuals (cases) that
answer all the 16 questions (variables). I don't need the invidual
cases that don't answer all the 16 questions (variables) completely.
That is, I want to conduct listwise deletion for missing data.

Might any expert please kindly tell me how to conduct the listwise
deletion in SPPS or Excel, step by step? Step by step, what do I need
to do to accomplish the work?

Thank you very much and look forward to any possible kind reply.
Please take care
Caroline

 

[Bruce Weaver]

Hi,

I asked similar question some time ago. However, I still can't
completely understand the feedbacks kindly provided. So, I try to ask
the question more specificly once again.

Imaging in the SPSS or Excel data set, the horizon from left to
right are variable 1 to variable 16; the vertical from top to down are
invidual reponses 1000 cases. Some individuals don't answer all the
16 questions (variables). I only want the individuals (cases) that
answer all the 16 questions (variables). I don't need the invidual
cases that don't answer all the 16 questions (variables) completely.
That is, I want to conduct listwise deletion for missing data.

Might any expert please kindly tell me how to conduct the listwise
deletion in SPPS or Excel, step by step? Step by step, what do I need
to do to accomplish the work?

Thank you very much and look forward to any possible kind reply.
Please take care
Caroline

In SPSS, see the nvalid and nmiss functions under COMPUTE. E.g.,

select if (nvalid(v1 to v16) EQ 16).
exe.

Or...

select if (nmiss(v1 to v16) EQ 0).
exe.

If you don't want to permanently delete the unwanted cases from the
file, set a filter rather than using "select if". E.g.,

compute f = (nmiss(v1 to v16) EQ 0).
filter by f.
exe.

 

[necia]

Another way would be to use the COUNT command.

Count Missing= V1 to V16 (missing).

That will create a new variable with the number of missing values across
the set of variables which I called V1 to V16. Using the word "missing"
will pick up missing whether it user-defined or system missing - therefore
all missing values.

Then do a SELECT IF (Missing eq 0).

to keep in the active file only those people for whom there were no
missing values across your set of variables.

Do where ever analyses you want. IF you want those folks for only one
analysis you culd put TEMPORARY in the line before the chosen analysis,
and after the analysis, all folks are back in the active file.

Caution. Do not save a system file while you have a SELECT IF active,
unless you really want just those folks, ie; do NOT overwrite your full
data file with one that has those folks "temporarily" selected out, or you
will loss data. Give the new system file a new name and then your
original folks are safe.

Cheers,
Necia


Necia A. Black, Ph.D. e-mail: (e-mail address removed)
246 Computing Center Web URL:http://www.acsu.buffalo.edu/~black
SUNY-Buffalo
Buffalo, New York 14260