We report on a randomized controlled trial used to examine the effect of dynamic pricing when applied to households with rooftop photovoltaic (PV) power-generation systems. Using high-frequency data on household-level electricity use, PV generation, purchases, and sales, we find that critical peak pricing induced significant usage reductions of 3-4% among households with PV systems, a quarter of the effect size seen among average households without solar PV systems. In addition, we investigate the influence of the amount of PV power generated on treatment effects and the potential heterogeneity caused by participating households' attributes. This is the first large-scale field experiment evaluating the demand response of households with PV generation capabilities.