Statistics – Final exam Spring 2017 Solutions Question 1 a) The probability of rolling a 4 is: ๐ 4 = ๐ ๐๐๐๐ ๐๐๐ 1 โ ๐ ๐๐๐๐ 4 ๐ค๐๐กโ ๐๐๐ 1 + โฏ + ๐ ๐๐๐๐ ๐๐๐ 5 โ ๐ ๐๐๐๐ 4 ๐ค๐๐กโ ๐๐๐ 5 Therefore, we need to determine the probability of rolling a 4 with each of the 5 dice and the probability of picking each of the 5 dice. We can easily calculate these probabilities for the three types of dice: • ! With a die with 6 faces, the probability of rolling a 4 is . There are two dice with ! ! 6 faces, so the probability of getting a 6-sided die is . Therefore, the probability • ! ! ! ! ! ! !" of rolling a 4 with a 6-sided die is โ = ! With a die with 8 faces, the probability of rolling a 4 is . There are two dice with ! ! 8 faces, so the probability of getting an 8-sided die is . Therefore, the probability ! ! ! ! ! !" of rolling a 4 with an 8-sided die is โ = • ! With a die with 10 faces, the probability of rolling a 4 is ! !" . There is only one die ! with 10 faces, so the probability of getting a 10-sided die is . Therefore, the ! ! ! ! probability of rolling a 4 with a 10-sided die is โ = ! !" !" Finally, the probability of rolling a 4 with a die chosen randomly from among the 5 dice is: 1 1 1 ๐ 4 = + + = 0.1367 15 20 50 b) We can apply the same strategy as above: • With a die with 6 faces, the probability of rolling an 8 is 0. Therefore, the ! probability of rolling an 8 with a 6-sided die is โ 0 = 0 ! • ! With a die with 8 faces, the probability of rolling an 8 is . There are two dice ! ! with 8 faces, so the probability of getting an 8-sided die is . Therefore, the ! ! ! ! ! !" probability of rolling an 8 with an 8-sided die is โ = • With a die with 10 faces, the probability of rolling an 8 is ! ! !" . There is only one die ! with 10 faces, so the probability of getting a 10-sided die is . Therefore, the ! ! ! ! probability of rolling an 8 with a 10-sided die is โ = ! !" !" Finally, the probability of rolling an 8 with a die chosen randomly from among the 5 dice is: 1 1 ๐ 8 = + = 0.07 20 50 Page 1 of 4 c) We can again use the same strategy: • With a die with 6 faces, the probability of rolling a 9 is 0. Therefore, the ! probability of rolling a 9 with a 6-sided die is โ 0 = 0 ! • • With a die with 8 faces, the probability of rolling a 9 is 0. Therefore, the probability of rolling a 9 with an 8-sided die is 0 ! With a die with 10 faces, the probability of rolling a 9 is . There is only one die !" ! with 10 faces, so the probability of getting a 10-sided die is . Therefore, the ! ! ! ! probability of rolling a 9 with a 10-sided die is โ = ! !" !" Finally, the probability of rolling a 9 with a die chosen randomly from among the 5 dice is: 1 ๐ 9 = = 0.02 50 d) In order to decide which bet to take, you need to compare the probability of rolling a 4 in a) to the probability of rolling an 8 or a 9, which is the sum of the probabilities in b) and c): ๐ 4 = 0.1367 > ๐ 8 or 9 = ๐ 8 + ๐ 9 = 0.09 Therefore, it is best to bet on a 4. Question 2 a) Although ๐ can take an infinity of values, we can count these values: value number 1 is 3! , value number 2 is 3! , value number 3 is 3! , and so on. Therefore, ๐ is a discrete random variable. b) We use the formula for the expected value of a discrete random variable: ! ๐ธ ๐ = ! ๐ฅ๐(๐ฅ) = !!! 2 3 โ != 3 ! ! !!! 2=∞ !!! Therefore, ๐ does not have an expected value. c) We now use the formula for an expected value of a function of ๐, which in this case is โ ๐ = ๐ !: ! ๐ธ ๐ ! =๐ธ โ ๐ = ! โ(๐ฅ)๐(๐ฅ) = !!! 3 !!! ! ! 2 โ != 3 ! 2 โ 3! = ∞ !!! Therefore, ๐ธ(๐ ! ) also doesn’t exist. d) Recall that the variance of ๐ is calculated as: ๐๐๐ ๐ = ๐ธ ๐ ! − ๐ธ ๐ ! However, neither ๐ธ ๐ nor ๐ธ(๐ ! ) exist (i.e., they cannot be calculated). Therefore, we cannot calculate the variance of ๐. Question 3 a) ๐ follows a Bernoulli distribution because it can take only two values: success (the individual would vote for FD) or failure (the individuals would not vote for FD). We typically indicate success with the value 1 and failure with the value 0. b) ๐ counts the number of successes from a Bernoulli distribution in a given number of tries (in this case, 100 tries). Therefore, ๐ follows a binomial distribution. The possible Page 2 of 4 values of ๐ are all integers between 0 and 100 (the number of people who would vote for FD lies between 0 and 100). c) The question tells us that ๐ follows a Bernouli distribution with parameter ๐ = 0.1. In this case, the probability function of ๐ is: ๐ 0 = ๐ ๐ = 0 = 1 − ๐ = 0.9 ๐ 1 = ๐ ๐ = 1 = ๐ = 0.1 d) Now we know that ๐ follows a binomial distribution with parameters ๐ = 100 and ๐ = 0.1. Therefore, its probability function is: ๐ ! 