Testing Retesting

What I’ve learned after a semester of allowing students to retake their exams, in a class where I though such a policy was impractical


About me and my school

When I started teaching AP Computer Science A in 2015, it was easy to approach a subject that was new to me in a way unlike how I had been teaching mathematics. Initially, I allowed students to retake their exams simply because I knew I lacked the experience to write fair exams on my own first try. I kept the policy because it seemed almost natural, the course itself seemed conducive to a policy that allowed students to resubmit code until it worked. At that time, I could not see how the same policy could carry over to my math classroom and a discipline that is notorious as a harsh exam environment. Finally, in 2017, having grown frustrated with feeling that I was perpetuating a culture that rewarded exam performance over learning, (mind the distinction: it’s one thing to take an exam for the sake of one’s learning, I take issue when students perceive they are learning for the sake of their exams) I decided I could try what I was doing in C.S. 


The Policy

All of the following practices were new this year:

  • Multiple modes of assessment: Paired Assessment, Group Assessment and Free Response Individual Assessment
  • Students may not retake Paired and Group assessments, which have review and revision built in. However, these exams are always open-note.
  • Students may retake Free Response Individual Assessments in a testing environment for 100% credit until they receive a satisfactory score. 
  • There is one version of each Free Response Individual Assessment, therefore, when a student retakes such an assessment, they take the same one.



This is how I explain the policy to my students at the beginning of the year: 

“In this classroom, learning is our priority. In ten or twenty years, when you look back at your math class and try to remember the most important thing you learned, it will not be a formula. It will not be a procedure. Hopefully, it will be the ability to solve problems, utilize resources, prove an argument and review the work of others. Traditional exams exercise only the first of these skills adequately, at best. Taking an exam in isolation, without any resources doesn’t just fail to exercise the other three skills, it misrepresents reality. Reality is not a testing room. For this reason, we will have three different types of assessment. Furthermore, I will allow you to retake your individual exams for full credit. The reason I am willing to do this is because I believe it supports your learning. There may be some of you who struggle with performing well on exams, to you I say: ‘practice makes perfect.’ I am convinced this policy will motivate you to go over your exams, figure out what you did wrong, and accustom you to an exam environment that is not as stressful as what you have known before today."



There are both good and bad things to report after one semester, which puts me in a conundrum.


The good

The culture of my classroom is highly preferable to what it was before. Students are far more collaborative. Students are far more comfortable sharing their work. Students are far more apt to approach me outside of class with a question or to seek extra help. To summarize in a single statement, it is a culture that says “If you make a mistake, it can be fixed.


The bad

Frankly, my students are doing worse on department-wide exams. Department exams are multiple choice, uniform exams that cannot be retaken, as per department policy. I’m going to make a broad statement and estimate that student exam scores are up to ten points lower than they were last year. This week, my students took the department-wide, mid-semester, 20-question exam and my highest score (out of 29 students) was a 90. The median was 70. (Granted, I have covered all of the material in the department curriculum map up until this point, whereas a unit’s worth of content was trimmed off the exam because most teachers had not yet covered it.) The same is not true in my C.S. course, where my classes have performed ahead of the department curve on the AP Exam.


The conundrum 

The conundrum I am facing is not a matter of what I could do to improve. I know what other teachers are doing that I am not: they are spending more time preparing students for exams and not letting them retake their class exams. The conundrum I am facing is not a matter of inner conflict. Personally, I know I would not trade my classroom culture for higher exam scores. The conundrum I am facing is this: supposing a student or parent expresses concern over how my class is preparing students for the Regents (end-of-year) exam, how would I justify my practice? I could ask them to weigh the benefits of what I am doing well against what I am giving up, I could even remind them that despite some lower exam scores, his or her child’s grade is relatively high. But this would not be satisfactory. Ultimately, I am afraid I am up against a fierce and unforgiving reality: parents and students perceive success as it is dictated by an exam score.



I mentioned students in my C.S. course are performing much better. That is a bit misleading because there are no department-wide assessments. Without a doubt, if I were to administer a standardized assessment to my C.S. students tomorrow, the scores would be disappointing. At this time in the year, my students are conditioned to work collaboratively on software projects, not answer test questions. In mid-March, we will begin a six-week intensive preparing students for the AP exam. During that time, students will take three entire exams, every free response question published in the last decade and will even write their own exam questions. They will breathe the AP exam like oxygen and then they will crush it when they take it in May. Eventually, my CS student will perform much better. If I were to test them today, they would probably perform at least as poorly as my geometry students.

The plan is to conduct my geometry class in the same fashion as my C.S. class. It may sound hypocritical that I would ultimately put so much emphasis on the exam at the end of the year, but let’s face it: while the exam is just a hoop students jump through, it’s an important hoop and it comes with its own subset of skills: an understanding of how distractors can be written to to deceive you, how the free response rubric tallies points in your favor, how to pace yourself, and how to handle the areas in the exam that may trip one up. It isn’t a pleasure to teach these skills, it is just part of the job.

After one semester, I’ve learned that the environment I am in is not conducive to a retake policy. I’ve learned that my students, while receptive of any policy that will enable them to get a higher grade, are still conditioned for the exam-centered system they’ve been inserted into. I’m frustrated that I can’t report stellar exam scores as a result of what I think is a better policy, yet I am convinced that there is a greater good to be achieved. I am hopeful Regents scores will turn out in the end, even if I do have to spend a few weeks centered on exam-taking skills. It is much better to spend the majority of the year teaching actual math.