In January WCET (WICHE Cooperative for Educational Technologies) successfully concluded the proof-of-concept phase of its work on the Predictive Analytics Reporting (PAR) Framework, a longitudinal data-mining project. Funded by the Bill & Melinda Gates Foundation, the PAR Framework employs predictive techniques commonly found in business to aid in educational decision making. The PAR Framework team and its partners created a single, federated dataset of anonymized student records, thanks to the contributions of six institutional partners, all WCET members, who worked together to create the dataset and to normalize 33 variables (and nine constructed variables) common to all of them.

The PAR Framework makes it possible to conduct predictive analyses of a massive collection of student records to better understand the variables affecting why students drop out or succeed in completing their programs. The ability to reliably predict behaviors and outcomes is enabled through the application of descriptive, inferential, and predictive tests and analyses on massive numbers of records that have been normalized around common data definitions. No such data set currently exists in American postsecondary education. The PAR Framework team has conducted a preliminary analysis, looking for patterns that provide new insights into our understanding of student loss and momentum. The analysis team is now working closely with institutional partners to explore intra- and inter-institutional data patterns and results.

March 2012  | Share this on Twitter | Post this on Facebook