The current study investigates the degree to which the lexical properties of students' essays can inform stealth assessments of their vocabulary knowledge. In particular, we used indices calculated with the natural language processing tool, TAALES, to predict students' performance on a measure of vocabulary knowledge. To this end, two corpora were collected which contained essays from early college and high school students, respectively. The lexical properties of these essays were then calculated using TAALES. The results of this study indicated that two of the linguistic indices were able to account for 44% of the variance in the college students' vocabulary knowledge scores. Additionally, the significant indices from this first corpus analysis were able to account for a significant portion of the variance in the high school students' vocabulary scores. Overall, these results suggest that natural language processing techniques can inform stealth assessments and help to improve student models within computer-based learning environments. [For complete proceedings, see ED560503.]