Developing Computational Thinking in Secondary-School Informatics through a Blended Project-Based Learning Model: A Quasi-Experimental Study

Computational thinking (CT) is widely recognised as a core competence of general education, yet school informatics instruction in many national systems remains oriented towards the syntax of a programming language rather than towards transferable problem-solving strategies. This study designs and empirically tests a blended project-based learning (BPBL) model for developing CT in lower-secondary informatics. The model integrates three sequential phases — a visual block-based phase, a hybrid block-to-text transition phase, and a collaborative project phase — with asynchronous online scaffolding and structured in-class feedback. A quasi-experimental pre-test/post-test design with non-equivalent groups was implemented over twelve weeks with 164 eighth-grade students (experimental group n = 82; control group n = 82) in four public schools. CT was measured with a 40-item criterion-referenced instrument covering abstraction, decomposition, algorithmic design, and debugging (Cronbach’s α = .87); learning motivation was measured with an adapted ARCS-based scale (α = .91). Groups did not differ at pre-test, t(162) = 0.29, p = .770. At post-test the experimental group significantly outperformed the control group, t(162) = 6.76, p < .001, Cohen’s d = 1.06; ANCOVA with pre-test scores as covariate confirmed the effect, F(1, 161) = 48.36, p < .001, partial η² = .231. The largest gains occurred in algorithmic design (d = 1.33) and abstraction (d = 1.04), the smallest in debugging (d = 0.70). Motivation also favoured the experimental group (d = 1.02). The findings indicate that sequencing representational transition and embedding CT components as explicit instructional objectives — rather than treating them as by-products of coding practice — produces substantial and broadly distributed learning gains.