Write your name and student ID # on the back of the back page. BA275 / Winter 2006 / Exam #2 / 80 Points For multiple choice problems circle the letter corresponding to the BEST answer provided. For calculation problems, SHOW ALL WORK. Credit is given for the method, not the answer. For short answer questions, keep your answers concise, and to the point. 1. The weights of artichoke shipping crates have a mean of 40.0 pounds and a standard deviation of 2.80 pounds. 1a. (5 points) A random sample of 32 artichoke shipping crates is selected. What is the probability that the sample mean is greater than 41 points. Draw a picture of the probability distribution that you use to find the answer, and show the area that corresponds to your answer. Also, find the mean and the standard deviation of this probability distribution. Answer: X 40 2.8 X .4950 n 32 For a sample mean of 41: Z X X 41 40 2.02 X .4950 P( X 41) P( Z 2.02) .5 .4783 .0217 1b. (3 points) If the sample size had been 4, rather than 32, carefully explain why it would not possible to find probabilities for the sample mean. Answer: Because we don't know that the population is normally distributed, and because the sample size is not large enough for the Central Limit Theorem to apply, we do not know the shape of the sampling distribution of the mean. If we do not know the shape of the distribution, we cannot find probabilities. 2. (3 points) Which of the following describes a Type I error: a. We reject the null hypothesis, when the null hypothesis is true. ***correct answer b. We fail to reject the null hypothesis, when the null hypothesis is false. c. We reject the null hypothesis even though the P-Value is greater than α. d. We fail to reject the null hypothesis, even though the P-Value is less than α. 3. (3 points) Which of the following describes a Type II error: a. We reject the null hypothesis, when the null hypothesis is true. b. We fail to reject the null hypothesis, when the null hypothesis is false. ***correct answer c. We reject the null hypothesis even though the P-Value is greater than α. d. We fail to reject the null hypothesis, even though the P-Value is less than α. 4. The quality control manager at a light bulb factory needs to estimate the average life of a large shipment of light bulbs. It is known that the population standard deviation of bulb life is 75 hours. A random sample of 50 light bulbs from the shipment has a sample mean of 340 hours. 4a. (5 points) Find the 94% confidence interval for the average life of light bulbs in the large shipment. Answer: Since we know the population standard deviation (σ), we use a Z statistic: X Z n 75 340 1.88 340 19.94 hours 50 2 4b. (4 points) Use an equation (a probability statement) to explain what this confidence interval means. Answer: P(340 – 19.94 < < 340 + 19.94) = .94 4c. (4 points) Which of the following interpretations of this confidence interval is correct? a. 94% of the sampled light bulbs had lives within the confidence interval. b. We are 94% confident that the average life of all the light bulbs in the large shipment is within the confidence interval. ***correct answer c. Of all the light bulbs in the large shipment, 94% of them will have lives within the confidence interval. d. We are 94% certain that the average life of the sampled light bulbs is within the confidence interval. e. 6% of the light bulbs in the large shipment have lives that are outside the confidence interval. 5. (4 points) The purchasing director for an industrial parts factory is investigating the possibility of purchasing a new type of injection molding machine. He determines that the new machine will be bought if there is convincing evidence that the new machine produces a smaller proportion of defective parts that the old, existing machine. Samples of 200 parts are gathered from each machine. For the old machine the sample proportion of faulty parts is .040. For the new machine sample, the sample proportion of faulty parts is .036. Identify the hypotheses (null and alternative) that would allow the purchasing manager to make this decision. Answer: H0: pnew pold, HA: pnew < pold Alternatively, H0: pnew - pold 0, HA: pnew - pold < 0 6. An engineer is evaluating whether or not to upgrade a machine component from the current all-metal component, to a new, polymer component. Because the upgrade is expensive, the upgrade will be made only if there is convincing evidence that the new, polymer component has a longer, average service life. Extensive experience with the existing, all-metal component indicates that the average service life is 200 hours. The engineer sampled 46 polymer components, and found a sample mean service life of 215 hours and a sample standard deviation of 21 hours. 6a. (2 points) For an alternative hypothesis of: H1: µ > 200, state the null hypotheses. Answer: H0: μ ≤ 200 6b. (3 points) Using = 1.0%, identify the critical value(s) for this hypothesis test. Answer: The critical value of t ( = .01, dof = 45) = 2.412 6c. (3 points) Calculate the test statistic for this hypothesis test. Answer: Test statistic: t X 0 209 200 2.9067 S 21 46 n 6d. (5 points) Draw a picture of the t-distribution for the test statistic (assuming that the null hypothesis is true). Show and label the critical value(s), the test statistic, the rejection region, and the area corresponding to the P-Value. 