About half of the points from my BIO181 class come from high stakes exams. I feel this is necessary to prepare students for their STEM degrees, MCAT, PCAT etc. I split the course content into 5 units with an exam for each unit. This means giving up 5 class meetings to exams, which for a TR class, is over 2 weeks of class time. I tried dividing the content into 4 units, with 4 exams. The last 2 exams remained the same, but I took the content from the first 3 exams and split it between 2 exams instead. I examined student scores to see if there was any impact to how well students did on the first two, new, "broader" exams. The cohort who took 4 exams scored over 9 points lower on their first 2 exams than those who look 5 exams. However, when comparing scores on the unchanged last 2 exams, the "4 exam" cohort also scored 6.5 points lower, indicating a generally weaker cohort. When controlling for this, no significant difference was found between taking 4 or 5 exams. This serves as a cautionary tale. A 9 point gain/loss is not neccesarily all it seems! To be truly informative, SOTL research needs controls.
Very interesting, of course we have always wondered about this, whether a 4 exam course would result in significantly lower success levels. Yes controls are pivotal and help to reveal the real story. Hopefully the increased class time helps with content.
Rachel, This is an excellent CATS for instructors to read who have struggled with the same issue of time being used for exams. It would be interesting to see if other disciplines have the same results. Thanks, Olga
Rachel, this is a perfect example of the Action Research Cycle. Most importantly, YOU learned from this and now know what path to take. Well done!
I have wondered this and have gone back and forth as well with the number of exams, but I haven't collected any data on it. It is difficult giving that many exams, I especially have difficulty with the hybrid courses. I hope the testing center will be able to change that.