How to code 'skipped questions' in SPSS

[arisai...@yahoo.com]

Hi all,

I have a few 'skipped questions' in my questionnaire.
e.g. Question 1: Do you smoke? Answer: 'YES' or 'NO'. If the
respondent answer NO, they will be asked to go to Question 5 and skip
Question 2,3,4.

How do I code Question 2,3,4 for those who answer 'NO' to question 1?

Thank you

[Rich Ulrich]

On Fri, 22 Jul 2011 17:00:47 -0700 (PDT), arisai...@yahoo.com
wrote:

missing values var2 to var5(9).
value labels var1 to var5 (1) yes (2) no (9) NA.

--
Rich Ulrich

[arisai...@yahoo.com]

Many thanks.

So, how do I calculate a scale that comprised the missing data
(skipped data)?

For example, there is 6 scales (6 sections) in my questionnaire. Each
scale consists a few items. I need to add the items to get the total
score of each scale. My sample size is 159. The total subject answered
the skipped questions were 41 only. So how do I calculate the total
score? I have tried calculating the items by including the skipped
data (which the N is lesser). However, the total score only appear for
those who did not skip any questions (N=41) whereas for those who
skipped, the total score was none.

Thank you

[Rich Ulrich]

On Sat, 23 Jul 2011 14:19:00 -0700 (PDT), arisai...@yahoo.com
wrote:

That is not a clear description of what you want, as a matter of
logic.

I will assume that you want to compute a pro-rated total - what the
total would be if the missing item matched the average of the items
that were there. This is a poor assumption when items have means
that differ by a lot.

It is easier to compute a pro-rated mean, and it is usually
nicer to describe pro-rated means in a report, especially when
scales have different numbers of items. For items with scales,
the results can be referred to the scale-anchors for reference.

Compute newvar= mean.3( item1 to item6).
* this .3 says, "return Missing if there are fewer than3 valid
* responses". I usually require 3/4 or 2/3 to be present.

For a prorated total, include "6* .... "

--
Rich Ulrich