Abstract:

Worldwide, university student completion rates are a pressing concern; considering solutions to this felt issue, it is widely theorized that development of connections to university people and resources may be impactful in promoting student persistence. This research describes the development and roll-out of a mobile platform, Connect Engage and Stay (CEAS), which sends targeted messages to university students identified as at risk for dropping out. These CEAS messages harness mobile technology to connect students with appropriate university people and resources that may help the students overcome barriers to continued study.

Our previous research analyzed university student text message data from a closed university social network to mine for patterns in student language use that may be cumulatively predictive of student drop out. Using corpus linguistics tools and systemic functional linguistic theory (i.e. computerized language analysis techniques), we found that, in comparing students who stay and students who drop out, students at risk for drop out asked fewer questions, expressed less certainty, used less punctuation, used fewer emoticons and expressed more negative sentiment.

Building on these previous findings of using student text messages to predict drop-out, our multidisciplinary university team turned our attention to the development of a system, Connect Engage and Stay (CEAS). CEAS sends automated, targeted, just-in-time mobile messages to university students identified as at risk for drop out – these messages give students a way to use their mobiles to connect with university people and resources needed to overcome financial, academic, and logistical challenges at the moment so as to promote student retention. The presentation will feature the results of CEAS implementation for a group of 998 students as well as brief technology demonstration; CEAS presents much promise as a cost-effective way to combat student drop out.

Keywords:

University student persistence, university student drop out, social media, corpus linguistics.