FINAL PRACTICE EXAM P (SOURCE: PRE AUT 07) Last updated 22nd October 2010

Cover Type B

TO BE RETURNED AT THE END OF THE EXAMINATION. THIS PAPER MUST NOT BE REMOVED FROM THE EXAM CENTRE.
SURNAME: FIRST NAME: __________________________________ __________________________________

STUDENT NUMBER: __________________________________ COURSE: __________________________________

__________________________________________________________________________________

FINAL EXAMINATION PRACTICE PAPER ‘P’ SOURCE: Past practice paper written Pre-Autumn 2007

SUBJECT NAME and NUMBER:

BUSINESS INFORMATION ANALYSIS 26133 BUSINESS STATISTICS 26134 3 Hours plus 10 minutes reading time

TIME ALLOWED:

INSTRUCTIONS TO CANDIDATES: • • This question paper MUST NOT be removed from the Examination Centre This is an OPEN BOOK EXAMINATION (any materials allowed including Lecture Notes) Calculators (including Programmable Calculators) are allowed Record your student name and number carefully on the multiple choice answer sheet provided Answer all questions on the multiple choice answer sheet provided Use the exam question booklet for any working. All exam question booklets will also be collected and MUST NOT be removed from the Examination Centre


• • •

FINAL PRACTICE EXAM P (SOURCE: PRE AUT 07) Last updated 22nd October 2010 Question P1: The owner of Fortee Bakery is interested in determining the difference between mean purchase amounts per customer at two locations. To estimate the difference, samples of 500 customer receipts from each location are examined. The following sample statistics are produced: Location 1 Location 2 Mean $5.26 $5.66 Standard Deviation $0.89 $1.05 The owner asks you to examine if there is a significant difference in the average purchase amounts per customer at the two locations. You determine that: (a) Using an independent samples t-test, the mean purchase amounts at each location are not significantly different at the 99% level of significance (b) Using an independent samples t-test, the mean purchase amounts at each location are significantly different at the 99% level of significance (c) Using a paired samples t-test, the mean purchase amounts at each location are not significantly different at the 99% level of significance (d) Using a paired samples t-test, the mean purchase amounts at each location are significantly different at the 99% level of significance Answer: B = independent; sales significantly different Justification: Ho: μ1 – μ2 = 0; Ha: μ1 – μ2 ≠ 0 Test method: independent samples b/c only have aggregate information about each store. No way of matching the two store sales at the individual observation level. Standard error: sqrt(s1^2/n1 + s2^2/n2) = sqrt(.89^2/500 + 1.05^2/500) = .061556 99% confidence means that α=1%. Since two-tailed test area in upper tail is α/2 = ½% = .005. There are n1=500 receipts and n2 = 500 receipts. Df = 500+500 – 2 = 998 d.f. We use t-distribution since σ is unknown. T(.005,998) = T(.005,Inf) = 2.576 (table available in exam) Critical value = 0 + T*SE = $0.158569 Evidence(Point estimate): xbar1 – xbar2 = 5.26 – 5.66 = -$0.40 Test-statistic = $0.40 / SE = -6.49 Since both the evidence in dollars and test-statistic are in the rejection region, reject Ho. Conclude that the sales at the two locations are significantly different at the 99% level. Question P2: As part of an internal auditing process, a firm wishes to estimate the mean proportion of its credit card holders having accounts that are overdue. The auditors have stipulated that the minimum margin of error must be no greater than 3% and they wish to use a 95% confidence interval estimate. What size sample should they select if they have no idea the percentage of accounts that are overdue but still wish the sample size to be large enough to meet their requirements? (a) 17 (b) 30 (c) 204 (d) 1068 Answer: D = 1068 Justification: n = (critical value)^2(sigma)^2 / (margin of error)^2 NOT n = (critical value)^2(standard error)^2 / (margin of error)^2 n = z^2 * p(1-p) / E^2. Since you do not know what p is, you maximise p(1-p), so assume p=0.5 = 1.96^2 * .5 (1-.5) / .03^2 = 1067.0719

