Delete cases with more than 80% of missing data; Duplicates

[Kseniya Katsman]

Hello, I'm new to SPSS and trying my best to find all sorts of manuals and guides to help me understand it better. Right now I am dealing with a substantial dataset (3000 cases and 1500 variables). It consists of respondents providing answers to various psychological measures. I need to be able to delete cases with more than 80% of missing data. How can I automate this process?
Also, I have about 500 duplicate cases. I need to compare cases with identical names to each other and delete the one that has less data filled out. Is there any way to do it without going manually through each pair over all 1500 variables?
Thank you!

[Bruce Weaver]

The NMISS() and NVALID() functions could be useful here. Something like:

COMPUTE #missing = NMISS(variable list).
COMPUTE #valid = NVALID(variable list).
COMPUTE #total = #missing + #valid.
COMPUTE propmiss = #missing / #total.
VARIABLE LABELS propmiss "Missing data proportion".
DESCRIPTIVES propmiss.
SELECT IF propmiss LE .8.
DESCRIPTIVES propmiss.

Replace "variable list" with your list of variables, bearing in mind you can use the key word TO (e.g., firstvar TO lastvar). If you don't want to completely remove cases with propmiss > .8, set a filter rather than using SELECT IF.

HTH.

[Bruce Weaver]

I forgot to add that this usenet group is relatively inactive nowadays, so you might consider joining the SPSS mailing list. It can be accessed via this Nabble site:

http://spssx-discussion.1045642.n5.nabble.com/

HTH.

[Bruce Weaver]

Aha...never mind. I see you already found it!

http://spssx-discussion.1045642.n5.nabble.com/Delete-cases-with-more-than-80-of-missing-data-Handling-duplicates-td5733845.html

[Kseniya Katsman]

On Tuesday, February 21, 2017 at 9:37:30 AM UTC-5, Bruce Weaver wrote:

Thank you, Bruce!