We hypothesised that college students take reading speed into consideration when evaluating their own reading skill, even if reading speed does not reliably predict actual reading skill. To test this hypothesis, we measured self-perception of reading skill, self-perception of reading speed, actual reading skill and actual reading speed to determine the relations that exist. The results supported our predictions. Primarily, self-perceived reading speed strongly correlated with self-perceived reading skill, whereas the correlation between actual reading speed and actual reading skill was low. Next, a multiple regression analysis showed that self-perception of reading speed significantly predicted self-perception of reading skill after controlling for actual reading skill. Lastly, how highly correlated reading skill was with self-perceived reading speed was found to affect how accurately one perceived his or her reading skill. Consequently, the study results suggest a negative impact of self-perceived reading speed on accuracy of self-perceived reading skill, as hypothesised.