4/17

Speed Dating with L-1 and D-3

You should know how to run 1-var stats on your calculator.

Casio reminder (thanks Sebastian!): if you set up your graphs to see List1 on Graph1 and List2 on Graph2, you can just press the 1-var button when looking at the graph. Easy!

I'm sure I'll have more to say about these problems after we finish them on Wednesday. But for today, that's it!

 

4/16

We worked on K-3. Important ideas!

  • Showing work on normal problems: normal, mu, sigma, boundary, direction. Write down your calculator menus with labels, and draw a sketch. (similar process for binomials)
  • You will never use the normalpdf button.
  • When finding probability with a mean, use sd/sqrt(n).
  • FRQ's with 4 parts and asking for "probability" are not inference problems. Problems with no parts or just (a) and (b) and asking for "statistically significant evidence" are full inference problems (blue sheets).

Last week

Reminders!

  • Reading the entire problem before you start it is a wise choice.
  • The AP exam will give you LOTS of space for your answers-don't feel pressure to fill all that space.
  • When asked to answer Yes or No, start your answer with Yes. Or NO!
  • When asked to choose between two options, say A or B, don't just say good things about the one pick. Also say why it is better than the option you didn't pick.

Also, we noted that the line y=x is often important. It is not a regression line. But it is handy for other uses.

Mock exam

FYI, our mock exam was the 2015 exam and you can find the rubric here. Thing we learned from this test:

  1. Use the statistical words we have learned to answer questions.
  2. If asked yes or no, start your answer "yes" or "no". Also remember that increases the sample size by X times decreases the margin of error by sqrt(X).
  3. Show work! Do NOT round expected value (1.73 ATM's. NOT 2.).
  4. Use the wording of the question and do not use past-tense to describe a parameter. Checking the independence condition is wrong for an experiment.
  5. Frequency means counts.
  6. "Sampling Distribution of Sample Means"--so dang simple man! Just state a center, shape, and spread.

Other tips:

  • Don't feel pressure to fill all the provided space.
  • Reading the whole problem before you answer the first part is often a great choice.

4/9 (or 4/11)

Today we reviewed inference. We completed a 1-prop z-test (B-5), a matched pairs t-test (A-5), and a chi-square 2-way table test (C-2). All of these questions are fair game for Friday's quiz.

Study L-2, super six version!

A few discussion points:

  • Defining the parameter in words is risky. You're best off using only symbols in the Ho/Ha and then careful wording in the conclusion.
  • You need the name OR the formula. Don't write the formula unless you know it is right. And don't write the formula with symbols (too much can go wrong). If you write the formula, just plug in the numbers.
  • Speaking of plugging in the numbers, you must do this on the conditions. Example:
    • n > 30 (not full credit)
    • 50 > 30 (full credit!)

Last week

We used a short answer powerpoint last week. It will be on School Loop later. It will provide you with a good check list of facts you should know.

We also starting reading the 30 pages of Top Tips in the prep book. You should plan on reading all 30 pages.

On this blog I said: "Don't say random residuals, say "the residual plot is random"". But many of you ignored this advice on the quest. Read. Think. Learn. Thanks!

4/4 Misc

Today in class our do-now included many important tips for the AP exam (and for your practice exams!)

  • Bar charts should be described with comparing groups.
  • Quantitative graphs: CUSS-center, shape, and spread.
  • Scatterplots: SOFA-strength, form, and association.

Experiments mean you applied a treatment. And if asked about this, you should specifically describe who got a treatment and what it was.

We also made our t-chart for parameters and statistics. Gosh that chart is important!

4/3 Expected value

The main skill of the day: Expected Value. That's the crucial skill from L-2, part (a). Don't forget that you can label your answer as a mu or E(x), but not x-bar.

Part (b) is another important fact: the law of large numbers. And finally part (d) reminds us that when a distribution is skewed right, the mean is bigger than the median. This fun fact is on the AP test Every. Single. Year.

For our do-now we reviewed 2 skills.

1. ALL the df formulas. Know them. Memorize them.

2. Using Table B and/or invNorm/invT to find z*/t*

Note: All of the above is fair game for Thursday's quest.

 

4/2 Regression

I class today we did problem C-1. A nice regression problem. I also passed out the first part of the Super Six. You should how to do every step of the Super C-1 and then you will be able to every basic regression skill that the AP test will ask.

Someone asked if it is OK to "random residuals". I know that I've put that phrase on the board. On further reflection, it is not clear communication. Instead please say "the residual plot is random". It makes no sense to say a bunch of numbers are random.

I also passed back your regression test. You should fix all your mistakes.

Regression will be a big part of Thursday's Quest.