Here’s something that has bothered me for a long time: Software that speeds up report card generation through the use of templates and boilerplate phrases generated by simple forms filled out by teachers.
Schools keep telling parents to take more interest in what their kids are doing in school and to help them in areas where they need work. I agree with this, but getting actual information out of the school on my daughter’s work is like pulling teeth.
There are two primary channels of information about how your child is doing: Parent-teacher interviews, and report cards. Leaving aside the parent-teacher interviews for this discussion, I want to talk about report cards.
When I was a kid, teachers didn’t have computers. This forced them to write personal assessments for every student. So my report card for science might read something like this:
“Little Sammy understands fluid flow very well, but he’s still a little weak on the concept of viscosity. I’ve noticed that he’s often distracted in class when the subject isn’t very interesting, and spends a little too much time talking to his neighbors. Also, the bunsen burner is not supposed to be used as a roman candle, and filling balloons with propane and igniting them is frowned upon. Overall, Sammy is an excellent but somewhat dangerous student who would make a good scientist if you can keep him away from combustibles.”
Anyway, there was enough information for a parent to get a picture of what the kid has been doing in class, and if the mark was anything other than an A, the teacher would always write down exactly what was lacking, and why.
But now, it’s apparent that teachers are saving time by using report-card generating software. They plug in the mark, and the software generates the written summary.
Here’s an example from my daughter’s report card for a subject she got an A in:
That doesn’t sound too bad - fairly specific. But then you read the math mark, which was also an A:
You can see right away that both teachers used the same software. They filled in a field for what was covered in the course, and then for each student they jsut had to click a radio button to indicate the overall mark, and the software generated the text for all the kids. Then maybe they edit the text a bit to add a phrase like “Good work!”.
But hey, she got A’s, so maybe I don’t need to know anything specific. But she got a C in phys-ed. I’d like to help her improve. So I went to the report card to see where she had problems. Here’s the summary:
Exactly the same boilerplate. A description of the units covered, and a summary paragraph that was clearly generated by software because the teacher selected ‘C’ as a grade.
This is useless information. You might as well just tell me the grade letters and leave it at that. There’s nothing I can learn from this report card that will allow me to help her in school, other than the letter grades themselves.
But the use of this kind of software is even more pernicious. The use of templates and boilerplate software generators like this allows the teacher to avoid having to sit and think hard about what to say about a child, and sometimes that process of thinking can help the teacher with understanding the needs of the children.
This issue has come up in health care digitization. One of the things stopping digitization is research that has shown that doctors who chart patients with computers start to develop habits like cutting and pasting diagnoses and treatment plans between patients, and they slowly begin to mentally disengage from their patients. They start fitting everyone into cookie-cutter boxes driven by the design of the software rather than by their own training. It’s a slow, pernicious effect.
The other problem with boilerplate text is that it can actually be misleading because it provides generalized responses for specific problems. It’s misleading in that it suggests that it’s describing personal traits of the child when it isn’t. Maybe my daughter’s problem in Gym isn’t lack of effort, but a physical defect or an inner ear problem causing lack of coordination. But the software can’t make that distinction, so it gives me bad info. Perhaps if the teacher had to sit and think about what she had done wrong, the teacher might have an epiphane about her needs and become a better teacher for her. That won’t happen when the sum total of the teacher’s per-child effort is clicking a damned radio button.
If if you’re a teacher and you’re reading this, put down the mouse and pick up a pen. Write personal student evaluations. Engage your brain, and think about what information the parents need to know to be better parents, and what next year’s teachers might need to know about the kid to be more effective. If you’re using an algorithm to generate the text of your report cards, you might as well just send home a list of letter grades and leave it at that.