THE UNIVERSITY OF CALGARY DEPARTMENT OF MATHEMATICS AND STATISTICS STATISTICS 217 - L01, L02 and L06 FINAL EXAMINATION

April 24
SURNAME

Winter, 2008
GIVEN NAMES

Time: 2 Hours
I.D.

STUDENT IDENTIFICATION Each candidate must sign the Seating List conrming presence at the examination. All candidates for nal examinations are required to place their University of Calgary student I.D. cards on their desks for the duration of the examination. (Students writing mid-term tests can also be asked to provide identity proof.) Students without an I.D. card who can produce an acceptable alternative I.D., e.g., one with a printed name and photograph, are allowed to write the examination. A student without acceptable I.D. will be required to complete an Identication Form. The form indicates that there is no guarantee that the examination paper will be graded if any discrepancies in identication are discovered after verication with the students le. A student who refuses to produce identication or who refuses to complete and sign the Identication Form is not permitted to write the examination. EXAMINATION RULES 1. Students late in arriving will not normally be admitted after one-half hour of the examination time has passed. 2. No candidate will be permitted to leave the examination room until one-half hour has elapsed after the opening of the examination, nor during the last 15 minutes of the examination. All candidates remaining during the last 15 minutes of the examination period must remain at their desks until their papers have been collected by an invigilator. 3. All enquiries and requests must be addressed to supervisors only. 4. Candidates are strictly cautioned against: (a) speaking to other candidates or communicating with them under any circumstances whatsoever; (b) bringing into the examination room any textbook, notebook or memoranda not authorized by the examiner; (c) making use of calculators and/or portable computing machines not authorized by the instructor; (d) leaving answer papers exposed to view; (e) attempting to read other students examination papers. The penalty for violation of these rules is suspension or expulsion or such other penalty as may be determined. 5. Candidates are requested to write on both sides of the page, unless the examiner has asked that the left half page be reserved for rough drafts or calculations. 6. Discarded matter is to be struck out and not removed by mutilation of the examination answer book. 7. Candidates are cautioned against writing in their answer books any matter extraneous to the actual answering of the question set. 8. The candidate is to write his/her name on each answer book as directed and is to number each book. 9. A candidate must report to a supervisor before leaving the examination room. 10. Answer books must be handed to the supervisor-in-charge promptly when the signal is given. Failure to comply with this regulation will be cause for rejection of an answer paper. 11. If a student becomes ill or receives word of domestic aiction during the course of an examination, he/she should report at once to the Supervisor, hand in the unnished paper and request that it be cancelled. Thereafter, if illness is the cause, the student must go directly to University Health Services so that any subsequent application for a deferred examination may be supported by a medical certicate. An application for Deferred Final Examinations must be submitted to the Registrar by the date specied in the University Calendar. Should a student write an examination, hand in the paper for marking, and later report extenuating circumstances to support a request for cancellation of the paper and for another examination, such request will be denied. 12. SMOKING DURING EXAMINATIONS IS STRICTLY PROHIBITED.

Page 1 of 15

Stat 217 - L01, L02 and L06 - Final Exam - Winter 2008 Introduction

Page 2 of 15

1. Blacken the course number (STAT 217) and section number (L01, L02 or L06), your name and ID on the corresponding space on the answer sheet with a No. 2 pencil. 2. This examination is scheduled for a 2-hour period. It has 38 multiplechoice questions. All questions are equally weighted. Blank answers and incorrect answers are scored as zero credit, while correct answers are worth 1 point each. For questions involving numerical answers, carry your computation to as many signicant digits as shown in the available answers and select the best answer that you think to be closest to the correct answer. 3. Read each question and its lettered answers. When you have decided which answer is correct, blacken the corresponding space on the answer sheet with a No. 2 pencil. Make no stray marks; they may count against you. 4. This is a closed book exam. A standard normal z-table, a 2 -table, a t-table and F -tables are provided. A non-programmable and nongraphing calculator is allowed. A two-sided, 8.5 x 11, formula sheet is allowed. 5. Hand in the exam papers with the answer sheet.

