(This is my fourteenth post in a series of entries I’m using to participate in Blaugust for 2026.)
Ars Technica recently reported on Mexico’s largest university doing a completely remote entrance exam, and the disaster that resulted from it.
I won’t be using this to rant about AI, though the anti-cheat cameras were “AI powered.” I won’t be using this to rant about anti-cheat lockdown procedures either, though I’m also not a fan of those.
Neither of those would be needed if we didn’t still rely on high-stakes testing as a measure for student aptitude and accomplishment.
Now I’m not opposed to the occasional quiz. More often is better, because one off day can be balanced out by the rest. If done well, quizzes can even be used as tools for reinforcing the information students need to know.
But the big chunks of a student’s assessment should be evaluated by looking at what they can do, and I’m not going to get that by looking at how often they picked the right answer out of four possible options.
Show me a movie they made, an image they painted, a book they wrote, a business they planned, a robot they built, or anything other than a series of A, B, C, and D answers.
There’s a reason why some colleges have stopped caring about SAT scores, and it might be related to the reason why there’s no peer reviewed evidence showing a correlation between high test scores and success after school. (I vaguely remember Pearson publishing a study on their website that showed it, but their bias alone would probably prevent it from getting through peer review.)
If there’s one thing the SAT is known to show, it’s how wealthy the students’ parents are.
But we still use them, because, and say it with me, now: “We’ve always done it this way.“
And of course we have.
Multiple choice questions are easy to grade. Easier, now that we can have computers do it for us, but I’m old enough to remember Scantrons and hand grading.
We picked the easy over the good, because we have so many other things we’re told to get done and, quite frankly, grading tests is tedious.
(I will not segue into complaining about AI here, but I want you to know there was effort involved in that decision.)
There’s so much more value and reward in seeing students make something. It can still be graded objectively, you just need a rubric.
What amazes me is how many teachers don’t even know how easy rubric grading can be. They’re taught the test and then told to teach to it. I don’t have an article to link for that one, but I recall my days as a traveling art teacher where the principal specifically asked me to show sample rubrics to the teachers in her building because the teachers didn’t know how to use them.
The Better Way: Rubrics
When I’m making a rubric, I ask a simple question: “What do I want the student to accomplish with this specific project?”
We’re not talking a course-long activity here, we’re frequently looking at a single day’s worth of work (maximum 2 weeks, for my older students).
This is where my ADHD comes into play. I’m going to have a short list, because if I start getting distracted while writing it there’s no way I should expect a student to focus through the whole list.
The result is going to be 3-5 things I want their work to include. My youngest students might only have 3 “look-fors,” and even the older students are usually going to have 4. Point values may or may not be distributed evenly, depending on what I want valued.
My older students, for example, will usually have a list of 4 requirements, each worth 25% of their grade. In the beginning of the year my youngest students might instead have a ranked list of scenarios, where managing to log into the computer without help gets them most of the way to an A.
“When the ink was still wet on my certification I sometimes made a list of 10 things I needed to see. Granted, I was teaching high school at the time, but that was still a lot to get done. I like my current method much more.
“But Aaron,” you might ask, “How do I make sure they know everything the test checks in Chapter 5 of the math textbook?”
Great question.
How do you know they’ll remember any of that the day after the test?
My students don’t forget how to make movies, because it’s interesting and compelling enough that the knowledge sticks with them. If I have them make a movie covering a math concept they might not memorize a whole chapter, but they’ll be more likely to have that part down.
Now have the rest of the class make videos about the other sections in that chapter.
Have everyone in the class critique each other’s videos. Give them the rubric, and ask for positive and constructive feedback.
I do this in my classroom. The Critique Days count as an assessment score. My classes aren’t small enough for everyone to watch everything in one day, so I limit them to 10 critiques, with any extras as extra credit.
(I sort them alphabetically by student name and have every student start with their own video, so ideally no two students are grading the exact same list of videos.)
If you want to be really daring, you can average the scores the students give together and have that be a portion of the grade. Reserve the right to override the result, though. I’ve found my students collectively tend to be on the mark for a lot of things but every now and then a misunderstanding results in what would result in an unfairly low grade without me correcting things.
(I’m not the only voice in my classroom, but I reserve the right to be the loudest one.)
So, yeah, ditch high stakes tests. They’re bad. The people who make us have to use them should feel bad for doing so. Assign a project. Use a rubric. Everyone’s going to have a lot more fun learning that way.