>>96273476contd.
For instance, in the case of sexual partners and mental illness, it would be nice to have a measure of some sorts indicating how often the number of sexual partners is negatively or positively made a topic in conversations the person has. I think it's quite obvious that while having many sexual partners has become more acceptable for women, it has always been acceptable for men. It is a positive thing for a man to have had many sexual partners and not many people would think of calling someone like that a slut. However, being called a slut and being associated with that kind of lifestyle has an influence on your behavior. A statistical model could account for this possible intermediate factor.
This is an inquiry at a societal level. There are many nonlinear interactions leading up to the behavior of this data. These kinds of problems are very often complex problems (not complicated, complex). Complex problems are distinct from simple problems in that they cannot be solved by simple solutions. Which is why we employ statistical models.
The kinds of assertions you and your source are making generalize over a huge group of people. If made into police, these kinds of assertions would have huge societal ramifications. A good rule of thumb: The larger the generalization and its consequences, the more careful and intricate you need to model the data.