Stat 217 - L01, L02 and L06 - Final Exam - Winter 2008

Page 3 of 15

Table 1: Correct answers to the multiple-choice questions Question Your Question Your Question Your number answer number answer number answer 1 C 16 C 31 B 2 D 17 B 32 A 3 C 18 C 33 D 4 A 19 B 34 B 5 C 20 B 35 C 6 B 21 C 36 D 7 D 22 A 37 C 8 A 23 C 38 B 9 B 24 B 10 A 25 B 11 A 26 A 12 C 27 C 13 B 28 B 14 A 29 C 15 D 30 A

Stat 217 - L01, L02 and L06 - Final Exam - Winter 2008

Page 4 of 15

1. The heights (in inches) of males in the U.S. are believed to be normally distributed. The average height of a random sample of 25 American adult males is found to be x = 69.72 inches with a standard deviation of s = 4.15. What is the standard error of x ? A) 0.17 B) 0.69 C) 0.83 D) 2.04 2. The hypotheses H0 : = 350 versus Ha : < 350 are examined using a sample of size n = 20. The one-sample t statistic has the value t = 1.68. What do we know about the P -value of this test? A) P value < 0.01 B) 0.01 < P value < 0.025 C) 0.025 < P value < 0.05 D) P value > 0.05 CAUTION: the next 3 questions refer to the following information: You wish to compare the prices of apartments in two neighboring towns. You take a simple random sample of 12 apartments in town A and calculate the average price of these apartments. You repeat this for 15 apartments in town B. Let 1 represent the true average price of apartments in town A and 2 the average price in town B. 3. What would be the hypotheses for this problem? A) H0 : 1 = 2 versus Ha : 1 < 2 B) H0 : 1 = 2 versus Ha : 1 > 2 C) H0 : 1 = 2 versus Ha : 1 = 2 D) None of the above. 4. If we were to use the unpooled t test, what would be the conservative estimate for the degrees of freedom? A) 11 B) 12 C) 14

Stat 217 - L01, L02 and L06 - Final Exam - Winter 2008 D) 25

Page 5 of 15

5. Suppose we were to use the unpooled t test with the conservative estimate for the degrees of freedom. The t statistic for comparing the mean prices is 2.1. What can we say about the value of the P -value? A) P value < 0.01 B) 0.01 < P value < 0.05 C) 0.05 < P value < 0.10 D) P value > 0.10 CAUTION: the next 3 questions refer to the following information: A political analyst was curious if younger adults were becoming more conservative. He decided to see if the mean age of registered Republicans was lower than that of registered Democrats. He selected a simple random sample of 128 registered Republicans from a list of registered Republicans and determined the mean age to be x1 = 39 years with a standard deviation s1 = 8 years. He also selected an independent simple random sample of 200 registered Democrats from a list of registered Democrats and determined the mean age to be x2 = 40 years with a standard deviation s2 = 10 years. Let 1 and 2 represent the mean ages of the populations of all registered Republicans and Democrats, respectively. Let 1 and 2 be the corresponding population standard deviations. Suppose it is reasonable to assume that the distributions of age in the populations of registered Republicans and of registered Democrats are normal with the same standard deviation. 6. What is a 90% condence interval for 1 2 ? A) 1 1.66 years B) 1 1.74 years C) 1 1.98 years D) 1 2.76 years 7. Suppose the political analyst had wished to test the hypotheses H0 : 1 = 2 versus Ha : 1 < 2 . What can we say about the value of the P -value?

