Design-based inference for incomplete block designs

NICOLE PASHLEY – RUTGERS UNIVERSITY 

ABSTRACT

Design-based inference relies on the random assignment of units into treatment arms as the basis for inference, avoiding standard model-based assumptions. This talk develops novel tools for conducting finite-population design-based inference for complex experiments, focusing on incomplete block designs. These designs are a natural alternative to the complete block design when resource or other constraints limit the number of treatments that can be assigned within a block. To assist practitioners in understanding the trade-offs of using these designs, precision comparisons are made to standard estimators for the complete block, cluster-randomized, and completely randomized designs.

This is joint work with Taehyeon Koo (Rutgers University).

RELATED PAPER:

• https://arxiv.org/pdf/2405.19312