In the current study, we utilize natural language processing techniques to examine relations between the linguistic properties of students' self-explanations and their reading comprehension skills. Linguistic features of students' aggregated self-explanations were analyzed using the Linguistic Inquiry and Word Count (LIWC) software. Results indicated that linguistic properties of self-explanations were predictive of reading comprehension ability. The results suggest that natural language processing techniques can serve as stealth assessments of abilities within intelligent tutoring systems. [This paper was published in: "Proceedings of the Seventh International Conference on Educational Data Mining (7th, 2014)" pp. 401-402 (see ED558339).]