100 ๐ ๐ = ๐ 1 − ๐ !!! = โ 0.1! โ 0.9!""!! ๐ ๐ The probability that 15 respondents answer that they would vote for FD is: 100! 100 ๐ 15 = โ 0.1!" โ 0.9!""!!" = โ 0.1!" โ 0.9!" = 0.033 15 15! โ 100 − 15 ! In other words, there is a 3.3% probability that 15 respondents answer that they would vote for FD. Note that you could calculate this probability in Excel using the formula: = BINOM. DIST(15, 100, 0.1, FALSE) Question 4 a) In order to construct a 95% confidence interval, we first need to find the critical value we need to use. Note that ๐ผ = 1 − 0.95 = 0.05. Therefore, the critical value is: ๐ง!!!/! = ๐ง!!!.!"/! = ๐ง!.!"# = 1.96 Note also that the text of the question gives us the sample average ๐! = 0.10 for a sample with ๐! = 50 elements, as well as the sample standard deviation ๐! = 0.15. The 95% confidence interval can then be constructed as: ๐ ± ๐ง!!!/! โ ๐! = 0.10 ± 1.96 โ ๐ 0.15! = [0.058, 0.142] 100 b) We want to conduct a hypothesis test of the hypothesis that stock A outperforms investments in riskless government bonds (in other words, that the return on stock A is higher than the return on government bonds). The null and the alternative hypotheses are: ๐ป! : ๐! ≥ ๐! ๐ป! : ๐! ≥ 0.03 , or ๐ป! : ๐! < ๐! ๐ป! : ๐! < 0.03 To conduct this test, we first need to calculate the test statistic: ๐! − ๐! 0.10 − 0.03 ๐= = = 3.300 ! 0.15 ! ๐! 50 ๐! This is a one-sided test with a “less than” alternative hypothesis, so we would reject the null hypothesis if the value of the test statistic is lower than the critical value ๐ง! = ๐ง!.!" = −1.645 (you can find this critical value in Excel using the formula = NORM. S. INV(0.05)). Given that 3.300 > −1.645, we cannot reject the null hypothesis that stock A outperforms investments in riskless government bonds. The p-value of this hypothesis test is: ๐ = Φ ๐ = Φ 3.300 = 0.9995 Page 3 of 4 This value is much higher that the significance level of 5%, which again tells us that we cannot reject the null hypothesis. This can also be calculated in Excel using the formula: = NORM. S. DIST(3.300, TRUE) c) Now we are given information about a second random variable. The sample mean of this new variable is average ๐! = 0.07 for a sample with ๐! = 30 elements, with a sample standard deviation ๐! = 0.1. We want to test the hypothesis that the population rate of return on stock A is higher than the population rate of return on stock B. This is a test of two population means: ๐ป! : ๐! ≥ ๐! ๐ป! : ๐! < ๐! The test statistic for this test is: ๐! − ๐! 0.10 − 0.07 ๐= = = 1.072 ! ! 0.15 0.10 ! ! ๐! ๐! + 30 + 50 ๐! ๐! This is again a one-sided test with a “less than” alternative hypothesis, so we would reject the null hypothesis if the value of the test statistic is lower than the critical value ๐ง! = ๐ง!.!" = −2.326 (again, you can find this critical value in Excel using the formula = NORM. S. INV(0.01)). Given that 1.072 > −2.326, we cannot reject the null hypothesis that stock A has a higher rate of return than stock B. Although not required, we can also calculate the p-value of this hypothesis test: ๐ = Φ ๐ = Φ 1.072 = 0.8581 As before, this value is much higher than the significance level of 1%, telling us again that we cannot reject the null. As before, the p-value can be calculated in Excel using the formula: = NORM. S. DIST 1.072, TRUE d) This time we have a two-sided hypothesis test: ๐ป! : ๐! = ๐! ๐ป! : ๐! ≠ ๐! The test statistic for this hypothesis test is: ๐! − ๐! 0.07 − 0.03 ๐= = = 2.191 ! 0.10 ! ๐! 30 ๐! We would reject the null hypothesis if the absolute value of the test statistic is higher than the critical value ๐ง!!!/! = ๐ง!!!.!"/! = ๐ง!.!" = 1.645 (you can find this critical value in Excel using the formula = NORM. S. INV(0.95)). Given that 2.191 > 1.645, we reject the null hypothesis that stock B has the same rate of return as government bonds. The pvalue of this hypothesis test: ๐ = 2Φ −|๐| = 2Φ −2.191 = 0.028 which is lower than the significance level of 10%, again indicating that we reject the null hypothesis at a 10% significance level (note that we would not have been able to reject the null at a 1% significance level). Similar to the previous parts, the p-value is calculated in Excel as follows: = 2 ∗ NORM. S. DIST −ABS(2.191), TRUE Page 4 of 4