摘要：Aggregate survey responses collected from students are commonly used by universities to compare effective educational practices across program majors, and to make high-stakes decisions about the effectiveness of programs. Yet if there is too much heterogeneity among student responses within programs, the program-level averages may not appropriately represent student-level outcomes, and any decisions made based on these averages may be erroneous. Findings revealed that survey items regarding students’ perceived general learning outcomes could be appropriately aggregated to the program level for 4th-year students in the study but not for 1st-year students. Survey items concerning the learning environment were not valid for either group when aggregated to the program level. This study demonstrates the importance of considering the multilevel nature of survey results and determining the multilevel validity of program-level interpretations prior to making any conclusions based on aggregate student responses. Implications for institutional effectiveness research are discussed.
