Advocates of school choice policies claim that private school competition would make public schools more efficient. I evaluate the effect of private school competition on public schools' test scores in Arizona. Private schooling is treated as endogenous. I use the Arizona K-12 tax credit to build the first instrument based on religious denominations likely to donate to private schools. Also, I use measures of the population of private school students and schools as new instruments for private schooling. Moreover, I build grade-specific measures of private school competition. I find a strong positive effect of private school competition on scores in grade 4, and no significant effect on scores in grade 7.