Although much is known about how students perform on standardized tests, little research exists concerning how students think and process while taking such tests. This mixed methods action research study was designed to investigate if a constructivist approach to test preparation could yield improved results for 37 English language arts freshmen preparing for the California High School Exit Exam (CAHSEE). The researcher explored student performances on Princeton Review practice CAHSEE tests (versions A, B, and C) over a 3-month period and she performed several statistical tests including repeated measures ANOVA by individual classes (advanced, basic, below basic) to determine differences between and among groups. In addition, she examined students' metacognitive processes as they participated in bi-monthly audio-taped classroom discussions of their thinking on specific multiple-choice items. After each of these discussions, all students independently wrote to a common prompt in their journals to articulate their thinking. In addition, six focus students recorded their thinking aloud while they tested on specific items and participated in end of study interviews about multiple aspects of the intervention. Emerging themes included how students read and interpreted test questions, eliminated answers, revisited the text, and felt about the intervention. While repeated measures ANOVA indicated a significant testing effect (p = .05) for advanced students early in the study, it was the below basic test-takers who actually made the greatest gains overall. And although students made improvement interpreting test questions and eliminating answers, all groups struggled with knowing how to strategically revisit the text. Approximately one-third of the students from all classes attributed laziness or tiredness as a factor influencing their performance. The more advanced focus students tended to respond to the test items in a cyclical pattern, practicing strategies repeatedly whereas the less advanced test-takers participated in more of a lock step, linear approach. Most focus students declared that the most useful aspect of the intervention was thinking aloud as they tested or listening to others during classroom discussions. Recommendations for future studies include more in-depth, multidimensional studies to capture students' thought processes in order to shape classroom instruction that assists students in becoming better test-takers.