January 2017
Beginner
882 pages
203h 41m
English
21. Possibly. Data from one day is not representative of variation in public transit time.
23. Pairing reduces the analysis to a one-sample analysis. Form the differences and compute the confidence interval for the average difference.
Take 40% of the two endpoints, [$200, $880] profits.
Yes, methods used by Group A sell statistically significantly more. Due to the randomization, confounding is not an issue.
Randomization makes this bias unlikely.
27. Compare the change in consumption from 2005 to 2006 in homes that were on daylight savings time to the change in homes that moved from standard to daylight savings time.