Screening Several Variables to Create Conditional Varible

[scvp]

Hello all,

I want to screen thought a set of 6 dichotomous variables (1/0), that each accounts for one disease (D1-D6)

From those I want to create another categorical variable (Dany) that will separate observations with '0-1 disease' and '=>2 more diseases'.

I imagine this can be done with an array but I am quiet unsure how.

Any tips are appreciated!

scvp

[Astounding]

Even easier than using arrays:

two_plus = sum(of d1-d6) > 1;

The new variable TWO_PLUS will be 1 (for two or more diseases) and 0 (for fewer than two diseases).

[ballardw]

And an excellent example of why coding dichotomous as 1/0 instead of Y/N, T/F or similar character coding schemes.

Another approach might be to just do the sum and use a format to create groupings on a single value such

proc format;

value disease2groups

0,1 = '0 or 1'

2-high ='2 or more'

;

value disease3groups

0,1 = '0 or 1'

2,3 = '2 or 3'

4 - high = '4 or more'

;

run;

Then use the format at the time of use for which grouping is desired on the sum variable.