Skip to content

Screening

Why It Matters

Job interviews, insurance menus, product lines, scholarship criteria - all are screening mechanisms. Rather than passively receiving signals, you design the game to extract information. Good screening mechanisms are incentive-compatible: people reveal their true type because it's in their interest.

The Key Move

Design a choice where different types prefer different options. High types self-select into one option, low types into another. The screen works when no type wants to mimic another - the options are designed so that misrepresentation isn't worth it.

Classic Application

Insurance menus. Offer a high-premium, low-deductible plan and a low-premium, high-deductible plan. High-risk people (who know they're high-risk) choose the first. Low-risk people choose the second. Self-selection reveals type.

Surprising Application

Job interviews as ordeal. Why do interviews involve stressful questions and long hours? Partly to screen for ability to handle pressure. But also: the very unpleasantness screens out people who aren't serious. Only those who really want the job submit to the ordeal.

Failure Modes

  • Designing screens where types don't separate - if both types prefer the same option, you learn nothing
  • Ignoring that screening is costly - you may reject good candidates or accept bad ones
  • Missing that people can game screens - if the test is known, preparation substitutes for underlying quality
  • Treating screens as fixed - good candidates learn which screens exist and how to pass them
  • Confusing screening for quality with screening for fit - different objectives require different screens

Go Deeper

Rothschild-Stiglitz on insurance markets. Standard personnel economics texts on job market screening.