Purdue University

Purdue's Signals is a real-time, early intervention system that provides students with timely, concrete steps to improve their coursework, targeting specific areas where their academic success may be at risk.

• By Meg Lloyd
• 08/01/11

When it comes to retention efforts, real-time data are vital in helping institutions identify struggling students early and intervene before it's too late. Purdue University's (IN) Signals academic-intervention program is built on that premise, developing actionable intelligence based on real-time data mining and a broad range of predictive factors.

Signals uses enterprise systems already in place at the university--Blackboard Vista, SunGard Higher Education Banner, and other course and student-information management tools--to mine relevant data. A sophisticated predictive algorithm developed at Purdue using SPSS (now IBM) then analyzes the data. Signals not only identifies at-risk students by analyzing quiz or test scores; it also factors in remedial or "improvement" actions taken by a student, such as help center usage or office hour visits with instructors. The result is a more complete view of a student's status and path toward potential success or failure. It's a real-time, early intervention system that provides students with timely, concrete steps to improve their coursework, targeting specific areas where their academic success may be at risk.

From a student's point of view, Signals is straightforward. Whenever students log in to a course's home page on the CMS, they see their status for that course in color-coded terms based on a traffic light: Green represents a high likelihood of success; yellow means potential problems; while red puts them at risk of failure. Feedback to students is frequent and authoritative: As early as the second week of a semester, students may begin to receive alerts via e-mail, text messaging, phone calls, or other personal contact from instructors.

The role of the instructor in Signals is critical. Faculty members are in the best position to intervene effectively, so they are carefully trained in--and must be committed to--the use of the Signals program. "The most satisfying part of the Signals project is having faculty, student services, and information technology work together to develop new approaches for identifying students at risk," observes John Campbell, associate vice president for academic technologies. "Signals demonstrates how information technology can be a partner in solving key institutional challenges such as student success."

Signals interventions have substantially boosted student success and resulted in better student satisfaction and retention rates. In spring 2010, students in courses using Signals scored up to 26.45 percent more A or B grades (depending on the class), up to 12.65 percent fewer C's, and up to 17.27 percent fewer D's and F's than students not using the system. Over the course of the semester, 55 percent of high-risk students pulled into the moderate-risk group (earning a C grade), while 24.4 percent actually pulled into the green group (achieving an A or B). Purdue is tracking the results of Signals over time and uses QSR International's NVivo to help evaluate the impact of the project.

Signals has grown to maturity from its beginnings as a research project in 2005 to an established program impacting 8,223 students at Purdue in 2010. Purdue now partners with SunGard Higher Education to install the system at other higher ed institutions. Schools that have implemented or have plans to roll out Signals include Metropolitan Community College of Kansas City (MO), Northeast State Community College (TN), Northeast Wisconsin Technical College, and Muskegon Community College (MI).