Page 2 of 20

twenty list a telephone number first (coded as two) and thirty list neither (coded as zero). X2 = 1 (celebrity appears). You cannot order the outcomes {web first .685 (c) $61.9287 0.963148 where X1 = the number of pages in the issue. Ordinal and Interval (d) Nominal. neither listed} could be written equivalently: { telphone first.P(BC) given independence.97 + 61650 + 35.97 + 2055*30 + 35. 70% of people use Artsel One as their carrier of choice. You predict the level of sales for an issue in which there are 30 pages and a well known celebrity appears on the cover to be: (a) $61.236 = 6.12 Question P4: A listing of advertisements indicates that five advertisements choose to list a website first (coded as one).000 Answer: C = $61. P(AC) = 1 – P(A) = 1 .236(X2) X1 = 30.24 = 61692. X2 = a dummy variable taking the value 1 if a well known celebrity is shown on the front cover and 0 otherwise. web first}. What is the expected probability a person does not have broadband connection and choose a different carrier other than Artsel One.97 2055 35.235 t Stat 11. telphone first.21 Page 3 of 20 .046227 P-value 5.30 = Probability of not using Artsel One C P(B ) = 1 – P(B) = 1 – .692 Justification: Sales = 6.846893 0. Question P5: The financial manager of a magazine has compiled the following table from a regression analysis used to make predictions about the number of sales (in dollars) per issue: Intercept X1 X2 Coefficients 6. Interval and Ratio Answer: A = Nominal only Justification: You can only nominate a response as either having the listing.5 762.FINAL PRACTICE EXAM P (SOURCE: PRE AUT 07) Last updated 22nd October 2010 Question P3: 60% of people have broadband connection.236*1 = 6..30 x .97 + 2055(X1) + 35.70 = . would be considered to have at best (in terms of ability to conduct statistical analysis) which measurement properties: (a) Nominal only (b) Nominal and Ordinal (c) Nominal.60 = .19318 0. Ordinal.650 (b) $61.97 + 61685.40 = .692 (d) More than $62.584 10637. assuming that the choice to have broadband is independent of the choice of carrier and visa versa? (a) 12% (b) 18% (c) 42% (d) 58% Answer: A = 12% Justification: P(AC and BC) = P(AC). Sales = 6. neither listed. The variable. P(AC and BC) = .236 Standard Error . type of listing in advertisement.40 = Probability of not having broadband.22E-29 0.

s(xbar) = SE = σ / sqrt(n) = 10 / sqrt(100) = 10 / 10 = 1kg mean = 17.8414 (d) .P(z < (15.7kg < xbar < 19. Question P8: The weight of goods being transported by an airline for each passenger is observed. and. respectively. a dummy variable representing whether Cracker Nuts had a promotional offer or not.FINAL PRACTICE EXAM P (SOURCE: PRE AUT 07) Last updated 22nd October 2010 Question P6: A pricing study examines the linear relationship between sales of Cracker Nuts and factors such as its price.2kg of the population mean weight? (a) .7 – 17. The population of weights for all passengers is known to follow a normal distribution with a standard deviation of 10kg. Hence.1586 (c) .7) – P(xbar < 15. 3.7kg) = P(xbar < 19. Based on this information you would suspect that the major concern is: (a) non-linear effects (b) multi-collinearity (c) outliers (d) too many variables Answer: B = Multi-collinearity Justification: Since the two companies both run promotional offers at the same time. Which statement is correct: (a) On average. this is mult-collinearity. median and mean on the data revealing the values 2. panel members appear to be employed in retailing (b) Most panel members appear to be employed in retailing (c) 50% of members appear to be employed in retailing (d) None of the above statements are correct Answer: D = None of the above Justification: The data has nominal properties so only the mode is interpretable. The coding scheme used is: 1 = telecommunications 2 = finance and banking 3 = retailing 4 = other industry You calculate the mode.9544 Answer: D = . its competitor's prices.7)/SE) . had a promotional offer or not.7kg. Question P7: An Excel spreadsheet displays a variable that is coded with values representing the area in which an online panel member is employed.9544 Justification: n = 100. which is finance and banking. the dummy variables will be correlated.0456 (b) . The regression output is very peculiar.7) = P(z < (19. The data revealed that Cracker Nuts runs promotional offers only when Sirius Nuts happened to be running a promotional offer as well.7 = P(15.7 – 17. The mode tells us that most people are employed in industry category 2. a dummy variable representing whether its closest competitor. Sirius Nuts.7)/SE) = P(z<+2/1) – P(z<-2/1) Page 4 of 20 . and 3. What is the probability that the sample mean of weights observed is within +/. N = unknown (infinite). s = 10. A sample of one hundred passengers reveals an average weight of goods to be 17.

