370 likes | 480 Views
Increasing parent engagement in student learning using an ITS with automated messages. A thesis presentation for the degree of Master of Science in Computer Science by Zachary Broderick Advisor: Neil Heffernan Reader: Carolina Ruiz March 1 st , 2011. Overview and Outline. Introduction
E N D
Increasing parent engagement in student learning using an ITS with automated messages. A thesis presentation for the degree of Master of Science in Computer Science by Zachary Broderick Advisor: Neil Heffernan Reader: Carolina Ruiz March 1st, 2011
Overview and Outline Introduction Background ASSISTments Component Development Exploratory Study Methods Results Discussion Experiment Methods Results Discussion
Introduction > Background Intuition and literature: Parental involvement in student learning is beneficial (homework too) Contingent on access to information Controlled by student Often limited to report card Direct communication with teacher costly
Introduction > Background Intelligent Tutoring Systems Tutoring in software Collects detailed learning data used by teachers and students Give access to parents Possibilities Carnegie Learning, Andes, IMMEX PowerSchool, Study Island
Introduction > Background ASSISTments Free, online ITS developed at WPI Teachers create virtual classroom Students complete assignments Data presented in reports
Component Development Parental Notification Component Teachers give parents accounts Access to reports Messages from teacher Automated Messages
Exploratory Study New area of research Does the code work? Will parents be able to use the data? Will it improve engagement? What is teacher/parent feedback? What will we need to do to test it?
Exploratory Study > Methods Fall Pilot Test Conducted at local middle school 2 ASSISTments (PIMSE) partner teachers 4 classes of 20 students, 8th grade math All parents invited to sign up Few logged in again after signing up
Exploratory Study > Methods Spring Experiment Send messages from teachers to remind parents to log in Pre/post survey to measure engagement Student survey as well Compare server logs against Fall pilot Monitor student performance
Exploratory Study > Results Parents logged in more Parents felt more engaged Student performance did not improve
Exploratory Study > Discussion Conclusion: Parents need to be reminded Need stronger intervention Automated Messages: More frequent, higher coverage Parents feel more engaged but… Need stronger experimental design to detect more reliably (noisy data) Might even detect performance gains Positive feedback
Experiment > Discussion Results indicated: Parent engagement increased Homework completion increased Performance did not increase Positive feedback or not needed Statistically: Only a few reliable results Would not stand up to correction All trended strongly in right direction High effect sizes
Experiment > Discussion Results present, hard to detect Noisy, real-world data IEP’s, internet access, ELL, abscenses Small sample sizes Ceiling effect Confound between units Causal chain
Experiment > Future Work More intervention Avoid spamming Text messages? Conduct in spring w/ similar units Use lower-knowledge students Only send messages when there is a problem?
Acknowledgements Thanks to: Oak Middle School Christine O’Connor, Courtney Mulcahy, Kevin DeNolf, Jen Dufaul, Christ Starczewski ASSISTments Team Neil and Cristina Heffernan Developers National Science Foundation GK-12/PIMSE Grant