Description
Analyses whether signals in a product (badges, reviews, verification, credentials) can credibly separate high-quality from low-quality senders, given their costs.
In a signalling game, a sender with private information (e.g. seller quality) chooses a signal, and a receiver infers type and acts. A separating equilibrium exists only if the signal is costlier for low types than high types; otherwise low types mimic and the signal becomes uninformative (pooling). Cheap talk, costless signals, is credible only when interests align. This explains why 'verified' badges, review counts or certifications succeed or fail, and why designs that make signals cheap destroy their value. Rigorous analysis models types, costs and incentives, then tests predictions with behavioural data or experiments.
Attributes
- Discipline
- Game Theory & Decision Science
- Loss type
- Unmigrated
- Data
- Analytical / expert
- Effort
- Days
Steps
- Identify the information asymmetry: who knows what about quality, honesty or intent.
- List available signals and estimate their cost to high and low types (money, effort, risk, time).
- Determine whether the single-crossing condition holds and whether separating, pooling or semi-separating outcomes are expected.
- Check empirical data: do signals predict outcomes (ecological validity) and do receivers rely on them?
- Redesign signals to restore differential cost (e.g. verification tied to accountability, history-based badges).
- Test with field experiments or lab games, monitoring for gaming over time.
Sources
- Spence, Job market signaling (1973), Quarterly Journal of Economics
- Connelly et al., Signaling theory: A review and assessment (2011), Journal of Management
- Crawford & Sobel, Strategic information transmission (1982), Econometrica
Also known as
- Signalling theory
- Costly signalling
- Spence signalling
Categories
- Strategy & Collaboration
- Game Theory & Decision Science