mean emissions not significantly higher than 2.9g/mi Justification: Ho: μ ≤ 2. one cannot make any of the above conclusions Answer: B = On average.. Rejection region/critical value: tdf=10-1=9. namely μ ≥ 40 minutes.05 = 1.05 = 1.833). the emissions are significantly higher than the 2. we can use this.1g/mi claimed (d) With only ten vehicles tested.58) < (t9. Another claim is the compliment of this.4g/mi. AGI Sales believe the emissions.7 .9g/mi (status quo) Ha: μ > 2. we cannot reject the null hypothesis.1g/mi with a standard deviation of 0.9g/mi. with a mean of 2. some students may use the following: P(xbar – j ≤ E(xbar) ≤ xbar + j) = 2*P(z ≤ j / s(xbar)) – 1. The mean emission are not significantly higher than the 2.05) of AGI Sales and conclude: (a) On average.9g until we find evidence to lend support for a contradictory view. Test statistic = (xbar – m)/(s/sqrt(n)) = (3.P(z<+2)] = 2*P(z<+2) – 1 = 2*.9g/mi claimed (c) On average... since the population standard deviation is unknown (but we know from sample that s = . Sampling ten of their vehicles.9g/mi (claim) Because the sample size is small (n=10) but the population parameter follows a normal distribution. on average.4/sqrt(10)) = 1. this will become the null.1 – 2. Question P10: Past studies show that the previous mean time to prepare a home-cooked meal was 40 minutes. may be significantly higher than that suggested by their supplier. What should the null hypothesis be to test the claims made in the new study? (a) μ ≥ 40 minutes (b) μ > 40 minutes (c) μ < 40 minutes (d) μ ≤ 40 minutes Answer: A = mu >= 40 minutes. the test statistic can be used and follows a t-distribution.4g/mi). you test the concerns (using alpha = . Also. the emissions are not significantly higher than the 3. Using the sample. Question P9: AGI Sales are informed from their car supplier that carbon monoxide (CO) emissions for a certain kind of car follow a normal distribution.9g claimed. j = 2kg.1g. AGI Sales finds that.9772 – 1 = . Justification: One claim or hypothesis is that μ < 40 minutes (significantly lower than 40 minutes). That is: Ho: μ ≥ 40.9544 Note: depending on how they have been taught. Suppose that a study is designed to test this by sampling 100 home-owners. the emissions are 3. As this contains the equals.833 (it is a one-tailed test) From sample observations: xbar = 3.581139 Since (test stat =1. A new study claims that the mean time to prepare a home-cooked meal has dropped to be significantly lower than this amount.FINAL PRACTICE EXAM P (SOURCE: PRE AUT 07) Last updated 22nd October 2010 = P(z<+2) – [1 . on average.9)/(. where E(xbar) = 17.9g/mi claimed (b) On average. We currently assume mu = 2. Page 5 of 20 . the emissions are not significantly higher than the 2.

