Journal of Educational Psychology, Vol 118(5), Jul 2026, 651-676; doi:10.1037/edu0001013
This study investigated whether reading medium (print vs. digital) influences early adolescents’ reading experience and performance, and whether these effects are moderated by individual differences in verbal and visuospatial working memory (WM). It also explored the potential mechanisms underlying any observed medium-related differences, drawing on three prominent hypotheses: the shallowing hypothesis, the metacognitive deficit hypothesis, and the haptic cues hypothesis. In a within-participants design, 141 seventh and eighth graders read two expository science texts, one in print and one on screen. Measures included processing time, perceived cognitive load, calibration bias, and text comprehension, along with a standardized assessment of verbal and visuospatial WM. Linear mixed models revealed that digital reading led to shorter processing times, greater overconfidence, and lower comprehension. We also found preliminary evidence that the influence of verbal WM on processing time and cognitive load varied by medium, suggesting a greater reliance on verbal WM in digital contexts. However, perceived cognitive load did not differ overall between media, and visuospatial WM played a minimal role, offering little support for the haptic cues hypothesis. Serial mediation analysis revealed that digital reading impaired comprehension through a sequential pathway: reduced processing time predicted greater calibration bias, which in turn lowered comprehension. Notably, this indirect effect was consistent across WM levels, suggesting that although WM influences how students allocate time, it does not substantially alter the downstream effects on comprehension. These findings support an integrative account of the screen inferiority effect and underscore the cascading interplay between cognitive and metacognitive processes. (PsycInfo Database Record (c) 2026 APA, all rights reserved)