Establishes assumption-lean weak limits for WIPW estimators in two-stage adaptive experiments, characterizes phase transitions across signal regimes, and provides a simulation-based method for hypothesis testing under non-normal limits.
By the Lipschitz property of min {1 − ¯l, max{¯l, x}} in x, we have |HN(s, W1) − H (2)(s)| ≤ | e(s, hN(W1)) − e(s, h(W1, c))|
1 Pith paper cite this work. Polarity classification is still indexing.
1
Pith paper citing it
fields
math.ST 1years
2025 1verdicts
UNVERDICTED 1representative citing papers
citing papers explorer
-
Assumption-lean weak limits and tests for two-stage adaptive experiments
Establishes assumption-lean weak limits for WIPW estimators in two-stage adaptive experiments, characterizes phase transitions across signal regimes, and provides a simulation-based method for hypothesis testing under non-normal limits.