I have a complicated (to me, anyway) scoring procedure and I am trying to determine the easiest approach. I think some kind of IF (...) THEN (...) set up will work. I have some dietary data with a frequency of consumption (codes ranging from 0 to 2.0). There are 15 different food items which results in 15 different frequency of consumption responses. I am supposed to take each one of those and multiply it by a corresponding portion size value. These portion sizes are not only dependent on the type of food (of the 15) but also the portion size value differs by the respondents gender (male or female) and age grouping (18-27, 28-37...78+). For example, for a person who reports consuming skim milk 2 times per day and is a 19 year old female, I would have to multiply 2 (consumption frequency) by 245.0 (the corresponding portion size for a 19 year old female); if everything else were the same but the respondent was male, I would have to multiply 2 by 366.7. In the end all the products for each food (15 total) will be added together for the final score.
I am thinking the best way is to do each food one at a time and then just have some sort of IF THEN statement for whether they are male or female and what age grouping they belong in to dictate what portion size to multiply the frequency by.
If it helps, this is an online site for the actual scoring procedures that I am using http://riskfactor.cancer.gov/diet/scree ... html...all
based in SAS. I am using SPSS 19.0.