The Problem with the Delta Cost Project Database

The Integrated Postsecondary Education System (IPEDS) collects data on Title IV institutions. The Delta Cost Project (DCP) integrated data from multiple IPEDS survey components into a public-use longitudinal dataset. The DCP Database was the basis for dozens of journal articles and a series of influential policy reports. Unfortunately, a flaw in the construction of the DCP Database may make it inappropriate for particular analyses. Specifically, the DCP Database often collapsed data from state systems, which consist of multiple Title IV institutions, into a single observation. For example, the University of Texas-Austin observation contained data from all Title IV Institutions in the UT system (e.g. UT-Dallas, UT-Brownsville). This research note investigates how many institutions were affected by this problem, identifies the extent to which published research used the DCP Database in potentially inappropriate ways, and conducts selected analyses to understand whether the problem can affect empirical results. Results show that the problem was concentrated in the public sector but only affected a small proportion of public institutions. However, analyses suggested that this problem can substantively affect empirical results. Therefore, we argue that the DCP Database should not be used to analyze public institutions. We conclude by discussing the creation of alternative databases.