Certainty Preference, Random Choice, and Loss Aversion: A Comment on "Violence and Risk Preference: Experimental Evidence from Afghanistan"