Pyequate: A Python Package for Test Equating

Test equating adjusts scores across alternate forms of a test so that they may be used interchangeably, and is a routine requirement of any testing program that administers more than one form. Software for equating is well established in R and in standalone programs. There is no current native option in Python, despite its role as the working environment for a growing share of assessment data and downstream modeling. Analysts working in Python must therefore leave their pipeline or reimplement methods themselves. The package pyequate closes this gap. It provides observed-score equating under random groups and common-item nonequivalent groups designs, including the mean, linear, equipercentile, Tucker, Levine, frequency estimation, and Braun-Holland methods, together with presmoothing and analytical standard errors of equating. It also provides item response theory parameter linking and true-score and observed-score equating.