# Collect answers and determine probability

#### alfred

##### New Member
Hello!

I'm trying to figure out a method to predict the most probable answer taking into account previous answers collected in a database.

Practically, I have 5 questions. For each question I present 2 possible answers (either A or B).
People answer to the 5 questions and their answers are stored in a database.

Just looking at the frequencies in the database I know that for question 1, X people answered A and Y people answered B. In question 2, X people answered A and Y people answered B, and so on...
On a very basic level, I suppose could say that e.g. if for question 1 ten people answered A and five people B, A is the "most probable" answer I would expect (66% probability). However this method looks to me quite poor, moreover it doesn't take into account the previous, following answers...

Therefore I'd like to figure out:
• For each question, which would be a more reliable way to compute the probability of people answering A or B, taking into account the data collected previously for the same question?
• Which approach / formula should I use to calculate the most probable answers accross the entire questionnaire, according to the previously collected answers? (i.e. which is the probability that people who answered A in question 1, will answer A in question 2. And which is the probability that people who answered A in question 1, and A in question 2 will answer B in question 3... and so on...)

Thank you so much in advance for any hint and help!!

Last edited: