Have you studied sampling theory or probability? As long as you have ~10+ in each category, your sample is large enough for a meaningful result. If you have le

Have you studied sampling theory or probability? As long as you have ~10+ in each category, your sample is large enough for a meaningful result. If you have less than 10, you might still have a good sample, but you might need to do a more accurate statistical test to find out. That is, if you sample 1000 people, you are good from 1% to 99% range (where you'd have 990 to 10, and 10-to-990 respectively). But you can't draw conclusions if your results say 99.5% (because then you'd only have 995-to-5, and 5 is smaller than 10). If you split the sample into further groups (yes, no, maybe), each grouping should have 10+ results in it if you want anything close to the truth.