Combined inner and outer loop feedback in an intelligent tutoring system for statistics in higher education. (7th September 2020)

• Record Type: Journal Article
• Authors: Tacoma, Sietske; Drijvers, Paul; Jeuring, Johan
• Is Part Of: Journal of computer assisted learning. Volume 37:Number 2(2021)
• Page Start: 319
• Page End: 332
• Publication Date: 2020-09-07
• Subjects: domain reasoner -- feedback -- inspectable student models -- intelligent tutoring systems -- statistics education; Computer-assisted instruction -- Periodicals; 371.334
• DOI: 10.1111/jcal.12491
• Languages: English
• ISSNs: 0266-4909

Abstract

Abstract: Intelligent tutoring systems (ITSs) can provide inner loop feedback about steps within tasks, and outer loop feedback about performance on multiple tasks. While research typically addresses these feedback types separately, many ITSs offer them simultaneously. This study evaluates the effects of providing combined inner and outer loop feedback on social sciences students' learning process and performance in a first‐year university statistics course. In a 2 x 2 factorial design (elaborate inner loop vs. minimal inner loop and outer loop vs. no outer loop feedback) with 521 participants, the effects of both feedback types and their combination were assessed through multiple linear regression models. Results showed mixed effects, depending on students' prior knowledge and experience, and no overall effects on course performance. Students tended to use outer loop feedback less when also receiving elaborate inner loop feedback. We therefore recommend introducing feedback types one by one and offering them for substantial periods of time. Lay Description: What is already known: Intelligent tutoring systems often provide feedback at the level of steps within tasks (inner loop) and at the level of multiple tasks (outer loop). Inner loop feedback may support learning by guiding students through the solution process of a single task Outer loop feedback may support learning by advising or selecting appropriate tasks and study material Whether both feedback types indeed support learning depends on many design choices and student characteristics. What this paper adds: Students who receive elaborate inner and outer loop feedback tend to use outer loop feedback less. In this implementation, inner and outer loop feedback did not influence each other's effects on course performance. Lower‐achieving students benefit more from outer loop feedback than higher‐achieving students. Students need time to familiarize themselves with elaborate inner loop feedback. Implications for practitioners: Providing both inner and outer loop feedback may increase the number of students receiving useful feedback. This study advises to implement feedback types one by one and allow sufficient familiarization time and guidance.