Transforming expansive “Check all that apply” type questions into random selection of lesser items

问题

I have been facing some difficulties in managing dietary data regarding cereal consumption. The question allows participants to select as many cereal types that they eat (among 300+ choices). I would like to collapse this into just two answers through random selection. The data structure is posing problems. I will outline my problem below…

1.) Structure of the data now: Respondents are allowed to select all cereals that they “usually” eat. The data looks like (1 is replaced with a code for the cereal)…

Cereal 1     Cereal 2     Cereal 3     Cereal 4     Cereal 5  
1             0           1            1            0

2.) I will need to collapse these choices into just two, even when there are more than two cereals selected. What I am thinking is that I would first collapse these choices into as many as each respondent selected. I want the data to be changed to the structure below…

Choice 1              Choice 2              Choice 3              Choice 4             Choice 5              
Cereal 1              Cereal 3              Cereal 4                 .                     .

3.) I would then want to randomly select two of these (since we cannot distinguish order of selection) to be displayed like the data below…

Choice 1                       Choice 2
Cereal 1 (randomly selected)   Cereal 4 (Randomly selected)

Does anyone have an idea of how I might accomplish this? I have been staring at my code for some time and can’t quite seem to figure this out. Please let me know if further explanation is needed.

回答1:

Assumptions:

1. Cereal variables are numeric
2. 0 is missing and otherwise the values for a cereal are larger than 0
3. Cereals have a unique code so variable names are not important

4. Everyone has selected at least two cereals.

   data want;
   set have;
   
   *set random seed to get the same 'random' results;
    call streaminit(999);
   
   *declare an array to simplify code;
    array cereal(*) cereal1-cereal300;
   
   *sort array so that cereals come first and 0s last;
   call sort(of cereal(*));
   
   *find the first 0;
   *this tells you how many cereals are selected;
   numCereals= whichn(0, of cereal(*)) - 1;
   
   *get a random number between those two numbers;
   rand1 = rand('integer', 1, numCereals);
   
   *ensure that your second random number isn't the same as the first random number;
   do until(rand2 ne rand1);
       rand2 = rand('integer', 1, numCereals);
   end;
   
   *get the values at those random index;
   randomCereal1 = cereal(rand1);
   randomCereal2 = cereal(rand2);
   
   run;
   

There is untested.

来源：https://stackoverflow.com/questions/61525204/transforming-expansive-check-all-that-apply-type-questions-into-random-selecti