Upload
matthew-d-pistilli
View
87
Download
1
Embed Size (px)
DESCRIPTION
Overview of analytics for the Gateways to Completion Community of Practice meeting, March 22, 2014
Citation preview
PREDICTIVE ANALYTICS OVERVIEW
/ PREVIEW
Matthew D. Pistilli, Ph.D.Research ScientistOffice of Institutional Research, Assessment & Evaluation Purdue University
[email protected] | @mdpistilli
March 22, 2014
Challenge: How do you find the student at risk?
http://www.youthareawesome.com/wp-content/uploads/2010/10/wheres-waldo1.jpg
http://www.youthareawesome.com/wp-content/uploads/2010/10/wheres-waldo1.jpg
Challenge: How do you find the student at risk?
http://classhack.com/post/76426005382/waldo
• Actionable intelligence• Moving research to practice• Basis for design, pedagogy,
self-awareness• Changing institutional
culture• Understanding the
limitations and risks
Analytics is about...
DEFINITIONS
Using analytic techniques to help target instructional, curricular, and support resources to support the achievement of specific learning goals (van Bareneveld, Arnold, & Campbell, 2012)
the process of developing actionable insights through problem definition and the application of statistical models and analysis against existing and/or simulated future data (Cooper, 2012)
http://www.gravitatedesign.com/wp/wp-content/uploads/SEO-data.jpg
THE BIG QUESTIONS
What can institutions do to improve student success?
How can institutions help students take advantage of existing campus resources?
What existing information on campus can be utilized to better identify students at risk?
How can students become self-aware of what effort is necessary to be successful in college?
How can analytics make a strategic impact at scale?
ANALYTICS IN G2C
OUR PREMISE
Ambient data Parsimony Focused on students
THE DEVELOPMENT PROCESS
Basic model constructed Four institutions to provide data for model
building and testing Model to be tested, revised, retested,
revised, etc. Anticipated roll out early summer Anticipated use by institutions this fall
MODEL BASICS
5 “buckets” of data Each bucket weighted
Largest weight placed on current academic performance and interaction with the course
The buckets: Student academic effort Current student performance Historical student performance Student demographics Student behavior out of class
Specific data to be used TBD based on model testing
WORTH NOTING…
http://i.imgur.com/nZArTnc.jpg
EXPECTATIONS REALITY
Plug and Play Immediate results Solve every problem –
ever! Universal adoption Everyone would love
it!
Fits, starts, reboots Mostly long term
outcomes Solve some problems,
create some new problems
Lackluster use Not everyone loved it
RESULTS… A LONG TIME COMING
Immediate Few Maybe noticed by instructors Possibly noticed by help centers
Short term (1 term out) Some Based in final grades earned compared to previous terms
Medium term (2 terms out) A few more Success of students in sequential courses One-year retention now available
Long term (3-4 years out) Retention over time knowable Graduation rates now available
INSTITUTIONAL CHALLENGES
Data in many places, “owned” by many people/organizations
Different processes, procedures, and regulations depending on data owner
Everyone can see potential, but all want something slightly different
Sustainability – “can’t you just…” Faculty participation is essential Staffing is a challenge
NEW POSSIBILITIES
Using data that exists on campus Taking advantages of existing programs Bringing a “complete picture” beyond academics Focusing on the “Action” in “Actionable
Intelligence”
PREDICTIVE ANALYTICS OVERVIEW
/ PREVIEW
Matthew D. Pistilli, Ph.D.Research ScientistOffice of Institutional Research, Assessment & Evaluation Purdue University
[email protected] | @mdpistilli
March 22, 2014