The AP Test 18

I told my students that I would write a blog post about the AP exam. We didn't have any school days together after the exam. So here goes. This blog post is written both to them and to any teacher that might read this, so it's a bit of mess. But you'll figure it out.

Overall I thought the 2018 AP stats exam a bit heavy on probability and mathematics this year, and a little too light on the topics that students really learn in an intro stats course.

Problem one (regression), number four (2-mean t-test), and number five (more below) were the type of questions that I think students who really work hard in AP Stats course should know how to do well. I especially liked number five, which I graded for four days. 

Number five had three parts. The first asked students take two medians and decide which of them belonged to two different histograms. I was really impressed with the different explanations that students used. Some counted into the median. Some made a really nice argument that compared the skewness, the means, and the medians. That was impressive. Part two asked students to calculate a weighted mean. That's an important idea. Finally part three asked students to use the histogram to find a probability. Unfortunately, many students used normality instead. In fact, most students did this (there's a teaching point here, probably worth a separate post). I'm really curious to see how my students answered this question.

Numbers 2 & 3 were a combination of probability, bias, and math/algebra. If you are strong at math, you probably figured out a lot of these problems. If not, they were tougher. We did not spend a lot of time on tree diagrams this year (that's probably worthy of another post), which was the best strategy for number 3. Hopefully that didn't put too big of a dent in my scores this year.

Number 6 ended up being too difficult for most students. The mean score was about 0.33 (out of 4). Thus if you did well on this question, it will definitely help your score. But if you didn't (I'd estimate that at least 75% of students scored zero) then it won't hurt your score.

Overall, this was not my favorite exam. However, I do think my students still scored well. Many of the items on the test we covered in class very thoroughly. And I'm also betting that many of students figured out more of the challenges than they suspected. We'll see on July 5!