Domestic approaches to compliance with international commitments often presume that international law has a distinct effect on the beliefs and preferences of national publics. Studies attempting to estimate the consequences of international law unfortunately face a wide range of empirical and methodological challenges. This article uses an experimental design embedded in two U.S. national surveys to offer direct systematic evidence of international law's effect on mass attitudes. To provide a relatively tough test for international law, the surveys examine public attitudes toward the use of torture, an issue in which national security concerns are often considered paramount. Contrary to the common contention of international law's inefficacy, I find that legal commitments have a discernible impact on public support for the use of torture. The effect of international law is also strongest in those contexts where pressures to resort to torture are at their highest. However, the effects of different dimensions in the level of international agreements' legalization are far from uniform. In contrast to the attention often devoted to binding rules, I find that the level of obligation seems to make little difference on public attitudes toward torture. Rather, the relative precision of the rules, along with the degree to which enforcement is delegated to third parties, plays a much greater role in shaping public preferences. Across both international law and legalization, an individual's political ideology also exerts a strong mediating effect, though in varying directions depending on the design of the agreement. The findings have implications for understanding the overall impact of international law on domestic actors, the importance of institutional design, and the role of political ideology on compliance with international agreements.