All simulators
Choosing a Design
Answer questions about your setting and get the design, estimand, assumptions and diagnostics.
Day 80: Phase 5 Capstone: Choosing a DesignCan you (or did someone) randomly assign the treatment?
Is there a lottery, rule or shock that changes take-up but affects outcomes only through treatment?
Is treatment (or its intensity) assigned by a known cutoff on a score, age, date or distance?
Does the rule change the slope of a schedule (a cap or kink) rather than switching treatment on?
Does crossing the cutoff switch treatment on for (almost) everyone?
Do you observe treated and untreated groups both before and after treatment starts?
Do different groups start treatment at different times?
Do you observe all the variables that drive both treatment and the outcome?
RECOMMENDED DESIGN
Fuzzy regression discontinuity
ESTIMAND
LATE for compliers at the cutoff
KEY ASSUMPTIONS
Continuity, monotonicity, exclusion of the cutoff
DIAGNOSTICS
First-stage jump (strength), all sharp-RD checks, weak-IV robust intervals
Covered on days 76–77
questions on the decision path
Answer the questions about your setting. The tool follows the decision path (highlighted questions are the ones that mattered) and recommends a design, the effect it estimates, its key assumptions and the diagnostics that make it credible, with the days where each is covered.