Substack update

I'm excited because, this week, I'm going to try inserting code into a Substack post to see whether I can truly make interactive quizzes. I have a mini-quiz, with its code, already here on my blog, so I'll first run the code through the CodePen website (a site where you can test code without harming anything else, sort of the like Mirror Dimension in "Doctor Strange" or the holodeck—with safety protocols—in Star Trek) and, if it works there, I'll next try pasting that into a Substack draft post. If the draft post also functions correctly, I'll unleash my mini-quiz upon the world of Substack, and that will open the door to all future quizzes. It's that first step that's the doozy. But I'm taking it this week. I'll likely post the mini-quiz on a day when I don't usually post anything. Normally, I schedule my posts to appear on Mondays and Fridays, so the mini-quiz will probably come out—for both free and paying members since it's an experimental feeler—tomorrow, Wednesday, or Thursday. Don't get excited: this is only an experiment, and the mini-quiz is the same mini-quiz that you know and love: it's about religion and has nothing to do with grammar. That mini-quiz also has a built-in randomizer to change the order of the answers so you can't simply memorize an answer pattern (what the fuck drives people to cheat, anyway? don't answer: I know).

My quizzes will generally be multiple-choice in nature. Multiple choice is not my preferred format. As I've argued in my book and elsewhere, the fundamental problem with that format—despite the convenience it offers in terms of ease of grading (which is nice for lazy teachers)—is that you can theoretically guess your way through an entire quiz or test, so how is this a measure of competence? Better to have fill-in-the-blank and short-answer formats. But for me, since I'm not interacting with any "students" directly, I have less choice in the matter unless I ask "students" to write essay answers that I must personally check. I can see that working while student numbers are low, but should my number of subscribers ever rise, essays could quickly become a problem, and I absolutely refuse to let AI "grade" essays. Then again, students of a mind to cheat will just use AI to generate their essay answers, so that's no good, either. Ultimately, I guess it's the honor system. If you cheat, you only cheat yourself. Lucky for me, I'm not "accredited" by any official institution, so there's a tacit agreement between me and my "students" that I qualify as an authority.

Regarding the upcoming quizzes and tests: The exams I'll be doing for Tasty Grammar will randomize the order of the questions plus the order of the answers to each question, and each question you see will be randomly selected from a pool of three questions, so there will be randomness upon randomness. This way, a person can hit "reset" five times and take five different versions of the same quiz. In theory, thousands of combinations of questions and answers ought to be possible, so there's almost zero chance of an answer pattern ever reappearing. Then again, if I'm leaning on the honor system, one could argue that I don't need to go through all of this nonsense to keep people from cheating. Hell, let them cheat! But here's the thing: some honest people will take the quizzes and tests, and they'll be unhappy with their low score. So they'll re-study the material, then retake the exam. By rights, the retaken exam should look different, otherwise they'll be able to plug in the same answers as before without having to think. So making this kind of extremely randomized testing environment is less about preventing cheating and more about providing motivated re-testers with plenty of opportunities to test and test again until they get a 95 or above. One more wrinkle: my multiple-choice questions will have "checkboxed" answers, not radio buttons (radio button = select only one). It'll work something like this:

Mark all that apply. This could mean marking one, some, all, or none.

1. Which conditional sentence(s) is/are incorrect?

◻︎a. If I could have assassinated him with my snot, I would've.

◻︎b. Were she the queen, you know there'll be nonstop parties.

◻︎c. She thought, If Hans were ten years older, I'd bang his brains out.

2. Find the incorrect sentence(s).

◻︎a. You can't wave that swollen thing around in here, it's forbidden.

◻︎b. In our culture; we call this the 'sneaky kitten.'

◻︎c. Rolo the veteran pilot, leaped into his fighter jet.

Just for shits and giggles—did you get the answers? To get a point for #1, you have to check both A and B. To get a point for #2, you have to check A, B, and C because all are incorrect. By using check boxes, multiple choice gets a lot harder, and you actually have to think.

A 95% is basically a 19 out of 20 on a 20-question quiz, or a 57 out of 60 on a 60-question test. If each question in a 20-question quiz is being randomly pulled from a pool of three questions, then I need to write 60 questions for every 20-question quiz, and 180 questions for every 60-question test. That's also going to mean writing up a ton of answer explanations for these questions, so I'm probably going to have to figure out a way to automate the explanations. Maybe I can make the explanations appear at the end of the quiz when the student hits "enter" to receive a score. "Here's why your answer was wrong, and here's the correct answer." Or maybe a better strategy, since the answer explanations are contained in the lessons themselves, is to offer just the final scores, highlight the wrong answers, and let the students go back to those lessons to figure out what went wrong, then retake the quiz/test. While I'm not the biggest advocate of learning through failure (failure is a poor motivator), some determined students thrive in an environment where they get things wrong but show improvement. The improvement itself can be a motivator. Feel free to leave a comment below debating the psychology. And be sure to suggest alternate, constructive solutions.

I also have to remember to incorporate humor into the process. If all of this starts to feel like a boring slog, I'm going to lose subscribers.

She stared in horror at the shit spattered all over the walls by her overcaffeinated, bloodshot-eyed demon baby.

a. In this context, shit is a countable noun.

b. In this context, shit is an uncountable noun. 

So that's where things are. I have a lot to think about and plenty to do. I'm currently cranking out enough Bad Online English posts (the ones for the free-subscription Substack) to last me through the end of November. I will then turn to cranking out the regular grammar posts for the paying subscribers, and I'll spend my weekends and free time working on quizzes and tests. You better hope I don't die before I get these things out.