Question

How to handle statistics questions where survey responses add up to more than the number of survey participants?

For example, the survey question asked 50 people, How do you deal with your garbage?

Collection = 37, Burning = 26, Composting = 24, Throwing anywhere = 18, Recycling = 9, Reusing = 3

Answer 1

See below:

Explanation:

One way to handle this kind of situation is to simply take the average of each category and know that you're not going to add to 100% - this kind of question is more a "ranking" kind of question.

For instance, let's put some numbers to the types of ways of dealing with garbage disposal:

Collection = 37
Burning = 26
Composting = 24
Throwing Anywhere = 18
Recycling = 9
Reusing = 3

So now we take the percentage of the number of "yes's" to each response against the number of people, which is 50:

Collection = 74%
Burning = 52%
Composting = 48%
Throwing Anywhere = 36%
Recycling = 18%
Reusing = 6%

The usefulness is that, if you've taken a large enough sample, you can apply these percentages to much larger populations. If we had a population of 100,000 people, we could estimate the number of people who deal with their garbage in the various ways. For instance, we'd estimate that 6,000 people reuse their garbage and 74,000 use collection services.