How new genetic variants originate in a population is usually challenging for students to understand because of the common misconception that mutations are harmful. Therefore, students often think that if a new trait originates by a mutation it must be because the organisms “needed” that trait. To reveal student misconceptions, constructed response (CR) assessments, in which students write short answers using their own words, are better than multiple-choice questions. Nonetheless, it is not always realistic to administer this kind of assessments in a large classroom setting. The Automated Analysis of Constructed Response research group (AACR) develops computer-automated tools that analyze students’ writing by creating scoring models that predict human scoring via a combination of computer trained scoring and statistical analyses. To better understand student thinking about the origin of a new trait in a population, we developed a two-part CR question about how an isolated population of buffaloes began to see a rapid increase in a new hair color: 1. How would a biologist explain the new hair color? and 2. How would a biologist explain the rapid increase in the new hair color? These questions were administered to introductory undergraduate biology students (n = 401) at a large research University. After running exploratory analysis of those responses, we developed a 5-category analytic rubric based on those responses, hand-scored all the responses, and created a computer scoring model trained with the human scored data set (Cohen’s kappa ≥ 0.7 for computer-human scoring agreement). Analysis of responses shows a majority of students correctly answered part one’s question while other students had alternative ideas for how genetic variation originates in the form of a mutation. Instructors can use this information to better help their students to master this concept.