Abstract Assessing student writing skills is a demanding task for teachers. Low-scope errors such as spelling mistakes have been shown to significantly distort assessments of written work. To date, no cognitive theory based explanation for why this happens has been empirically proven. The current study used an experimental design to test whether low-scope spelling errors affect the attention paid to and assessment of high-scope errors (e.g., errors of coherence and vocabulary). The reading behavior of 58 Swiss preservice teachers was recorded by measuring their fixation time on high- and low-scope errors while they assessed eight texts. All the preservice teachers assessed the same texts, which contained systematically manipulated errors. Low-scope errors attracted more attention and influenced text assessments. High-scope errors had no effect on either fixation or assessment. This indicates preservice teachers seems to be so focused on low-scope errors that they barely notice high-scope errors. The implications of these results for training preservice teachers to properly assess student texts are discussed.