This study used a multi-theoretical approach to examine the associations between institutional policies on the assessment of faculty pedagogy and faculty's use of learner-centred assessment (LCA) practices in their undergraduate classrooms in the United States. We found strong evidence that it was not the number of methods but the types of methods used by institutions to evaluate teaching that were significantly associated with the extent to which faculty use LCA practices in their undergraduate courses. Our results returned conflicting patterns: Assessments that were conducted top-down--those by deans and department heads--were associated with a reduced rate of LCA use in classrooms, whereas assessments that relied on data from students, peers, or faculty themselves were associated with a greater rate of LCA use. Yet, most student-centric methods to evaluate teaching, including the popular use of student evaluations of teaching, were unrelated to the incorporation of effective pedagogy. The pattern of results did not vary significantly across different types of higher education institutions.