An Aptitude-Treatment Interaction Analysis of Structured Reflective Problem-Based Learning among Pre-Service Teachers: Levelling the Playing Field

Adi irawan, Turino
Nazurty, Nazurty
Effendy, Muhammad Haris
Anggereini, Evita
lestari, Putri Ayu
Maharani, Della
Agustiyara, Agustiyara

Indonesian Educational Administration and Leadership Journal (IDEAL)
Vol. 8 No. 2 (2026): On Progress

Problem-based learning (PBL) is a pedagogical approach that has gained significant traction in higher education, with a focus on fostering higher-order thinking skills. However, unstructured inquiry has been observed to have the unintended consequence of widening achievement gaps by placing excessive cognitive demands on students who demonstrate lower levels of academic achievement. The objective of this study was to investigate the differential treatment effects of Structured Reflective Problem-Based Learning (SR-PBL) - an instructional model integrating nine sequential problem-solving phases, explicit cognitive scaffolds, rotating inquiry roles, and iterative cyclical reflection - against conventional PBL. The research was grounded in Aptitude-Treatment Interaction (ATI) theory and Cognitive Load Theory. A non-randomised quasi-experimental pretest-posttest control group design was conducted among a sample of 113 pre-service elementary teachers (PGMI) enrolled in an undergraduate Environmental Education course (n = 85 across three SR-PBL cohorts; n = 28 in a conventional PBL control cohort). Moderation regression analysis revealed a statistically significant pretest-by-treatment interaction (b=-0.173, SE=0.070, t=-2.47, p=.015), confirming that the treatment effect was systematically moderated by baseline aptitude. However, conventional PBL yielded uniform, low-level improvements regardless of incoming ability (b=-0.029, p=.637). In contrast, SR-PBL exhibited a strong compensatory pattern (b=-0.202, p<.001), with the lowest-baseline learners (≤13/28) demonstrating the largest absolute gains (M=6.96 vs. 2.25 in control) and closing the post-test gap with top performers. Sub-skill decomposition across Facione's dimensions revealed that this compensatory leap was driven by targeted gains in inference a mean score gain of +1.68 raw points overall and +1.93 among low-baseline learners) and analysis +1.09 overall and +1.11 among low-baseline learners), domains where conventional PBL generated negligible growth (+0.07 and +0.25, respectively). The findings demonstrate that structured cyclical reflection acts as an indispensable cognitive equaliser in ill-structured inquiry learning.