Prior studies conclude that climate plays one of the most important roles in driving variations in residential electricity consumption. While some past studies have quantified sensitivities of electricity use to ambient temperature, 1) few previous studies utilize both high temporal and spatial resolution electricity data, and 2) no research to our knowledge has investigated how the temporal and spatial resolution of electricity data, and choice of ambient temperature indicators, affects quantification of these sensitivities. In this study, we use smart meter data records of electricity use for 1245 households across California, along with hourly ambient temperature records, to compute electricity-temperature sensitivities using a segmented linear regression approach. We find that electricity use and temperature show the strongest relationships when computed using daily accumulated electricity use and daily average temperatures; using these metrics results in a mean electricity-temperature sensitivity of 0.11 kW degrees C-1. This value is higher than corresponding sensitivities computed using spatially aggregated data, with values ranging from 0.097-0.10 kW degrees C-1 depending on the amount of spatial aggregation. Through presenting probability density functions of household-level electricity-temperature sensitivities, we illustrate insights that can be gleaned using high resolution electricity datasets such as that used here. We note that values of electricity -temperature sensitivity reported here are representative of the 1245 households under investigation. (C) 2018 Published by Elsevier B.V.