The amphibian metamorphosis assay (AMA) is an in vivo screen to assess potential interactions of chemicals with the amphibian thyroid system. Tadpoles are exposed for 21-days, then assessed for development and growth after 7 days and at test termination. This paper presents data from studies performed to satisfy test orders from the US EPA's Endocrine Disruptor Screening Program. Data Evaluation Records were used to collate the control variability and performance of biological endpoints in AMAs conducted in different laboratories, then supplemented with recent studies. We examine the statistical power of AMA endpoint analysis and assess whether historical control data (HCD) can assist evidence-based interpretation of the endpoints, with 52 studies from 7 different laboratories. HCD can be used to understand assay performance post validation. The analysis identifies some need for flexibility in the interpretation of the Test Guidelines' performance criteria, including latitude with analytical variability and statistical analysis of late-stage animals. Additionally, more guidance is suggested for feed regiments and the selection criteria for batches of animals to initiate the assay. Potential Guideline refinements that improve interpretation of the data and have potential to reduce the number of vertebrate animals used in the conduct of AMAs are identified and discussed.