Yup, it's just a matter of determining what those assumptions have to be. If a model's used outside of its relevant scope--well, that's bad news.
Originally Posted by Jennifer
If the set of assumptions winds up including lots and lots of unrealistic caveats, then it'll become useless in terms of real-world applicability.
"Okay, we've determined that the Big Five's agreeableness factor doesn't apply to Belgium."
"Okay, we've further determined that the Big Five's agreeableness factor doesn't apply to any of Europe."
"... okay... it looks like three of the factors are just total bullshit."
".. okay. Turns out the Big Five only applies to people who lived on Mars in the year 1873."
".. uhgh. Fuck the whole thing."