I have what is probably a simple problem.
I'm trying to say something about sells of candy bars. I got data of sells for a population of children. There are three parameters (factors) to this data: age, income of parents, and name of a child. What I can compute from this data is, say that 34% of all sold candy bars were bought by children of age 12, 25% by children of age 10, etc. At the same time, I can compute that 52% were bought by children coming from families with monthly income X, 35% by children coming from families with monthly income Y, etc. And finally, I know that say 40% of the bars were bought by Sams, 25% by Abigails, etc.
Would I would like to deduce from these two characteristics is, given a child of certain age and name, and coming from a family with monthly income Z, what is the "probability" that he/she will buy a candy bar.
You will need to use conditional probabilities to calculate this. The basic formula for 2 events A and B is: P(A|B) = P(A and B) / P(B), where "|" reads "given that"
For three events & given your question, you will need to rename event B as a joint event of 2, say C and D. So the formula becomes:
P(A|B) = P(A|(C and D)) = P(A and C and D) / P(C and D).
Now, if you write your data into a 3-D table (or, easier, a set of 2-D tables, where each new 2-D table is for a different value of the 3rd variable), you can easily calculate all these factors, based on subtotals of rows and columns, as in the case of a single 2-D problem.