Comparison of Computer-based and Paper-based Exams: Evidence from PISA

Abstract

We investigate the effects of the exam mode change - from paper-based to computer-based- on PISA on students’ exam performance. This change took place in 2015 in 57 countries, while 15 countries continued to take the paper-based PISA. Using this change and the difference-in-differences estimation method, we find that the computer-based format reduced Turkish students’ math, science and reading scores by 28.85, 29.52, and 39.975 points which correspond to a 5.9 percent decrease in math and science, and 8.1 percent decrease in reading scores compared to the average scores in OECD countries in the corresponding fields, respectively. We also investigate the heterogeneity of our results by gender and computer possession and show that there is no differential effect except a significantly negative impact for males in reading tests.