Chapter 19 Solutions

Problem 3.

a) It means that the proportion answering "Yes, Should" in the sample is within 4% of the proportion of all adults nationwide who would answer "Yes". This can be verified by noting that the margin of error is 0.0394 or about 0.04.

b) 31% +/- 4% or 27% to 35%.

c) It probably represents a real difference. If the margin of error for each poll is 4%, then a confidence interval for the June poll is 15% to 23%, which does not overlap at all with the confidence interval given in part b.

Problem 4.

a) 0.0229

b) 2 standard deviations is about 4.5.

c) a 95% confidence interval is 70.5% to 79.5%; we are 95% confident that the percentage of adult Americans who think that Congress should maintain the ban is between 70.5% to 79.5%.

Problem 5.

a) 68%

a) 90%

a) 95%

a) 99%

Problem 6.

a) decrease

b) remain the same

c) decrease

Problem 11.

a) 56% to 64%

b) yes, we can be 95% sure that the true percentage of people who prefer the quarter system is 56% to 64% which is a majority.

c) 46% to 74%. Because the CI goes below 50% this does not provide convincing evidence that a majority prefer to remain on the quarter system.

d) The larger sample size in part a yielded a smaller standard deviation and thus a smaller CI. In general the larger the sample size the smaller the 95% CI we can get, so the sample size alone can determine whether we accept that the majority of a population indeed has a certain trait.

Problem 12.

a) A 95% CI is 54.6% to 63.4%. This means that we are 95% confident that the portion of American Catholics who believe that women should be allowed to be priests is between 54.6% to 63.4%.

b) The standard deviation is about 0.022, so a 95% CI is 0.546 to .6334, the same as in a.

c) Using the SD, a 95% confidence interval is about 86.2% to 91.8%. Using the margin of error, th CI is 84.6% to 93.4%, much wider than the first.

d) the intervals in parts a and b were quite similar because the proportion was close to 0.5. In part c the proportion was further away from 0.5, so 2*SD is smaller than the margin of error.

Problem 13.

a) sqrt(1/1000) = 0.03

b) We are 95% confident that the portion of American adults who feel that the world will come to an end is between 56% and 62%.

Problem 14.

a) The SD = 0.02, so the 95% CI is 29% to 37%.

b) You can't use the +/- 3% margin of error because this is based on the 1000 sample size.

Problem 15.

a) the 95% CI is 20% to 24%. This means that we can be 95% confident that the percentage of American college student binge drinkers having unprotected sex is between 20% to 24%.

b) The sample proportion is 0.02; the SD is 0.001. A 95% CI is 1.8 to 2.2%.

Problem 16.

The SD is about 0.04, so a 99.7% CI is about 0.37 +/- 3(0.04) or 25% to 48%. A critic would prefer to use this interval because it does include the 25% value that would be expected if there were no ESP at work.