CP Stats: Sample report form · May 23, 06:23 PM
My answers to the 2008 test · May 19, 12:20 PM
Here are my answers.
Link to the Against All Odds case study · May 15, 08:20 AM
Augggggh! · May 2, 03:01 PM
PHANTOMS and PANIC are important. I’m posting a copy of my answer to question 5b (Night owls and early birds) along with a few comments. The grading rubric breaks the problem into these three parts:
- Parameters and hypotheses
- Mechanics
- Conclusion in context
Those three parts connect directly to our mnemonic like this:
- PHAN
- TO
- MS
Each part is either correct or incorrect; there is no partially correct to assign. Therefore, if you skipped the assumptions, you got part 1 wrong. If you didn’t conclude in context, you got part 3 wrong.
Well, I hope to see a lot of you tomorrow. Let’s fix this.
Quick fixes to drastically improve your score · May 1, 12:11 PM
I’ve scored everyone’s paper. I’ll give them back after questions 5 and 6, tomorrow. Prepare to use inference procedures.
As a class, your scores are a little — and I mean only a little — lower than I’d like to see. Using your average score on questions 1 through 4 as your scores on questions 5 and 6, the class median AP grade is a 3 with four 4’s, seven 3’s, three 2’s, two 1’s and 2 incompletes. A significant portion of you are near a border with the grade above.
A major part of low scoring is a lack of attention to detail. I can read between the lines to see that you essentially get something, but the scoring scheme doesn’t allow for much of that.
Here are my general comments about how each question went.
Question 1: Einstein vs. Newton on curvature of space
- When the question asks “how the precision . . . has changed over time” your answer should say the precision has (in this case) increased over time. Some people described the change correctly without tying it back to the word “precision.”
- Many of you nailed this question.
Question 2: Boot experiment
- Random assignment to treatments. Some of you didn’t say that subjects would be randomly assigned to treatments. The rest of you did, but you didn’t describe how you’d assign digits and use the table of random digits or a computer to actually assign subjects to treatments. This is an easy thing to fix. Generally speaking, this fix alone would add nearly 2 points to your score.
- The groups that must be blinded for an experiment to be double blind are the influencers (like the volunteers) and the boot evaluators. Neither group needs to know which coating the boot received so this could be double blind. Another easy fix and another 2 points.
Question 3: New High School runners
- Show your work. If you’re adding four numbers up, write them down and put plus signs between them.
- The problem states that the team time is the sum of the individuals times. Therefore the mean for the team is the sum of the individual times (18.9 minutes) and not the mean of the means.
- To combine standard deviations of independent random variables, you add the variances. Remember your “Pythagorean Theorem.” (The answer is 0.30).
- Do what you can do. If your answers to part b were wrong, use them in part c and get full credit. This would net you another 2 points (nearly) on your quest to change your AP grade a level.
Question 4: Commercial airlines
- The program reports r 2. You need to take the square root to find r = 0.755.
- r tells us the strength and direction of a linear association. Give answers in context: There is a strong, positive linear association between the number of seats in a plane and the cost of operating a plane.
- I know that you’re rusty on regression, but this question is well within reach of what you can do. You simply need to know how the computer output matches up with the numbers you obtain on the calculator. Moreover, you need to know which numbers, like R-Sq (adj), to ignore entirely.