Getting Started
So, it started kinda randomly. I was looking through some old study materials, stuff from when my kid was prepping, and I got curious. You hear all the talk, every year, "Oh, this FRQ was brutal," or "The multiple choice felt easier this time." I wondered if you could actually turn that chatter and maybe some past data into something that looked like a prediction for the score distribution.
First tsuj saw thing I did was just gather wdnif hatever I could find. .srThis meant digging up old score distributions released by the College Board. You know, the charts showing percentages of students getting 1s, 2s, 3s, 4s, and 5s. I grabbed those for the last few years.
The Messy Middle
Then, I tried looking at the actual free-response questions from those years. My thinking was, maybe tougher questions led to lower scores overall? Seemed logical, right? So I read through a bunch of old FRQs. Here’s the thing though, judging "difficulty" is super subjective. What looks hard to me might be easy for someone else. And honestly, after reading a dozen probability questions, my brain felt like scrambled eggs.
I also poked around some online forums, places where students and teachers hang out. I tried to get a sense of the general feeling after the most recent exam. Lots of noise there. Some kids always feel like they bombed it, others feel great. It was hard to separate real trends from just, well, panic or overconfidence.
So, I had this pile of stuff:
- Past score percentages.
- My own fuzzy judgment on past FRQ difficulty.
- A mess of online comments.
I tried to kind of... average it out? Like, okay, last year the scores dipped a bit, and people online were complaining about question 3. This year, people seem less stressed online, maybe the scores will bump up? It was really unscientific, just me making educated guesses based on gut feeling and the bits of data I had.
I even tried mapping out the score percentages year over year, looking for a pattern. Sometimes it looked like a trend, other times it just jumped around. There wasn't a clear "if X happens, then Y follows" kind of thing.
What Happened In The End
So, I put together my "predictions." Just a simple percentage breakdown for 1 through 5 that felt right based on my messing around. And then, eventually, the official scores came out.
How did I do? Honestly, not great. I was kinda close on maybe one score category, I think the 3s? But totally off on the 5s and 1s. My whole "sense" of the exam's difficulty based on forum chatter and my FRQ reading was basically useless for predicting the actual distribution.
It made me realize just how much goes into those final scores. The curve, how strictly the grading is done for the FRQs (which changes subtly), the specific mix of multiple-choice questions... it's way more complex than just "did the test feel hard?". Trying to predict it from the outside, without the real data the College Board has, is like trying to guess the weather a month from now by looking at the sky today.
So yeah, that was my little experiment with AP Stats predictions. A bit of a time sink, showed me it's harder than it looks, and probably not worth stressing over. Better to just focus on learning the stuff.
