Wednesday, August 5, 2015

Assessment Twists

In my classroom, I use a variety of nontraditional mechanisms to assess my students. Some of the assessments have some interesting benefits, including virtually wiping out student cheating... by letting them cheat.

At least once in each unit of my AP Statistics class, I give students the opportunity to work together on some type of product. Sometimes these are just group activities or projects, but sometimes they're an actual group in-class test. Students get to work together on these test questions. And there are very good pedagogical reasons to do so.

I think there's an inherent value in students collaborating together in all aspects of classroom learning. When teachers build in opportunities for student interaction, there is opportunity for more learning and deeper thinking. Students rely less on me, the teacher, and learn to rely more on themselves. In real-life situations, problems are rarely solved by one person working alone. It seems almost silly to me now to think that assessment has to be done only for students working alone.

Group tests and/or assessments can help students who are struggling to keep up with the pack, while still providing a way for those students to contribute. They help students with traditional test anxiety by creating a less stressful environment. They help students clarify their ideas verbally before committing to writing. They help build the notion of community and teamwork in the classroom. They help build opportunity for interesting and amazing discussions between the students. Depending on the strategy, they can help lighten the load for teachers with a large number of students in a course. And, yes, they pretty much eliminate any incentive for students to cheat.

Here are three different ways I use this strategy:
  1. Test with a Brainstorm Session -- In this strategy, that I borrowed from a colleague, students are given test questions individually. In groups, students get time to read silently to themselves and think about the question, then they get some time to discuss the question(s) with each other. During the brainstorm session, they are not allowed to write or take notes. That time is solely to discuss, listen, and ask questions. Once that time is over, they separate and answer silently by themselves. This is a fantastic way to test more difficult or new tasks. Students find that they get over the hump of how to begin or what procedure to use. Bonus Option: Give students questions that are the same but differ in some small way (different values, using different dataset, etc.).
  2. Group Test -- Students are given one test for the entire group. Similarly to the first strategy, I give students time to read the question(s) silently or aloud in their group and think about it. Each student must contribute to the answer(s) and be actively engaged. An easy way to demonstrate this is to require each student's handwriting on the paper. [In the photo above, you can see my students working on a group test. In this case, I allowed students to use any resource in their notebooks. Each student took turns writing, calculating, and looking up information so that the conversation about the problem was nonstop by all members.] Bonus Option: Grade the test as the students are working and allow them to correct their mistakes before they have to submit.
  3. Pyramid Test -- Students are given a test individually as you might do traditionally. On a subsequent day, before they have been given feedback on the first test, they work on part or all of the questions given individually. They discuss and agree on a solution to submit as a group. Just as in the previous method, each group submits one copy with all students' handwriting on the paper. This can be useful to do after-the-fact if you find that many students struggled on a particular question on a test, for example. In terms of grading, you can decide what weight to give each version. Bonus Option: Have students discuss as a class and then submit one class solution. 
I think these methods would work well in any math or science classroom. You'll notice that in each of these methods, there is still accountability for each student's performance but with a little twist to allow for the group collaboration. While I think the last strategy can be time consuming, they all can be integrated into a normal testing routine nicely. A very nice outcome of these is that I observe students engaging and learning as they are working, so these tests become learning tools as well as assessment tools.

Just as a note, if you choose to do any kind of group test, it'll likely take longer than a traditional test. I find it takes groups at least twice as long to complete a task than if it were assigned individually due to the time spent discussing until they reach a consensus. In addition, you'll want to make sure students get back all group assessments. I make copies of all graded assessments so that each student has his/her own copy for reference.


  1. I have been thinking about this post, and I have been especially thinking about this statement that you made: "Group tests and/or assessments can help students who are struggling to keep up with the pack, while still providing a way for those students to contribute."

    Can you clarify more about what you mean? Do you mean that it helps struggling students get a better grade? Or a better understanding? Or both? And do you have more evidence to add to that? I could definitely see it helping with a grade, but I'm trying to see how it will help them with understanding without sacrificing or obscuring the conversation between them and me (which is how I view assessments in my classes).

    I'm not sure I could adopt any of these right now as part of an assessment scheme (with the way that I am grading now, it is very important to me that any work that counts for their grade is done individually even though a lot of other work is done together), but I'm interested in thinking about it more. My biggest concern is a student fooling himself (and fooling me) into thinking he understands something well when he actually is has a gap (filled through some group work) that could have been fixed if I'd forced him to work individually. (I use standards-based grading, so students take multiple assessments on the same objectives and can continue working and revising until they reach the level of mastery that they want—or until they run out of time in the year.)

    Okay, I hope you were looking for rambling, thinking-through-things comments here! :)


    1. Kelly, I don't use standard-based grading so you'll forgive me if I'm not addressing what you need. (I'm interested to see how this could be integrated into a SBG scenario.) I can clarify, however, how this might work. I'll use an easy example-- let's say students are asked to solve an equation that involves 3 steps. A struggling student might be proficient at two of those steps, but shaky on the third. Now imagine, that that student has his/her peers to give feedback on that third step. Or can ask questions of his/her peers about the work that they're doing. The advantage here is the immediate peer feedback they are receiving while they are working.

      I think that's a very powerful tool for that student's learning. They can still demonstrate what they know while having the support of peers who can help with what they might still need help. I found that many of my students were more comfortable with doing tasks on their own (on a subsequent test or activity) if they were initially able to do them as a group first. Most of the skills (and thinking) that students use in my classroom build on each other, so a student is continually revisiting past topics in new ways.

      Could you see this in your classroom as an activity without a grade, but to give feedback? That might be a good first step. I'd be interested in hearing feedback from teachers who use SBG to see how that might be implemented.

  2. I too teach AP Statistics. I am considering having students use the AP Scoring guidelines to score a classemates work. Have you ever tried this? I am considering having everyone do an AP question individually. Then, for homework, each student would have to score a different students work using the AP guidelines. Then, I would randomly select about 5 sets to grade. I would be grading both the grader and the question. The reason for only grading about 5 sets is to reduce the time that I spend grading. I am afraid that I won't do this too often if I am overwhelmed by the paper work. In the end, if the graders of the randomly selected papers graded well, I would award the entire class some points for grading properly (I am hoping that this random method would prevent kids from just giving their buddy a good grade). Then, each individual in the class would also get the grade that the student who graded their work awarded them. I know this sounds really confusing but I think it may be worth a try. I really want my classes to see how strict the AP standards are.

    1. I do this regularly but I have students do the peer assessment in class. We go over the rubric together. I add some advice here and there. They can help each other to assess and also me (& the whole class) questions. I usually have students rate papers from another period, to make it easier to discuss student work and responses. If you only have one class, this will be a little weird perhaps. You might start with taking one paper and grading it as a class. I do look at students' rating of their peers and give them a grade for their assessment. If a student doesn't assess their peer properly, I reserve the right to reduce their assessment grade. It's been a great way to get students familiar with the rubric, scoring, and also critically analyze another student's work.

    2. Thanks for your advice! You sound very experienced teaching AP Stat. I like the idea of swapping papers from other sections. That will work for me. I will definitely start by grading one as a class.