We all know that statistical inference is the end goal of AP Statistics. Inference also makes up the largest part of the AP Exam (up to 40%). While our students often do well within each chapter of inference, they often struggle on the AP Exam when they have to read a question and then decide on the correct inference procedure (and the appropriate conditions and formulas).

**The Big Idea**

Each student will individually be writing an exam that will contain two questions for confidence intervals and two questions for significance tests. The students create a unique context for each question and must research (or make up) the data to be used in the question. The students create a rubric for each question that will be used to grade the exam. On exam day, each exam is randomly assigned to another student. Once completed, the exam is returned to the student who created it so they can use their rubrics to grade it.

**Why does this work?**

__Students choose the context.__

Encourage students to create questions with a context that they care about: acceptance rates at their favorite college, Snapchat streaks, or their favorite NBA team. Engagement in this activity is very high as students are using their own interests and passions to create questions.

__Students have to create and use a rubric.__

The process of creating a rubric reminds students about all of the parts needed to get full credit on the AP Exam. Because students have to create rubrics, they are more likely to remember all the details needed when they have to take another student’s exam.

__Creating a question versus answering a question__

When students are forced to write the questions, they have to think about all the information that is needed for conditions to be met and for the question to be answerable. This increases students’ ability to choose the right inference procedure when reading new questions on the AP Exam.

**Teaching Tips**

__How long does this take?__

We suggest 2-3 class periods for students to create questions and rubrics, 1 class period to take the exams, and 1 day to grade the exams and reflect.

__Approve each question. Don’t check every single rubric.__

During the 2-3 work days, require students to get each of their questions “approved”. This means that you have read the question and you know that is can be answered. You will find that students often forget necessary information when writing questions. As for checking rubrics, have students hand in the rubrics on the exam day and then “spot check” them. There is no way you have time to check them all….but don’t worry, if there is an issue with a rubric, it will surface on the day exams are being graded and you can deal with it then.

__How to grade it__

You don’t grade it. Students do. Each exam will be scored out of 30 points.

do you give the students the doc attached to this post or is this just for the teacher?

In the confidence interval example provided, it is stated that the Sampling Distribution of x-bar is approximately normal because the population is Normal. But the random sample is drawn from the population = Mr. Wilcox's AP Stats students which are not clearly defined as a Normal population. Is it ok to assume that subset is normal since the overall population is Normal?