Gender differences in Norwegian PIRLS 2016 and ePIRLS 2016 results at test mode, text and item format level

Gender differences in reading are a common finding in international assessments with girls usually outperforming boys. This article investigates such gender differences by looking at test modes(paper-based versus digital assessments), reading purpose (literary versus informational), text features (associations between reading scores and how much students like a text) and item formatcharacteristics (multiple choice versus constructed response items). All analyses are based on dataof Norwegian fifth-grade students (n = 3610) from the most recent cycle of the Progress in International Reading Literacy Survey (PIRLS and ePIRLS) 2016. The results point towards a generalmode effect between the paper-based and digital assessment for constructed response items. Thiseffect seems to be less strong in boys, indicating that boys may be motivated to type responses on akeyboard as opposed to writing with a pen on paper. For text features, we found that boys might bedisengaged from reading when the text shows female characteristics such as a female protagonist,leading to boys’ lack of interest and, subsequently, to lower scores. The results are discussed in thelight of the test design of PIRLS and ePIRLS.