Can We Judge People by What They Look Like? In Fact, Yes

Re:Can We Judge People by What They Look Like? In Fact, Yes by jimhefferon on Tuesday April 17, 2018 @06:59AM (#16857180)

Suppose that you could believe studies of the kind that finds athiests have more of a tendency to have bad skin. And suppose that it found 80% of athiests do have skin quality that is below the median for whatever group of fresh-persons was being tested. Now I meet a person whose skin appears to me to be sub-median. How is it actionable?

Even if I restrict myself to things that many of us might think have some basis, such as: people with glasses are often more studious. People come in for a job interview. Do I have under the strengths column a "glassses" checkbox?

[ Parent | on HN ]

Re:Can We Judge People by What They Look Like? In Fact, Yes by DoofusOfDeath on Tuesday April 17, 2018 @07:03AM (#16857200)

Supposing that the book's claims are accurate, it would provide some Bayesian justification for police profiling at traffic stops, choosing whom to approach at a bar, etc.

(Standard caveats apply.)

[ Parent | on HN ]

Re:Can We Judge People by What They Look Like? In Fact, Yes by AstralStorm on Tuesday April 17, 2018 @07:27AM (#16857323)

What is the prior in this logic?

Daily reminder: Bayesian logic does not deal in quantifiers. Trying to represent quantifier based logic in Bayesian systems results in absurd. (The update function does not work.)

Thus it cannot be used to justify or counter existential statements.

It can tell you: given prior A and conditional statements B through W, the chain probability assuming distribution Z may be Y. It cannot easily deal with truncated chains either.

[ Parent | on HN ]