Stat 217 - L01, L02 and L06 - Final Exam - Winter 2008 A) P value < 0.025 B) 0.025 < P value < 0.05 C) 0.05 < P value < 0.10 D) P value > 0.10

Page 6 of 15

8. Suppose we wish to test the hypothesis that the groups are equivalent in how variable their ages are. To do this, we wish to test the hypotheses H0 : 1 = 2 versus Ha : 1 = 2 . What can we say about the value of the P -value? A) P value < 0.05 B) 0.05 < P value < 0.10 C) 0.10 < P value < 0.20 D) P value > 0.20 CAUTION: the next 2 questions refer to the following information: A sociologist is studying the eect of having children within the rst two years of marriage on the divorce rate. Using hospital birth records, she selects a simple random sample of 200 couples who had children within the rst two years of marriage. Following up on these couples, she nds that 80 couples are divorced within ve years. Let p = the population proportion of couples who had children within the rst two years of marriage and are divorced within ve years. 9. What is a 90% plus four condence interval for the proportion of couples who had children within the rst two years of marriage and are divorced within ve years? A) 0.402 0.035 B) 0.402 0.056 C) 0.402 0.067 D) 0.402 0.088
1 10. Suppose the sociologist wishes to test H0 : p = 3 versus Ha : p = 1 . 3 At the 10% signicance level, are the data statistically signicant for testing these hypotheses? A) Yes

Stat 217 - L01, L02 and L06 - Final Exam - Winter 2008

Page 7 of 15

B) No C) This cannot be determined from the information given. CAUTION: the next 4 questions refer to the following information: A simple random sample of 100 college students was interviewed. They were asked what size pizza they usually order and what their favorite topping is. The results are presented below. Topping Pepperoni Veggie 18 11 14 12 3 9 35 32

Size Small Medium Large Total

Cheese 6 7 20 33

Total 35 33 32 100

11. What would be null hypothesis for a chi-square test based on these data? A) Pizza topping and pizza size are independent. B) The average pizza size is the same for pepperoni, veggie, and cheese pizzas. C) The distribution of pizza topping is the same for small, medium, and large pizzas. D) The distribution of pizza size is dierent for the three dierent pizza toppings. 12. Under the null hypothesis that there is no association between pizza size and pizza topping, what is the value of the expected count for a small pepperoni pizza? A) 11.67 B) 12 C) 12.25 D) 13 13. What are the appropriate degrees of freedom for the chi-square statistic? A) 3

Stat 217 - L01, L02 and L06 - Final Exam - Winter 2008 B) 4 C) 6 D) 8

Page 8 of 15

14. What is the contribution to the chi-square statistic from the cell of a large cheese pizza? A) 8.4 B) 9.3 C) 20.4 D) 52.8 CAUTION: the next 3 questions refer to the following information: A particular paperback mystery book is published with a choice of three dierent pictures on the cover: a photograph of the actor playing the main character in the movie version of the book, a drawing of the mansion where the story in the book takes place, or an embossed graphic of the murder weapon. A certain bookstore keeps copies of this book with each of the pictures on the cover on its racks. To test the hypothesis that sales of this book are equally divided among the three choices, a simple random sample of 120 purchases of this book is obtained. The numbers are displayed in the table below. Picture on the cover Photograph Drawing Embossed graphic Total 31 47 42 120 15. Under the null hypothesis that sales are equally divided among the three dierent covers, what is the value of the expected count for sales of the book with the photograph of the actor on the cover? A) 0.258 B) 0.333 C) 31 D) 40 16. What is the value of the chi-square statistic? A) 1.225

Stat 217 - L01, L02 and L06 - Final Exam - Winter 2008 B) 3 C) 3.35 D) 40

Page 9 of 15

17. At the 10% signicance level, do we reject the null hypothesis? A) Yes B) No C) This cannot be determined from the information given. CAUTION: the next 3 questions refer to the following information: The data referred to in this question were collected on 41 employees of a large company. The company is trying to predict the current salary of its employees from their starting salary (both expressed in thousands of dollars). The Minitab regression output is given below as well as some summary measures: Descriptive Statistics Minimum Maximum Mean Starting Salary 10.95 44.10 17.9818 Current Salary 19.65 100.00 36.2543 Model Summary R Square Adjusted .833926

