The Mentors Matter Recruitment initiative leveraged state administrative data to recommend and successfully recruit more instructionally effective and experienced teachers to serve as clinical mentors, which, in turn, increased pre-service teachers' readiness to teach. Prior results, however, focus on self-reported outcomes and stem from a single field experiment, raising questions about their replicability. In this paper, we replicate the previous study with a second cohort, finding again that the initiative led to the recruitment of more instructionally effective and experienced mentors. In addition, we examine new outcomes from administrative and program data, observing that candidates assigned to mentors recruited through our lists were rated as significantly more instructionally effective during student teaching. Given these promising results, we tested the same initiative at three new programs under less oversight from research partners and policymakers, finding that even under these more typical conditions, the initiative successfully produced significantly more instructionally effective and experienced mentors for a third time, though with smaller effects. [This is the online version of an article published in "Journal of Research on Educational Effectiveness."]