I've talked about the plan and how things went wrong. It is probably only fair to myself and my students to do the opposite as well.
Another thing I decided to do was only unannounced assessments because if people only do well because they cram for something, it doesn't give me useful information about their proficiency. Also, telling them in advance tends to get them stressed out. This was a good idea.
I also called all my quizzes "comprehension checks." Not saying the Q-word turned out to be a great idea. Even though they were effectively quizzes, even halfway through the year I still had students bragging to the Spanish students that we never had quizzes. They still stressed out once they got the papers, but...
since I allowed unlimited retakes it wasn't really a big deal for most of them. Next year I will definitely do retakes again, but I'd like to make it more extra-help oriented (i.e. you have to come work with me or do this extra assignment or something before you're allowed to retake it.). I also need to make multiple versions so they're not retaking the same thing over and over.
Initially my comprehension checks were all in Latin including the questions. I changed that mainly because of what Martina Bex describes in #1 here:
When assessing reading comprehension of target language texts, we almost always ask the questions in English so that an incorrect answer can only be attributed to a misunderstanding of the text, as opposed to a misunderstanding of the question or an ability to ‘hunt and peck’ to find the answer. There are a few exceptions, and they are noted throughout this packet.Hunting and pecking was DEFINITELY happening before I started doing this. English comprehension questions made it much clearer to me what my students actually were able to understand, and I think many of them liked the change because they didn't have to stress out about the question's meaning. The difficulties here are that it's hard not to "give away" answers by using English questions- e.g. if the target structure was vult and I ask "what did Mary want?" a clever student should be able to figure out vult = want. On the whole I'm okay with that, though, because that in itself is a kind of reading comprehension skill.
Another great idea was no error correction. I based this off some things I read in Conventiculum Bostoniense's excellent pedagogy course, especially this article by Corder. If you've ever graded papers and watched students throw them away after looking only at the grade, this work comes as a breath of fresh air... and it's from 1967. If you don't feel like reading, try this Black Box video from Musicuentos & Indwelling Language on written corrections specifically. Justin has promised there's one coming on Corder's article which talks about oral corrections too I believe. Basically, and you already know this in your heart, most students don't get much out of corrections except a sense of failure. This is so, so so so so so contrary to how we teach Latin traditionally. Mark every tense error! Every voice error! No long marks?! F! F MINUS!
That sucks. It sucks to be that teacher because it's so discouraging. Even the best students make mistakes. And it sucks to be that student even more. "If my paper is going to be covered in red pen regardless, why even try?" So, cut it out, praise instead of blame, and "correct" errors by selective recasting and mostly by providing more correct input and more, and more, and more. Learning language is a slow, weird process. A three year old's English-- no matter how precocious the child-- is charmingly inept. Our students have had MUCH LESS input than a three year old. Why hold them to a higher standard? Additionally, students' errors tell us a lot about what they do know- so dwell on the positive! Here's some more links about how errors can tell us what students know, and how they're NOT random: one, two, three. I could go on, so let's leave this here.
Whew. I feel like that's enough on assessment. Oh, nope, one more.
I didn't do this enough, but it's a five finger system. 1 finger = "That might as well have been Chinese," 5 fingers = "I could teach that right now." 3 fingers is something like "I got what was going on more or less." After a reading, or after I've been speaking with the students, I ask them to rate it by "giving me fingers" (this was not an intentional phrasing and no one noticed that it sounded like I was asking them to gesture inappropriately). I like this rather than thumbs up thumbs down thumb in the middle because it has more of a range. A "1" and a "2" are pretty different in terms of self-perception. A "1" sometimes means "I've given up," whereas a "2" means "I'm still engaged but I really want some help." On a thumbs up/down/side scale, those would both be thumbs down. I think I'll use up/down for agree/disagree next year, but stick with five fingers for comprehension self-evaluation. hat tip to Justin Schwamm, from whom I got this system.
Okay, so this is a lot to chew on for one entry, so I think I will leave it here and do another one on other kinds of success I had. Assessment is one of the weirdest scariest parts of getting into CI-based Latin teaching, so I hope this is helpful.
If you're curious, you can dig around in this folder for comprehension checks. It's a big pile of everything I've been doing for Latin this year so it's not user friendly, but if you search for "CC" you will probably find a bunch.