# Statistics in a Modern World 800 Homework 5 NAME:_________________________________

Homework 5 Exercises Assigned from Chapters 12-15 (14 pts.) due Fri., October 16 in Lecture.

### CHAPTER 12

#3 (3 pts.) According to the University of California at Berkeley Wellness Letter (Feb 1994, p.1), only 40% of all surgical operations require an overnight stay at a hospital.

1. The proportion of surgical operations requiring an overnight stay is ____?
2. The risk of requiring an overnight stay is ___?
3. The odds of requiring an overnight stay are ___?

#13 (3 pts.) Reporting on a study of drinking and drug use among college students in the U.S., a Newsweek reporter wrote: "Why should college students be so impervious to the lesson of the morning after? Efforts to discourage them from using drugs actually did work. The proportion of college students who smoked marijuana at least once in 30 days went from one in three in 1980 to one in seven last year [1993]; cocaine users dropped from 7% to 0.7% over the same period." (19 December 1994, p.72)

1. What was the relative risk of cocaine use for college students in 1980 compared with college students in 1993?
2. Are the figures for marijuana use (for example, "one in three") presented as proportions or odds?
3. Is the statement that "efforts to discourage them from using drugs actually did work" justified? Answer yes or no and explain briefly.

### CHAPTER 13

#14 (2 pts.) Compute the chi-squared statistic for the relationship between bird ownership and lung cancer, based on the data in Exercise 15 p.237.

Is there statistically significant evidence of a relationship?

### CHAPTER 14

#1 (1 pt.) The price of a first-class stamp in 1970 was 8 cents, whereas in 2002 it was 37 cents. The Consumer Price Index for 1970 was 38.8, whereas for 2002 it was 172.2. If the true cost of a first-class stamp did not increase between 1970 and 2002, what should it have cost in 2002? In other words, what would an 8 cent stamp in 1970 cost in 2002, when adjusted for inflation?

### CHAPTER 15

#1 (2 pts.)For each of the following time series, do you think the long-term trend would be positive, negative, or nonexistent?

1. The cost of a loaf of bread measured monthly from 1960 to 2004.
2. The temperature in Boston measured at noon on the first day of each month from 1960 to 2004.
3. The price of a basic computer, adjusted for inflation, measured monthly from 1970 to 2004.
4. The number of personal computers sold in the U.S. measured monthly from 1970 to 2004.

#2.(1 pt.) Which of the four time series in Exercise 1 would have the strongest seasonal component? [Circle one.] (a) (b) (c) (d)

#18 (2 pts.) According to the World Almanac and Book of Facts (1995, p.380), the population of Austin, Texas (reported in thousands), has grown as follows:

Year 1950 1960 1970 1980 1990

Population 132.5 186.5 253.5 345.5 465.6

1. Of the three nonrandom components of time series, which do you think would be most likely to explain the data if you were to see the population of Austin, Texas, by month, from 1950 to 1990? (a) trends (b) seasonal (c) cycles
2. The regression equation relating the last two digits of each year (50,60, and so on) to the population for Austin, Texas, is: population = -301 + 8.25(year) >Use this equation to predict the population of Austin for the year 2000.
3. Consider the method you used for the prediction in part b. Do you think it is likely to be accurate? (Answer yes or no.)
4. Would the same method continue to give accurate predictions for the years 2010, 2020, and so on?