Introduction: Many teachers consider real-life tasks problematic for students because of their linguistic demands. However, it is unclear how the linguistic demands of real-life tasks actually affect students' solution rates for these tasks. Method: Against this background, this experimental study systematically varied the linguistic demands of real-life mathematical tasks. It examined (1) the extent to which teachers (N = 72) estimate the influence of linguistic modifications on students' solution rates, (2) the extent to which linguistic modification influences actual student test results (N = 1, 346), and (3) whether students' language skills mitigate this effect. Results: The results showed that the teachers expected linguistic modification to strongly influence the students' problem-solving processes (effect sizes: 0.73 < d < 1.67). However, the effect size for the actual student performance (d = 0.12) was considerably lower than the teachers' expectations. Discussion: The findings indicate that in mathematics teachers' education, additional attention should be paid to the role of language in solving reality-based tasks.