I hope this makes sense.
First the mathematical trick. The mean is total divided by no of
cases. Thus the total is the mean multiplied by the no of cases. A
easy way to do mean substitution on a case basis is to calculate
the mean of the valid subscales and then multiply it by the actual
number of subscales in the scale.
You are half-way there.
75% of 17 question subscale means you require at least 13
subscales to be present. That implies that you want the number
after the '.' in the mean statement to be 13. This also works for
sum but because you need mean substitution it is easier to work
with the mean and multiply out.
Thus you need the statement like
COMPUTE xbar = MEAN.13(pjss1 to pjss17)*17.
If you want to do this via Compute dialogue then you need to type
in the numeric expression box.
MEAN.13(pjss1 to pjss17)*17
Jean M. Russell
Date sent: Tue, 15 May 2001 16:02:23 -0400
Send reply to: "Cohen,Diane" <dc42@EXCHANGE1.DREXEL.EDU>
From: "Cohen,Diane" <dc42@EXCHANGE1.DREXEL.EDU>
Subject: Compute total with 25% missing answers
> Dear Folks, I have a subscale with 17 items for which I calculated a
> total score (with ease, remarkably enough!). Now I need help with for
> the cases in this scale in which the respondant answered 75% or more
> of the items. I saved the syntax you all sent on how to do this with
> another scale where the total score was a mean of all responses:
> COMPUTE xbar = MEAN.12(pjss15 to pjss30). EXECUTE I thought if I
> fiddled around with it somehow I could figure it out, but no cigar. I
> would like to try it thru the compute command rather than syntax if
> anyone has the patience to tell me how to do it. I imagine it has an
> "if" statement, but when I went into that section I became even more
> lost. Thanks in advance, Diane d'Newbie
Jean M. Russell M.A. M.Sc. J.Russell@sheffield.ac.uk
Corporate Information & Computing Services,
University of Sheffield
285 Glossop Road
Fax : 0114-222-3040