Variance 57.472 305.297

R (or r) .913195

R Square .829667

Source Regression Error Total

ANOVA df Sum of Squares Mean Square F p-value 1 10183.816 10183.816 195.835 .000 39 2028.084 52.002 40 12211.900

Coecients Predictor Coecients Std. Error t p-value Constant -1.59266 2.92961 -.544 .590 Starting Salary 2.10473 .15040 13.994 .000

Stat 217 - L01, L02 and L06 - Final Exam - Winter 2008

Page 10 of 15

18. What is an (approximate) 95% condence interval for the slope 1 ? A) (-7.57, 4.39) B) (-4.52, 1.34) C) (1.80, 2.41) D) (1.95, 2.26) 19. Suppose we wish to test the hypotheses H0 : 1 = 2 versus Ha : 1 = 2. Together with an insignicant constant in this model, this would imply that the employees currently earn about twice as much as their starting salary. At the 5% signicance level, would we reject the null hypothesis? A) Yes B) No C) This cannot be determined from the information given. 20. John Doe works for this company. He started with a salary of $15,300. Predict his current salary with a 90% interval. Express the interval in the appropriate units. A) ($15,683; $45,537) B) ($18,205; $43,014) C) ($28,580; $32,640) D) ($31,516; $32,885) CAUTION: the next 4 questions refer to the following information: A researcher is investigating possible explanations for deaths in trafc accidents. He examined data from 1991 for each of the 50 states plus Washington, D.C. The data included information on the following variables: Deaths The number of deaths in trac accidents Income The average income per family Children The number of children (in multiples of 100,000) between the ages of 1 and 14 in the state As part of his investigation he ran the following multiple regression

Stat 217 - L01, L02 and L06 - Final Exam - Winter 2008 model

Page 11 of 15

Deaths = 0 + 1 (Children) + 2 (Income) + i , where the deviations i were assumed to be independent and normally distributed with mean 0 and standard deviation . This model was t to the data using the method of least-squares. The following results were obtained from statistical software. Source Model Error Variable Constant Children Income Sum of Squares 48362278 3042063 df 2 48

Coecient 593.829 90.629 -0.039

Standard Error 204.114 3.305 0.015

21. Suppose we wish to test the hypotheses H0 : 1 = 2 = 0 versus Ha : at least one of the j is not 0 using the ANOVA F -test. What is the value of the F statistic? A) 12.45 B) 17.3 C) 381.5 D) 402.5 22. What can we say about the P -value for the ANOVA F -test? A) P value < 0.002 B) 0.002 < p value < 0.01 C) 0.01 < P value < 0.02 D) P value > 0.02 23. What is a 95% condence interval for 1 , the coecient of the variable Children? A) 90.629 5.55

Stat 217 - L01, L02 and L06 - Final Exam - Winter 2008 B) 90.629 3.305 C) 90.629 6.68 D) 90.629 4.65

Page 12 of 15

24. What proportion of the variation in the variable Deaths is explained by the explanatory variables Children and Income? A) 0.380 B) 0.941 C) 0.980 D) 0.840 CAUTION: the next 4 questions refer to the following information: A store owner wishes to compare the average amount of money high school and college students spend on CDs. He randomly selects ten students from three dierent student populations: high school students, undergraduate students, and graduate students. The statistical assumptions required to perform a one-way ANOVA to compare the means of these three groups are reasonable based on the data. A partially completed ANOVA table is provided below.

