Suppose a new drug treatment for headaches is compared to placebo ...

[saliriagwu]

Suppose a new drug treatment for headaches is compared to placebo in a two independent group matched-pairs study (measuring difference in pain levels before and after for the placebo group and for the drug group). Suppose the severity of symptoms was substantially reduced for both the placebo and the treatment groups. What should be done next?
 a. The placebo group should be ignored from this point on, focusing only on the drug group.
  b. The placebo effect should be taken into account by comparing the two groups directly, so we can examine the additional influence of taking the drug on the reduction of symptoms.
  c. We should conclude that the drug doesn't do any better job than the placebo does.
  d. None of the above.

Question 2

Are all confidence intervals (whether they be for population means, proportions, or anything else) symmetric around their respective sample values? Explain.

[tanna.moeller]

Answer to Question 1

B

Answer to Question 2

NO. THE RULE FOR SAMPLE PROPORTIONS AND THE RULE FOR SAMPLE MEANS SAY THAT THIS IS TRUE FOR MEANS AND PROPORTIONS, IF THE SAMPLE SIZE IS LARGE ENOUGH. BUT FOR OTHER POPULATION VALUES, THEIR SAMPLE ESTIMATES MAY NOT HAVE A SYMMETRIC SHAPE.