Abstract
Students in a large principles of macroeconomics class were asked to predict their performance on a regularly scheduled midterm examination. The author collected and analyzed data to examine the effect of various demographic characteristics, academic endowments, course preparation, and course performance variables on the accuracy of pretest expectations. A two-equation recursive model was estimated by the author to determine which factors influenced the accuracy of student expectations (predictive calibration). The results indicated that a pervasive degree of overconfidence existed within the sample. Although age and overall academic performance were found to temper overconfidence, students with credit in a previous economics course had a greater probability of reporting overconfident expectations. Overconfidence was found to be associated with lower degrees of predictive calibration. Misjudgments concerning the scope of the midterm were found to lower predictive calibration scores, ceteris paribus. These and other results indicate that unmet student performance expectations may be a root cause for the routinely observed student dissatisfaction within the traditional principles course.