Background: This study examined the relations between students' expectancies for success and a physiological component of test anxiety, salivary cortisol, during an authentic testing setting. Aims: The aim of the study was to better understand the connection between shifts in students' control appraisals and changes in the physiological component of test anxiety. Sample: The study comprised 45 undergraduate engineering majors in the United States. Methods: Survey data concerning students' expectancy for success and saliva samples were taken before, during and after the practice midterm examination prior to their actual in-class examination. Results: Students' expectancy for success declined during the examination while cortisol levels declined from the beginning to middle of the examination and began to increase again as a function of time. Although students' initial levels of expectancy for success and cortisol were not correlated, there was a negative relation between change in cortisol and change in expectancy for success. Conclusions: Our study demonstrates a relation between salivary cortisol, a physiological component of test anxiety and students' expectancy for success in an authentic testing context. Most students saw a decrease in cortisol during the examination, suggesting anticipatory anxiety prior to the test and a return to homeostasis as the examination progressed. Some students, however, did not see a declination in cortisol, suggesting they may not have recovered from pre-examination anxiety. The negative relation between change in cortisol and expectancy for success suggests that students who had the greatest decrease in expectancy for success saw the smallest recovery in cortisol.