Did you start yet? (the remix...added to this post Monday night)

Start your take home test yet? You don't want to have to sit home and do work all day on Wednesday...

Feel free to email with questions! See the post below for one hint..

In about 4 seconds I plan to start making our chapter 13 vocab quiz, which you'll have Friday. Then, you'll get this back Tuesday and we will have our vocab matching test on Wednesday--I also think I'll be adding some multiple choice questions from past AP exams to this test (so it'll be part MC, part vocab). In the meantime (Thursday, Friday, Tuesday) we'll wrap up our learning with experiments. Then, on Thursday and Friday, we'll do a little review/start probability.

As always, this plan is totally open to change, but that's my idea for now...

So, to recap:

• Chapter 13 Vocab Quiz Friday
• Unit Test on Wednesday (vocab matching and some multiple choice)

Need some extra stamps?

Check out the statistics rap songs linked to the right--the "Stat Rap" and the "Stat Rap (Remix)." As you listen, write down any lines (not just one word, write out the entire line of the rap) that reference or are related to statistical concepts we've learned thus far. For every 3 you write down, you'll get a stamp. You can do this to earn up to 10 extra stamps if you like!

Chapter 13 Vocab List (slightly narrowed down...):
Observational Study, Retro/Prospective Observational Study, Experiment, Random Assignment, Factor, Response, Level, Treatment, Principles of Experimental Design, Blocking, Statistically Significant, Control Group, Blinding, Single/Double Blind, Placebo, Confounding, Matching

And here's an email response I wrote to clarify the ideas behind blocking a little...
Blocking is similar to stratifying a sample. With blocking, we want to take our experimental units or participants, and split them into groups based on some characteristic that we know is related to what we're measuring. Let's consider an example...

Suppose you plan to do an experiment testing a new pill to prevent dogs from getting fleas. You have 200 dogs (and dog owners) who have volunteered to participate. One way to do the experiment would be to use a completely randomized design--that is, take the 500 dogs, give 100 the new pill, give 100 no pill, and then measure (over time) how many fleas each dog gets. Then, compare to see if the dogs who took the new pill had a statistically significantly lower number of fleas (if so, then the pill worked).

Now, let's consider a blocked design.

It might not be the best idea to just randomly assign 100 dogs the new pill. For instance, maybe the type of dog (the breed) also affects the number of fleas the dog gets. So, we can block by breed. This means we will first split up our 200 dogs into their breeds. Then, within each block, (or within each breed of dog), we randomly assign some of the breed to get the new pill, and some of the breed to get no pill. We then repeat this process for each block, or each breed.

Breed isn't the only variable that might be related to the amount of fleas a dog gets. Maybe we could instead block by the length of the dogs hair (maybe fleas like longer hair?). Similarly, if we blocked by hair length, we could split the dogs into three blocks--long, medium, and short hair. Then, within each block, we can randomly assign some dogs (some with long hair, for example) the new pill and some (with long hair) no pill. We'd then repeat for each block.

Think--what are some other variables we might block by? What characteristics of dogs might be related to the number of fleas a dog gets?

That's the idea--don't worry, we'll be doing some practice with blocking in class when we get back.