(no subject)
From the first chapter of Diego Gambetta's "Codes of the Underworld":The probability of making identification mistakes is 'frequency dependent': the higher the proportion of criminals in the search environment, the lower the risk of approaching the wrong people. Where corruption is known to be widespread, for instance, corrupting others or signaling one's willingness to accept bribes is not much of a problem. If the probability of encountering a corrupt agent is correctly believed to be high, criminals will rationally try more and bolder approaches, and will easily uncover corrupt partners.
laterOne way to acquire good evidence of someone's criminality, which intersects with the referral method, exploits law enforcement itself. In terms of how effectively they can mix and match kindred sprits, the hangouts that criminals freely choose cannot compete with the places they are forced to go by the agents who fight against crime: there is nothing like prison to mix like with like. Prisons promote crime in obvious ways, teaching criminals new skills and brutal modes of behavior, but they also do so, less obviously, by shouldering the costs of advertising, and identifying who is a criminal to begin with.
Referred there by Chis Blattman
laterOne way to acquire good evidence of someone's criminality, which intersects with the referral method, exploits law enforcement itself. In terms of how effectively they can mix and match kindred sprits, the hangouts that criminals freely choose cannot compete with the places they are forced to go by the agents who fight against crime: there is nothing like prison to mix like with like. Prisons promote crime in obvious ways, teaching criminals new skills and brutal modes of behavior, but they also do so, less obviously, by shouldering the costs of advertising, and identifying who is a criminal to begin with.
Referred there by Chis Blattman