Source Sum of Squares DF Groups Error 3240 4450 Total

Mean Square

F p-value

25. What are the degrees of freedom for the F statistic? A) 2 and 10 B) 2 and 27 C) 3 and 27 D) 3 and 30 26. What is the value of the pooled standard deviation? A) 10.95 B) 24.60 C) 56.92

Stat 217 - L01, L02 and L06 - Final Exam - Winter 2008 D) 120 27. What is the value of the F statistic? A) 3.36 B) 3.73 C) 5.04 D) 5.99

Page 13 of 15

28. At the 1% signicance level, would we reject the null hypothesis? A) Yes B) No C) This cannot be determined from the information given. CAUTION: the next 3 questions refer to the following information: Twenty high school-aged students are randomly selected from three dierent school districts: a district in the city, a district in the suburbs, and a district in a rural area. Each group of twenty students consisted of 10 boys and 10 girls. Each of the students was asked what price they paid for their last haircut. The two-way ANOVA output is shown below in an incomplete form.

Source Sex Region Sex*Region Error Total

Sum of Squares DF Mean Square 2674.005 1 2674.005 703.583 2 351.791 231.803 115.901 2143.037 39.686 5752.427 59

F p-value 67.379 0.000 8.864 0.000 2.920

29. What is the value of the sum of squares for the model? A) 2143.037 B) 3377.588 C) 3609.391 D) 5752.427

Stat 217 - L01, L02 and L06 - Final Exam - Winter 2008

Page 14 of 15

30. The degrees of freedom for the interaction term and the error term are missing from the table. What are their values? A) 2 and 54 B) 6 and 54 C) 2 and 56 D) 6 and 56 31. Is the interaction eect statistically signicant at the 5% signicance level? A) Yes B) No C) This cannot be determined from the information given. CAUTION: the next 4 questions refer to the following information: Does income increase when you get married? To investigate this, a college student has randomly selected people from his college town and collected income data for these people. They were categorized according to their marital status. The following results (in thousands of dollars) were obtained: Marital status Income ($1000) Married 136 116 82 166 72 Not married 33 78 76 125 80 The data are to be analyzed using the Wilcoxon rank sum test. 32. The null hypothesis is that income has the same distribution for married and unmarried people. What is the appropriate alternative hypothesis? A) The distribution of income for married people has values that are systematically larger than for unmarried people. B) The distribution of income for unmarried people has values that are systematically larger than for married people. C) The distribution of income for one of the two groups of people has values that are systematically larger than for the other group. D) None of the above.

Stat 217 - L01, L02 and L06 - Final Exam - Winter 2008

Page 15 of 15

33. Arrange the ten observations in order and assign ranks. What rank does the observation with an income of $116,000 receive? A) 3 B) 4 C) 5 D) 7 34. What is the standard deviation of W under the null hypothesis that the two distributions are the same? A) 4.56 B) 4.79 C) 22.9 D) 38.67 35. What is the observed value of W ? A) 6 B) 21 C) 34 D) 55 CAUTION: the next 3 questions refer to the following information: At what age do babies learn to crawl? Does it take longer to learn in the winter when babies are often bundled in clothes that restrict their movement? Data were collected from parents who brought their babies into the University of Denver Infant Study Center to participate in one of a number of experiments between 1988 and 1991. Parents reported the birth month and the age at which their child was rst able to creep or crawl a distance of four feet within one minute. The resulting data were grouped by month of birth: January, May, and September. Birth month January May September Crawling age (month) 26 29 33 27 29 31 31 34 37

The data is to be analyzed with the Kruskal-Wallis test.

Stat 217 - L01, L02 and L06 - Final Exam - Winter 2008

Page 16 of 15

36. The null hypothesis is that the crawling ages have the same distribution in all groups. What is the alternative hypothesis? A) Not all three mean crawling ages are equal. B) The mean crawling age in September is higher than the other two months. C) The crawling times in September are systematically higher than in the other two months. D) None of the above. 37. What is the value of the Kruskal-Wallis statistic H for these data? A) H = 2.77 B) H = 3.10 C) H = 3.76 D) H = 5.99 38. At the 5% signicance level, would we reject the null hypothesis? A) Yes B) No C) This cannot be determined from the information given.

Have a nice summer!