Nine of 48 items on Indonesia's new digital test show gender bias.

Tests should measure ability fairly, without favoring one group over another. The PISA 2025 exam added a new area called Learning in the Digital World. It tests digital problem-solving and self-regulated learning using computer simulations. Because this test is new, researchers wanted to check if its questions treat all students equally. They analyzed responses from Indonesian students to see if boys, girls, or students from different types of schools had an unfair advantage. When two students have the same overall skill, they should have the same chance of answering a question correctly. If they do not, the question is biased. The researchers found that some questions were biased. Out of 48 items, nine showed bias by gender, with eight favoring boys. Six items showed bias by school type, consistently putting students in private, government-dependent schools at a disadvantage. Unequal past experience and school resources might explain these unfair gaps.

Why it matters

Comparing test scores across groups only makes sense if the questions measure the same thing for everyone. In Indonesia, the new PISA digital test is used to judge skills like self-regulated learning and computational thinking, so biased questions could make average scores unfair. The study flags fifteen such questions and urges corrections before using the scores.

Questions answered