Chapter 8-10 ReviewMULTIPLE CHOICE Answers
1.ANS:APTS:1TOP:Significance test for p1-p2: test statistic expression
2.ANS:APTS:1TOP:Conclusion—association not causation
4.ANS:BPTS:1TOP:Choosing a procedure
7.ANS:CPTS:1TOP:Confidence interval for mu1-mu2 (formula)
8.ANS:CPTS:1TOP:Significance test for mu1-mu2: calculate test statistic
9.ANS:EPTS:1TOP:Conditions for confidence interval for mu
10.ANS:BPTS:1TOP:Choosing sample size (proportions)
12.ANS:C
/C/Correct! Formula for two-sample t-statistic for the difference of two means is .
13.ANS:CPTS:1TOP:Using t-table to find critical t* values
14.ANS:APTS:1TOP:Conditions for confidence interval for p
16.ANS:CPTS:1TOP:Stating hypotheses
17.ANS:DPTS:1TOP:P-value only give evidence against null
18.ANS:B
/B/Correct! I is true because power is greater if the alternative (actual) value of the parameter is farther from the null. III is true because P(Type I error) = and P(Type II error) depends on . II is false, since P(Type I error) = , not P(Type II error). PTS: 1
19.ANS:BPTS:1
TOP:Conservative versus software values for degrees of freedom
20.ANS:DPTS:1TOP:Interpret P-value
21.ANS:DPTS:1TOP:Using tests wisely (poor design)
22.ANS:E
/E/Correct! While this is a probability sample of sorts, it is not an SRS. Because she sampled for only one two-hour period, the sample is subject to bias arising form the specific day and time she sampled.
23.ANS:CPTS:1TOP:Factors influencing width of confidence interval
24.ANS:D
/D/Correct! For tests of significance, use the combined estimate in the formula .
25.ANS:EPTS:1TOP:Confidence interval for p1-p2 (computation)
26.ANS:CPTS:1TOP:Paired vs. independent samples
28.ANS:BPTS:1TOP:Type I and II error
30.ANS:EPTS:1TOP:Probability calculation for sampling distribution of p1-p2
35.ANS:B/B/Correct!
36.ANS:E
/E/Correct! Since both sample sizes are less than 15, t-procedures are not recommended in the presence of skew or outliers. This is especially true when the two sample sizes are quite different.
38.ANS:DPTS:1TOP:Conditions for paired t-test
39.ANS:A
/?/Correct! Since the mean of all the distributions is 0, the probability that a value is less than 3 is always greater than the probability it is greater than 3. The t-distributions have “heavy tails” in comparison to the standard Normal distribution, which means . By the complement principle,
40.ANS:C
/C/Correct! Critical z* for 95% is 1.96; critical z* for 98% is 2.33.