Hsu et al and Huettel et al have both argued that these two processes are qualitatively different and have correspondingly different neural substrates (although they disagree on what these substrates are).
I can't figure out how risk and ambiguity are qualitatively different things. It just seems like what we call ambiguity is really just "risk that takes more mental effort to evaluate".
I like concrete examples, so here are two:
1. From Colin Camerer
A. (risk) Was the temperature in New York on Oct 17, 1996 at 3 pm greater or less than 60 degrees?
B. (risk+ambiguity) Was the temperature in Bishkek, Kyrgyzistan on Oct 17, 1996, at 3 pm greater or less than 60 degrees?
2. Ellsberg Paradox
There are three balls in a bag. If I draw a green one, you win $15, if I draw a red one, you win $0. What is the value of the gamble if the ratio is:
A. (risk) Two green, one red.
B. (risk + ambiguity) One green, two unknown.
In the risk situation of 2, the EV is $10. In the ambiguity situation, the EV is also $10, although the computation is more difficult to perform in one's head. So normatively, the choices are the same (I'm saying that ambiguity is not a valid normative concept). Descriptively, valuation of a gamble is inversely correlated with ambiguity.
It seems intuitively that ambiguity aversion is explained by well-known heuristic biases. For example, people have been to NYC, so NYC is more mentally accessible - so people are more likely to be overconfident in their judgements about its temperature. Similarly, people are more likely to be overconfident about simple calculations than more complex ones.
Many people would probably not be bothered by the contention that ambiguity aversion = risk aversion + some additional biases. But this contradicts the claims made in Huettel et al, e.g. "[overcoming ambiguity aversion] may require specialized neural mechanisms" and "decision making under ambiguity does not represent a special, more complex case of risky decision making; instead, these two forms of uncertainty are supported by distinct mechanisms."
I personally find it hard to imagine that risk and ambiguity are different, simply because risk without ambiguity is not natural. Foraging animals never have really have pure risk (i.e. stochastic outcomes without ambiguity) to deal with, right?
Thoughts?
