Research / BFI Working PaperSep 09, 2019

Panel Data and Experimental Design

Fiona Burlig, Louis Preonas, Matt Woerman

How should researchers design panel data experiments? We analytically derive the variance of panel estimators, informing power calculations in panel data settings. We generalize Frison and Pocock (1992) to fully arbitrary error structures, thereby extending McKenzie (2012) to allow for non-constant serial correlation. Using Monte Carlo simulations and real-world panel data, we demonstrate that failing to account for arbitrary serial correlation ex ante yields experiments that are incorrectly powered under proper inference. By contrast, our “serial-correlation-robust” power calculations achieve correctly powered experiments in both simulated and real data. We discuss the implications of these results, and introduce a new software package to facilitate proper power calculations in practice.

Additional Materials

More Research From These Scholars

BFI Working Paper Jan 3, 2022

Blackouts in the Developing World: The Role of Wholesale Electricity Markets

Akshaya Jha, Louis Preonas, Fiona Burlig
Topics:  Energy & Environment
BFI Working Paper Feb 25, 2024

Long-Range Forecasts as Climate Adaptation: Experimental Evidence from Developing-Country Agriculture

Fiona Burlig, Amir Jina, Erin M. Kelley, Gregory Lane, Harshil Sahai
Topics:  Energy & Environment
BFI Working Paper Feb 8, 2021

Low Energy: Estimating Electric Vehicle Electricity Use

Fiona Burlig, James Bushnell, David Rapson, Catherine Wolfram
Topics:  Energy & Environment