missing values

[Chelminski, Iwona]

Hi Group,
Quick question--do you have any idea how to quickly identify a case that has
more than 20% of missing values? So for example my ID #1 case has 200
variables, corresponding to 200 item questionnaire and I want to eliminate
it if s/he didn't answer 20% of the items (i.e., 40 variables). I could do
it with recoding, summing up, etc, but there must be an easy way....is
there?

[Keith McCormick]

Hello,

If you are describing a case with lots of missing, it is indeed easy:

COUNT
  nummiss = var1 to var100  (MISSING)  .

then you can do a select if, etc.

Best, Keith

www.keithmccormick.com

On 9/12/06, Chelminski, Iwona <[hidden email]> wrote:
> Hi Group,
> Quick question--do you have any idea how to quickly identify a case that has
> more than 20% of missing values? So for example my ID #1 case has 200
> variables, corresponding to 200 item questionnaire and I want to eliminate
> it if s/he didn't answer 20% of the items (i.e., 40 variables). I could do
> it with recoding, summing up, etc, but there must be an easy way....is
> there?
>