spss, How to compute a dummy variable

I have four dummy variables (yes/no). I would like to compute a variable
(Transform > Compute) that is also a dummy variable. It should be 0 when
none of the four dummy variables is 1. Otherwise it should be a 1. Only when
all four dummy variables have no value (no response), the new variable
should also have no value. How can I compute such a variable?

Does this do what you want?

compute newdum = sum(dum1,dum2,dum3,dum4) GT 0.
exe.

--
Bruce Weaver
XXXX@XXXXX.COM
www.angelfire.com/wv/bwhomedir

Thanks!
When I enter the function "sum(dum1,dum2,dum3,dum4) GT 0" in the all the
values of the new variable are "1".
However, I now see that the initial dummy variables are value "1" when no
and "2" when yes, instead of "0" and "1". Perhaps this causes this mistake?
How can I correct this and what does the "GT 0" mean?

"Bruce Weaver" < XXXX@XXXXX.COM > wrote

when

Yes, it does. Dummy variables have values of 1 and 0. Try this:

recode dum1 dum2 dum3 dum4 (2=0) (else = copy).
exe.

Then try again.

Normally, it is not a good idea to recode a variable back into itself.
But in a case like this, there are no bad consequences if you happen to
run the RECODE command twice. Nothing would happen the second time,
because there would be no 2's left.


GT 0 means greater than 0.

--
Bruce Weaver
XXXX@XXXXX.COM
www.angelfire.com/wv/bwhomedir

As far as Bruce's response goes:

The GT means "Greater Than" or ">".

For your specific situation (1's, 2's versus 0,1s), change Bruce's
suggested syntax to the following:

compute newdum = sum(dum1,dum2,dum3,dum4) GE 5.
exe.

Where the GE means "Greater than or equal to" or ">=".

Not the prettiest solution, but it should work.

Other

COUNT newdum = Dum1 Dum2 Dum3 Dum4 (2) .
RECODE newdum (1 THRU HI=1).
EXECUTE .


And other

COMPUTE Newdum=(Dum1=2) OR (Dum2=2) OR (Dum3=2) OR (Dum4=2).
EXECUTE.


MJ

And a neat form for the latter statement is

Compute Newdum= any(2, dum1, dum2, dum3, dum4) .

ANY() is dual purpose; that is, it is possible to say,
"any(value, varlist)" as well as "any(var, value-list)"

--
Rich Ulrich, XXXX@XXXXX.COM
http://www.pitt.edu/ ~wpilib/index.html