International treaties face the challenge of selection bias in measuring their impact.

But treaties can also constrain state behavior, which von Stein's analysis didn't fully capture.

Critics question von Stein's estimator due to its strong assumptions and statistical issues.

However, his estimate doesn't explain the differing results between von Stein (2005) and Simmons (2000).

The authors frame this as a model dependency problem instead.

Using preprocessing matching techniques reduces that dependency significantly.

This approach yields treaty effects estimates that are both substantively meaningful and statistically robust, aligning closely with Simmons's findings.