70 likes | 185 Views
Theories of Learning and Judgment. Last time: signalling models prospect theory On-line processing. Signaling models. Cheap talk under what conditions can a listener learn about the world from a sender who pays no costs and bears no risk for communicating?
E N D
Theories of Learning and Judgment Last time: signalling models prospect theory On-line processing
Signaling models • Cheap talk • under what conditions can a listener learn about the world from a sender who pays no costs and bears no risk for communicating? • key condition is similarity of preferences • if sender’s goals and receiver’s goals don’t conflict, no incentive to mislead; if they are congruent, this is a coordination game • true cheap talk is rare in politics because repeated interactions/overlapping generations allow for reputations to develop • reputations/brand names are commitment mechanisms; a candidate/commentator with one can be persuasive even without similarity of goals
Costly Talk • Reputation is an example of risky talk • sender puts valuable reputation up as bond; loss of reputation is example of a penalty for lying • Other examples of learning from signals • costly action: sender pays a fee to send a signal • verification: sender’s statement can be evaluated by a third party who is motivated to expose falsehoods (e.g., the press) • Preference for biased information sources • when a liberal says something good about a conservative, or vice-versa
Priming and Framing • Kahneman & Tversky’s prospect theory • conventional expected utility maximization models tend to use symmetric utility functions • K&T hypothesize an evaluative asymmetry between the domain of losses and the domain of gains • decision frames describe the psychological context surrounding an individual’s choice • priming is raising the salience of a particular issue or context • framing is the manipulation of the decision frame surrounding a person’s opinion-formation on a question
Prospect theory and politics • Implication: spin control is potentially important • binary choices can be reversed via subtle changes in decision frame • Who controls the policy/issue agenda? How much room is there for priming by strategic actors in politics? • Who controls the selection of decision frame? • do different people choose different frames? If so, why and under what conditions?
On-line processing • How do we know what people know and how they make judgments? • Lodge, McGraw and Stroh: 2 models of candidate evaluation • memory-based model: current judgment is based on a (perhaps unbiased) assessment of recalled data • impression-driven model: current judgment based on a retrieval of a “running tally” of marginal judgments; the data bits on which the judgments are based may be disposed of
Relating Prospect Theory to Judgment • “Agenda control” in campaigns can matter • If candidate A’s position on issue 1 is more popular than candidate B’s position on issue 1, and vice-versa for issue 2; and voters are on-line processers, • A focus on issue 1 raises A relative to B; staying “on message” may have a cumulative effect • If two alternative decision frames α and βcan be used to describe issue 1, with α more favorable to candidate A and β more favorable to candidate B • Messages that use framing α will tend to raise A relative to B; staying “on message” may have a cumulative effect