it is the number of pairs that form the number of observations).09 Round up to 217. How large a sample of assembly times should be taken to estimate the mean assembly time to within 2 seconds. Ha: μd ≠ 0. It is believed that the population standard deviation is 15 seconds.2–0)/(8.228 (Note. Hypothesised mean = 0 table critical value = ta/2. The sample mean difference was found to be 4200 units with a standard deviation of 8800 units sold.11-1 = t.8 test statistic = t = (dbar – mud)/(sd/sqrt(n)) t = (4. a measure of central tendency.583 Since t = 1.2.05/2. Page 6 of 20 .10 = 2. is sensitive to outliers because it relies on which measurement property of the data to be calculated? (a) Integer (b) Nominal (c) Ordinal (d) Quantitative Answer: D = Quantitative Justification: The mean relies on the numerical value or quantitative properties of the data b/c it sums the data.8/sqrt(11)) = 1. Ho: μd = 0.025. with 95% confidence? (a) 153 (b) 216 (c) 217 (d) 865 Answer: C = 217 Justification: n = (critical value)^2(sigma)^2 / (margin of error)^2 NOT n = (critical value)^2(standard error)^2 / (margin of error)^2 1. size and product sales. Question P12: The mean. These eleven pairs of supermarkets (22 in total) were selected that were similar in terms of geographic location. observed: dbar = 4. (with α=. Justification: Hence. this time promoting the product on the basis of its energy producing benefits. Examining this evidence. sd = 8.05) you can conclude: (a) The average number of units sold was not significantly different across the two sets of stores (b) The average number of units sold was significantly different across the two sets of stores (c) A conclusion about differences in average units sold cannot be made since we do not know how each supermarket performed (d) People appeared to really enjoy the promotion involving low fat content Answer: A = no significant difference in number of units sold.n-1 = t. Another eleven stores were identified.96^2 * 15^2 / 2^2 = 216. The difference in units sold was calculated for each pair.FINAL PRACTICE EXAM P (SOURCE: PRE AUT 07) Last updated 22nd October 2010 Question P11: A production manager of Lotzafun Toys needs to estimate the average time taken to assemble products using a new manufacturing technique.583 < t-table we cannot reject Ho at the 95% level. Question P13: Eleven supermarkets introduced a promotional display for Grand Baked Beans promoting the product on the basis of its low fat content.

39093 0.05. better than chance. the sample revealing an average premium to be $550.43666 MS 3.021768 0.90018) – 1 P(548. The following formula was provided to students in previous semesters to save time.905 Question P15: A real estate agent believes a regression will be useful to predict auction prices by including various factors.084869 F 44.53E-17 Using the F-statistic (with 95% level of significance) provided in the ANOVA table only.5 / .537164 8.21E-38 Coefficients 60.FINAL PRACTICE EXAM P (SOURCE: PRE AUT 07) Last updated 22nd October 2010 Question P14: The premium amounts of all 2500 insurance payers (i.021616 0.74425 64.738482 0.50kg.5% Answer: D = 90. What is the probability that this average premium is within +/.5% (b) 11. (d) None of the above can be established using the statistics listed in the ANOVA table Answer: D = none of the above are correct.948683) See question 8 solution for the full step-by-step theoretical approach..23617 0. An internal audit selects a sample of 250 premiums.1 which is NOT <= . s(xbar) = [s / sqrt(n)][sqrt(N-n) / sqrt(N-1)] s(xbar) = [15 / sqrt(250)][sqrt(2500-250)/sqrt(2500-1)] = [. hence finite correction factor reqd.12484 3. which statement is correct: (a) Both distance to shops and the number of bedrooms in a dwelling are significant in predicting auction sale prices.4% (c) 88. j = 1.5% Justification: n = 250.021086 0. the population) follows a normal distribution with a standard deviation of $15.948873]= .90018 (NOT .5 ≤ E(xbar) ≤ 550+ 1.068209 -0.50 of the population mean premium amount? (Hint: consider the issue of finite correction) (a) 9. better than chance. E(xbar) = 550 . P(550– 1.001692 0.e.8495 1. but not sure which one (Ha). hence. better than chance.948683][.012512 0.05016 Significance F 4.6% (d) 90. The following regression output is produced: ANOVA Regression Residual Total Estimates Intercept Square Metres Distance to Schools Distance to Shops Bathrooms Bedrooms df 5 539 544 SS 18.$1.769 3. (b) Only distance to shops is significant in predicting auction sale prices.50) = 2*P(z ≤ 1.5) = 2*P(z ≤ 1.033752 0. Page 7 of 20 .11E-25 0.017512 6.021957 0. but you do not have to learn this: P(xbar – j ≤ E(xbar) ≤ xbar + j) = 2*P(z ≤ j / s(xbar)) – 1. (c) Only the number of bedrooms in a dwelling is significant in predicting auction sale prices. N = 2500 (finite).155526 -2. Justification: F-statistic in ANOVA only tells us whether none (Ho) or at least one of the mean coefficients are significantly different from zero.38312 -10. s = 15.50≤ E(xbar) ≤ 551.05025 -0.187133 Standard Error 0.69241 45.9525 – 1 = .666333 ) – 1 = 2*.021467 t Stat 4826. n/N = 250 / 2500 = .71703 P-value 0 0.

