It is well known that inhalation of 222Rn and 222Rn decay products increases the risk of lung cancer. While the occurrences of high radon areas in the United States are generally known, studies examining the temporal yearly radon variation in homes across different regions are lacking. This information is essential to assess the ability of a year-long radon measurement to predict the future radon concentration in a home or reconstruct the retrospective residential radon concentration. The purpose of this study is to help fill this gap by examining the temporal variation of residential radon concentrations in homes over several years as well as to explore factors that affect the yearly temporal variability of residential radon concentrations. The coefficient of variation was used as a measure of relative variation between multiple measurements performed across homes over several years. Generalized linear model analyses were applied to investigate factors affecting the coefficient of variation. The median coefficient of variation between the first and second test period was 12%, while a median coefficient of variation of 19% was found between the first and third test period. Factors impacting the coefficients of variation were found to vary for different types of homes and by floors of a home. This study provides important insights into the uncertainty of residential radon gas concentrations that can be incorporated into the sensitivity analyses for the risk estimates of both the North American and global pooling of residential radon studies to improve risk estimates.