Page 313, 2006 AP problem answers

So I'm posting a few things...first off, below are the answers to the even problems from your homework.

Second, tomorrow I'll be assigning a worksheet in class to design our experiments discussed in class (I would've given them to you today, but didn't get a chance to pass these out). This will be due Friday, and we can work on it at tomorrow's review. If you want a head start, it's under the Classroom Powerpoints link under "Experimental Design Practice."

Finally, I've posted both the AP rubrics for the 2 problems you did with the sub under the same link entitled. "2006 #5, and 2006B #5."

Enjoy!

Page 313:
24. The "control" in this experiment is not the same for all swimmers. We don't know what "their old swim suit" means. They should compare their new swim suit to the same suit design. The order in which the swims are performed should be randomized. There may be a systematic difference from one swim to the next. For instance, swimmers may be tired after the first swim (or more warmed up). Finally, there is no way to blind this test. The swimmer will know which kind of suit they have on, and this may bias their performance.

26. a. Self-selection could result in groups that are very different at teh start of the experiment, making it impossible to attribute differences in the results to the diet alone.

b. The meals were prepared by dieticians to ensure that the diets were followed and that all subjects received comparable treatments.

c. The researchers can compare the change in blood pressure observed in the DASH group to the control group. They need to rule out the possibility that external variables (like the season, news events, etc.) affected everyone's blood pressure.

d. We would like to know the standard deviation of the changes, as well. If the standard deviation is very small, then 6.7 points would seem like a significant change. If not, 6.7 points could be due to naturally occurring variability.

28. Answers may vary. Suppose you select the next 12 batches of sake for the experiment, in order to replicate the experiment. For each batch, divide the yest into two parts. Randomly select half of the yest and play Mozard for that half and no Mozart for the other half. When the sake is done, have trained sake testers taste samples from each group and rate the sake. In order to eliminate potential bias on the part of the testers, blinding should be used. Do not tell the tasters whether or not the sake has been brewed with Mozart playing next to the yeast. Compare the ratings for the Mozart and the non-Mozart sake. Try to keep other variables (type of yeast, size of batch, etc.) under control.