Dalton State College-[node:field-date:custom:Y]--Predictive Analytics

Goal 4 – Intrusive Advising -- Strategy 4.2 – Predictive analytics

In order to understand reasons for low retention and graduation rates, several sources of data were investigated.  In an effort to increase the effectiveness of academic advising, we determined a need for predictive analytics to help advisors improve guidance of students in selecting the appropriate major based on their abilities and skills, identify students who are off track, and help students understand the likelihood of success in a given program of study.  We believe that better matching of students with majors as early as possible will reduce the number of D/F/Ws, course repeats, major changes, and credit hours in excess of what is required for a given degree or certificate, thus reducing time to degree and increasing the likelihood of successful completion.  First, primary reasons for withdrawals from classes was determined.   Data collection on these points revealed that over a period of two years (2013-2015), student claimed to withdraw from classes primarily for academic-related reasons, not personal ones as commonly assumed; 25.3% reported withdrawing in order to protect GPA, 21.8% due to academic difficulty, 18.4% due to work conflicts, and 9.6% due to “course not needed.”  Additionally, IPEDS data was accessed to discern the number of graduates who earn 150% of required hours for the credential (or less), which indicates the number of graduates who are taking excess hours.   For the 2006 cohort, 20/134 earning the bachelor’s degree earned within the 150% benchmark (14.9%); for the 2007 cohort, 24/116 (20.7%); for the 2008 cohort, 40/196 (20.4%).  In terms of all credentials (bachelor’s, associate of science, arts, applied science, and certificates), the respective numbers are 114/662 (17.2%), 110/768 (14.3%), and 102/745 (13.7%).  Finally, the number of major changes was found to be significant:  766 students changed their majors in Fall 2013 and 508 in Fall 2014. 

These three sources of DWF rates and low retention:  excess hours, reasons for withdrawing, and--changes in major can at some level be attributed to advising gaps; therefore, attention to advising forms a basis for our strategy.  In an effort to increase the effectiveness of academic advising, we determined a need for predictive analytics to help advisors improve guidance of students in selecting the appropriate major based on their abilities and skills, identify students who are off track, and help students understand the likelihood of success in a given program of study.  We believe that better matching of students with majors as early as possible will reduce the number of D/F/Ws, course repeats, major changes, and credit hours in excess of what is required for a given degree or certificate, thus reducing time to degree and increasing the likelihood of successful completion. 

The initial impediment to this strategy of improving advising was funding to purchase and implement the necessary software.  Thus we requested and were approved for funds in our FY15 budget to join the Education Advisory Board’s (EAB), which included purchase and implementation of their predictive analytics software.  The EAB platform, which is being used at other USG institutions successfully, provides advisors with relevant student data that is formatted to expedite and facilitate the advising conversation.  The interface indicates risk levels of students in terms of likelihood of successfully completing their programs, the strength of the advisee in different academic area, likelihood of successful completion of courses, and recommendations and information about majors that are deemed a good fit for the student based on past academic behavior.

During FY2015 and 2016, an EAB dedicated consultant has worked closely with Office of Academic Affairs, professional advisors, informational technology staff, and department heads to install and implement the predictive analytics advising program, create success markers for academic programs, and train faculty in its use.  After a pilot stage in Summer 2014 involving some professional advising staff and some faculty in the STEM disciplines, the EAB dedicated consultant visited the campus for a kickoff in August.  At this plenary session, the faculty were introduced to the platform’s functionality. Beginning January 2015 through July Fall 201515, 16 multiple training sessions were held for faculty and new professional advisors, allowing 102 160 personnel to become proficient in the basic use of the platform.  EAB sends DSC monthly usage reports; as of now, the professional advisors are the primary users, but by April 20 there had been 1386 logins by users, with 167 in March alone.

Now that over 960% of the faculty have been trained and have been provided with follow-up materials, the next step is to continue to grow usage of the platform.  The first target is to increase utilization of the EAB platform.  EAB provides monthly reports of utilization.  To this point, the primary users of EAB have been the eight professional advisors assigned to the five academic schools.    Additionally, a student survey of satisfaction with advising has not been performed in several years, so a survey will be administered in early Fall 2015 to gauge students’ attitudes on strengths and weaknesses of advising in general.  Because faculty and advisor utilization of the EAB platform is beginning to gather steam now in 2015, we we currently have no long-term data to support its efficacy, but the college will continue to track DWF, course repeats, major changes, and credit hours at graduation as well as utilization to assess EAB’sits value. 

However, as of this point, DWF rates have declined since 2011 due to several reasons, one being that students are required to obtain the faculty member’s signature before dropping a class.  Other efforts at improved teaching, raising entrance test scores, emphasizing DWF rates in faculty evaluation, addressing “killer courses,” and improving remediation have also addressed DWF rates. In Fall 2011, over 20% of hours resulted in DWF grades, so that the completion rates of courses was 79.26% (46,834 hours completed out of 59,090 attempted).  By Fall 2013, the course completion rate was 85.32% (46971 hours completed out of 55054 attempted), and in Fall 2014, 85.82%, thus, a 6.5% improvement.