6e. (3 points) In the context of the problem, state the conclusion(s) of the hypothesis test. Answer: Because the test statistic is greater than the critical value, we reject the null hypothesis and conclude the alternative hypothesis. “At a 5% level of significance, there is sufficient statistical evidence to conclude that the average service life of the new, plastic component is greater than 200 hours. 6f. (2 points) Carefully explain whether the P-Value for this hypothesis test is greater than or less than 1%? Answer: Since we rejected the null hypothesis, the P-Value must be less than alpha = 1%. 7. The following hypothesis test was carried out: H0: = 0.0, HA: ≠ 0.0 STATGRAPHICS provided the following as part of the analysis: Power Curve Power alpha = 0.1, sigma = 1.82 1 0.9 0.8 0.7 0.6 0.5 0.4 0.3 0.2 0.1 0 -1.2 -1 -0.8 -0.6 -0.4 -0.2 0 0.2 0.4 0.6 0.8 1 1.2 True Mean 7a. (3 points) If the true population mean = 0.20, the probability of making a Type II error is: a. .00 ≤ probability < .08 b. .08 ≤ probability < .18 c. .18 ≤ probability < .28 d. .28 ≤ probability < .38 e. .38 ≤ probability < .53 f. .53 ≤ probability ≤ 1.00 *** correct answer 7b. (3 points) If the true population mean = 0.20, the probability of making the correct decision is: a. .00 ≤ probability < .08 b. .08 ≤ probability < .18 c. .18 ≤ probability < .28 ***correct answer d. .28 ≤ probability < .38 e. .38 ≤ probability < .53 f. .53 ≤ probability ≤ 1.00 8. (5 points) The following hypothesis test was carried out: It is known that the population standard deviation (σ) = 15. A random sample of n = 100 had a sample mean of 88. Find the P-Value for this hypothesis test. H0: µ≥; µ Answer: Since we know the population standard deviation, we use a Z statistic: Test Statistic: Z X 0 88 90 1.3333 15 n 100 P-Value = P(Z < -1.333) = .5-.4082 = .0918 9. (5 points) A researcher calculated the following confidence interval for a population mean: [21.0, 29.0]. The confidence interval was based on a sample of size 45, a sample mean of 25.0, and a known population standard deviation of 10.417. What α did the researcher use to calculate the confidence interval? Answer: e Z 2 n ; 4.0 Z 10.417 ; Z 2.5758 2 2 45 α/2 = .005, so α = .01 10. An independent testing agency has been contracted to determine whether an experimental (new) formulation of gasoline improves fuel economy. The original (old) formulation is tested in 200 cars giving a sample mean of 18.5 mpg and a standard deviation of 4.6 mpg. The experimental formulation is tested in 100 cars, giving a sample mean of 19.34 mpg and a standard deviation of 5.2 mpg. The (partial) results of a hypothesis test ( = .05) are shown below: Hypothesis Tests ---------------Sample means = 18.5 and 19.34 Sample standard deviations = 4.6 and 5.2 Sample sizes = 200 and 100 Null Hypothesis: difference between means = 0.0 Alternative: less than Computed t statistic = -1.4266 P-Value = 0.0773714 (Equal variances assumed). 10a. (4 points) Restate the hypotheses in the context of the problem. In other words, restate the null and alternative hypotheses in terms of : µOLD: “average mpg using the old formulation”, and µNEW: “average mpg using the new formulation”. Answer: H 0 : OLD NEW 0 H1 : OLD NEW 0 OR H 0 : OLD NEW H1 : OLD NEW 10b. (4 points) Using an α of 5.0%, use the reported P-Value to state the conclusions of the hypothesis test in the context of the problem. Answer: Because the P-Value is greater than α = .05, we cannot reject the null hypothesis. "There is insufficient statistical evidence to conclude that the average mpg using the new experimental formulation is greater than the average mpg using the old, original formulation. 10c. (3 points) If the average mpg is exactly the same with either formulation, what is the probability that we would get a sample result where: X OLD X NEW 0.84 Answer: This is just the P-Value: P-Value = 0.0773714 11. (4 points) A mutual fund manager would like to investigate whether companies that are “employee friendly” provide higher stock returns. She identifies 50 "employee friendly" firms, and measures the 5year stock return for each. For each of the 50 "employee friendly" firms, she then identifies a "nonemployee friendly" industry peer (a similar company in the same industry), and measures the 5-year stock return for each of those firms. She then carries out the following hypothesis test: H0: µ1 - µ2≤ µ1 - µ2 where µ1 : the average 5-year stock return of employee friendly firms, µ2the average 5-year stock return of non-employee friendly firms Using the same data (5-year stock returns for the same 100 firms) briefly describe an alternative approach to this hypothesis test that would provide a more powerful test. Answer: This is an example of a “related samples” situation. This allows us to define a new variable such as: DIFFERENCE = return of employee friendly firm – return of non-employee friendly industry peer This would give us 50 observations for the variable DIFFERENCE. We would then just do a one sample hypothesis test such as: H0: µDIFFERENCE ≤0.0 H1: µDIFFERENCE > 0.0 The rejection of the null hypothesis would allow the same conclusion as the other approach, but this is a more powerful test… when the null hypothesis is false, it is more likely that we would reject the null hypothesis in the related samples approach.