Setting Graduation Rate Thresholds

Abstract

The federal government, with the ill-fated State Postsecondary Review Program (SPRP), required the State Postsecondary Review Entities (SPREs) to develop quantitative performance standards. An essential part of the requirements was to set performance thresholds for targeting institutions to be reviewed. The work of the SPREs on one standard in particular, the idea of a graduation rate threshold, may still be useful in assessing an institution's graduation rate compared to other institutions or a selected group of peers. Although the SPRP no longer exists, many states have moved in the direction of either incentive funding or performance funding based on various performance indicators. One of the indicators nearly always selected to be associated with incentive or performance funding is graduation rate. This article is based on work done by the authors as members of a committee to set graduation standards for the SPRE. The article reviews the completion/graduation rate thresholds developed by several other states and discusses the advantages and disadvantages of the various methods. Alternative statistical approaches to establishing performance thresholds for graduation rates are also discussed, including the use of the one standard deviation lower bound method, the logit prediction bound method, the linear regression method, and the logistic regression method.