This is the first study to explore the impact of premium variation across individuals, states, and time on enrollment in the State Children's Health Insurance Program and their transitions to private insurance or uninsurance in response to higher premiums. With a sample of income-eligible children from the Medical Expenditure Panel Survey, I evaluate the effect of premium changes on public and private insurance enrollment and uninsurance using a wide array of methods: Regression-Discontinuity Design for the study of the within-state variations in premiums, cross-sectional analysis for evaluating the response using across-state variation in premiums, and difference-in-differences strategies that exploit temporal variations in premiums. The main regression-discontinuity estimates point to significant declines in public enrollment along with significant increases in private take-up and no change in the rate of uninsurance. The cross-sectional results support the finding that higher premiums are associated with statistically important decrease in public enrollment and increase in private. I find no evidence of increases in the rate of uninsurance as a result of public premium increases. These results are reinforced by the longitudinal findings. They indicate a statistically significant decline in public enrollment, significant increase in private and no change in uninsurance for children in the higher-income group in response to a per dollar increase in premium over the course of 2003 year.