We present a validation study of fifteen new math anxiety scale items designed to augment the widely-used Revised Mathematics Anxiety Scale (RMARS). While the RMARS and other standard instruments measure students' anxiety in response to computation, test situations, and math course activities such as buying a textbook or watching a lecture, the new items present scenarios that reflect current pedagogical recommendations for college mathematics classrooms, specifically solving of non-routine problems and explaining thinking to peers. Anxiety in response to these kinds of scenarios we name Problem Solving Anxiety and Explanation Anxiety, respectively. With a survey sample of 132 future teachers enrolled at colleges and universities, we use exploratory factor analyses to demonstrate distinctness of these constructs from existing constructs and provide evidence for construct validity. Cronbach's alpha for the overall scale, as well as for each dimension individually, are all between 0.9 and .95, indicating internal consistency.