percent of cases 解释起来方便。

谢谢您，您能具体点解释一下么?谢谢

There are two kinds of multiple response sets in SPSS Statistics, and they are not interchangeable. The MULT RESPONSE procedure defines the set within the procedure syntax and is part of the Base system. The other, newer set definitions are used by Custom Tables (an add-on option) and, in the Base in the Chart Builder. The Ctables set definitions are saved with the data.

What most people don't know is that there are situations where Mult Response will give different results than the Multiple Response in Custom Tables. The Custom Tables approach is much better, but of course you can only use it if you have that module. Also, Also, understanding the base when there are multiple dichotomies can be tricky, depending on the circumstances, so be careful.

Which procedure are you talking about anyway?

I am not sure if I understand your question properly. Anyway My answer to your question is as following:

it depends what you want to do with the data, what you are trying to find out. It you wanted to know (for example) the reasons why people stopped using a product then I would use percentage of respondents so I could say that 10% of people left because of price, 15% left because of quality etc. The same person could be in both groups but that is fine.
If I was looking at content recalled from advertising however I might like to also look at the share of mentions that one feature achieved eg 20% of people mentioned "the woman in the red dress" and 15% of all mentions were "woman in the red dress". I would say that typically I would be looking at the answers based on respondents, but there may be times I would base on responses.

I hope this helps.

个案百分比----就是在所有选择（回答）这个题目的人中有多少人选择了某个选项（回答）的百分比。
响应百分比----就是在所有选项（答案）的选择数量中，某个选项（答案）被选择的百分比。