Friday, November 8, 2013

Predictability of behavior

Predictability of behavior.

Analyzing a character is basically profiling. Bootstrapping. You start with the behavior, and work back towards personality based on your understanding of psychological mechanisms. The problem with this, and it's a problem I struggle with every day, is that profiling assumes a stability of personality characteristics, or traits. It must, to work at all. If people who behave the same way do not share the same personality traits (the homology assumption) then the whole thing is out the window.

Unfortunately, there is nothing to prove that personality traits are stable. Personality constructs however, like Eisenck's Extraversion or Hare's Primary Psychopathy, seem to hold constant. These are organizing constructs however, not traits. They can help you understand how traits are articulated, and they can be somewhat predictable, they can be stable. But these aren't traits.
I think it's a matter of boundaries. You can narrow down a range of behaviors someone would be likely to engage in, or not engage in for that matter, if you have enough psychological, sociological, economical, ecological and biological data on them, but as for predictability-- Predictability implies consistency. It implies that placed in similar circumstances, a person would behave in similar ways.
I link this to the dichotomy between personality traits and constructs. Traits (actions) are unstable (unpredictable) but constructs (themes) are stable (predictable).

I can only speak from personal experience (and admittedly I deal with criminals) but the consistency hypothesis is in bad shape. Our best test of consistency is the observation of serial offenders. Controlling for fluctuations in environment and victimology, it's not like on TV. Serial rapists don't go after the same type of victim, using the same modus operandi, in the same setting, engaging in the same range of sexual acts. It happens, but it's the exception rather than the rule. If you look at behaviors, it's absolutely not consistent (hence not predictable). However, if you look at themes (like under a microscope, pulling back to see the specimen rather than just a cell) then some amount of predictability can be reached. A serial rapist might always engage in sadistic acts. Different actions, same connotation to each. Or always engage in controlling acts... Like I said above, boundaries. You can't predict the actions, but you can perhaps predict their range.

There's always a difficulty in saying anything was predictable after you saw the outcome. We've been shown how John behaved, and had we not, could we have predicted it beforehand?

No comments:

Post a Comment