Hi all Not completely sure this is the right place for this, but anyway... Does anyone know what kind of algorithms those personality quizzes you see use? You know the ones...What animal are you? etc. Obviously a basic one could be used that simply associates each answer with a profile and then returns the profile with the most amount of answers. But there would be an issue if more than one profile gets the same amount of answers associated with it. For example: 10 questions, 3 possible profiles (A, B and C) 4 questions are answered matching with profile A 4 match with profile B 2 match with profile C In that scenario, how do you decide between profile A and B? Any thoughts on this would be appreciated. If you've ever designed a quiz like this, how did you do it? Thanks!
What you'd normally do is weight the questions and answers. To copy your example: 10 questions, 3 alternatives, A, B and C We want to get a specific value, so we weight the questions first, or simply just do a scaling - start with the least important and increase importance, or vice versa. Then you weight the answers for each question. In the end, you'll end up with something like this: Question Answer 1. . . . . . . . A - 1 2. . . . . . . . A - 3 3. . . . . . . . B - 5 4. . . . . . . . C - 6 5. . . . . . . . C - 7 6. . . . . . . . B - 8 7. . . . . . . . A - 10 8. . . . . . . . A - 15 9. . . . . . . . B - 20 10. . . . . . . B - 25 This leaves us with the same problem that you had originally, except this time, the answers are weighted. So, the total score are as follows: A - 29 B - 58 C - 15 Hence the result will be B However, there are lots of ways to weight, and there's also the possibility of having more results than just one for each answer - you can also create connections between answers, so that if a user answers A-B-B-C that person get a different score from someone answering A-A-B-C