Development of a Computational Thinking Assessment Instrument for the Renewable Energy Topic Dian Purnama (a*), Lilik Hasanah (a), Irma Rahma Suwarma (a)
a) Department of Physics Education, Faculty of Mathematics and Natural Sciences Education, Indonesia University of Education
Jalan Setiabudi No. 229, Isola, Bandung, Indonesia
*purnamadian[at]upi.edu
Abstract
This study aimed to develop and validate a computational thinking (CT) assessment instrument for high school students within the renewable energy topic, using the Computational Thinking for Science (CT-S) Framework. The instrument, consisting of 13 multiple-choice items with reasoning options, measures five CT aspects consist of abstraction, decomposition, algorithmic thinking, evaluation, and generalization and integrates Arduino-based block programming in Tinkercad as a contextual science application. The development followed the Research and Development (R&D) model adapted from Plomp, including the stages of problem analysis, design, realization, implementation, and evaluation. Validation used expert review with empirical testing through Rasch model analysis. The results show that all items met model fit criteria (MNSQ = 0.77-1.46, ZSTD = -1.38 to 2.42, PTMEA-CORR = 0.29-0.62), and the instrument demonstrated unidimensionality (raw variance explained = 31.5%, eigenvalue = 1.60). Reliability indices were acceptable, with Cronbach^s Alpha = 0.7, item reliability = 0.93, and person reliability = 0.68, supported by separation indices (item = 3.59- person = 1.39). Item difficulty ranged from -0.53 to +0.65 logits, with 20% easy, 60% moderate, and 20% difficult items, indicating balanced coverage. The test information curve formed a bell-shaped distribution, showing the highest measurement precision around 0 logit, which represents students of moderate ability. Overall, the developed instrument is valid, reliable, and effective for assessing students^ computational thinking abilities in renewable energy-based physics learning.
