I took the AP Statistics exam May 8, 2002. This writeup will focus on the exam itself, rather than my experience with the class (which is a whole 'nother story).
The exam is divided into two parts, multiple choice and free response. The exam is three hours long, with time split evenly between each section.
You are allowed to use calculators on both sections of the test. Pretty much anything that doesn't have a actual keyboard (like the Texas Instruments 90 series) is ok. You are not limited to one calculator either. Some people in my class used two so that they could continue to calculate if they ran into a problem that required a lot of processing. This is not as much of a problem in statistics as it is in calculus though. I didn't have any problem at all taking the test with only my trusty TI-83.
The test covers a good bit of material. The AP Statistics website divides the test content into four conceptual blocks:
- Understanding properties of sets of data
This covers basic stuff like the mean, median, mode, five number summary, the shape of sets of data, and regression techniques.
- Data gathering
Understanding methods of data gathering, like surveys, observational studies, and experiments. This includes knowing how bias can affect a set of data, and how bias can be controlled.
- Probability
Understanding how probability is used in statistics to draw conclusions about sets of data.
- Statistical Inference
Being able to correctly apply techniques of hypothesis testing to sets of data to determine statistical significance.
The multiple choice section covers each of these generally, you can assume that there will probably be a little bit of everything in that section.
The content free response section is a bit more predictable than that of the multiple choice section. There is always a question on hypothesis testing, one on general analysis of data, and one on experiment design. The free response is all about words. The test will always ask for you to describe things in words, rather than just numbers. I think the idea behind this is that anyone can memorize the calculator function that automatically analyze data, so they need to ask you to write about the meaning of the data to make sure you actually know something about statistics.
The free response section is graded holisticly. If you use a method that is not quite optimal, or is in fact totally wrong, you can still score well on a question if you are consistent.
I found the exam to be easy. The calculator does most of the analysis, so it was just a question of remembering concepts and calculator functions.