(d) The soles are significantly different in terms of the amount of variability exhibited in wear and tear. Page 8 of 20 .1954 213. Hence we do not rewind. Answer: The soles are not significantly different in terms of mean wear and tear.FINAL PRACTICE EXAM P (SOURCE: PRE AUT 07) Last updated 22nd October 2010 Question P16: A town planner obtains a sample listing the actual heights of buildings (in metres) in the local central business district.375494 What would you conclude at the α=. but since we know the assumption of independence does not hold. Ha: At least one of the means is significantly different. Using p-value: p-val = . Durability is assessed based on wear and tear.05.782 533420. Justification: Ho: μ1 = μ2 = … = μ5 = μ*.05 level based on the Analysis of Variance (ANOVA) output: (a) The soles are not significantly different in terms of mean wear and tear. Ho refers to the equality of the means not the variances. what is the observed conditional probability that the town planner selects a building that is built too tall.311 > a = . P(T and R) are unobserved.6 df 4 2495 2499 MS 255.7959 F 1. (c) The soles are not significantly different in terms of the amount of variability exhibited in wear and tear.20 P(River) = 0.19364 P-value 0. However. (a) 6% (b) 20% (c) 30% (d) Unable to be determined Answer: B = 20% Justification: P(Too Tall) = 0. The following results were reported testing the null hypothesis that mean wear and tear for all five soles is equal. given it is built within a close proximity to the river? Assume that the excessive height of a building relative to its approved height is independent of its proximity to the river. 30% of buildings fall into the category of being in close proximity to the river. P(TT | R ) = P(TT)*P(R) / P(R) = P(TT) = . The town planner determines that 20% of buildings are above the actual height approved by the town planning committee. P(TT and R) = P(TT)*P(R) under independence. ANOVA Source of Variation Between Groups Within Groups Total SS 1020.8 534441. where higher ratings indicate greater wear and tear. The town planner also categorises buildings as being within a close proximity (less than 2km) to the river running through the city. Randomly selecting one building for further investigation. (b) The soles are significantly different in terms of mean wear and tear. the value can be determined.30 P(TT | R ) = P(TT and R) / P(R). Question P17: A sports manufacturer tests the durability of five different soles. we do not rej Ho.311512 F crit 2.20 = 20% Note that under independence we’ve shown that P(TT | R) = P(TT) – this should make theoretical sense to students: we are saying that the height of a building is not conditional upon whether it is near the river or not.

expertise (4).0428 0.5602 P-value 0.4685 INDEX current 1 3 4 2 6 2 5 39.0461 -0. the charity should consider that the following characteristics are significant (at the α=.1156 0. except income (d) All of the variables Answer: B = Income only.05 level) in explaining donation behaviour: (a) None of the variables (b) Income only (c) All of the variables.7331 1.6032 0.6426 improve 1 5 4 2 6 6 5 46.0000 0.5432 2.05 level since p-values are all > .0915 t Stat 9. trustworthiness (6).0864 5.7238 3.0882 0.6029 2. we cannot rej Ho.6139 4.0032 Edyr Hospital has the following ratings: empathy (3).0457 0. The dependent variable results in an index of performance. When testing at α=.5202 -1.2562 0.05 hence rewind. What will be the impact on their overall index of performance as a result of their endeavour: (a) The index will improve by 5 points. where 1=Poor and 7=Excellent. all else constant.018 Question P19: A charity was attempting to determine if the number of donations being made to its foundation (dependent variable) was somehow related to the characteristics of donors.0172 0. Factor Intercept Empathy of staff Expertise of medical staff Administrative efficiency Trustworthiness of staff Cleanliness Quality of facilities Coefficients 2.5840 1.6050 0. all else constant.0001 0.4756 0.1427 Standard Error 5.018 impact change 1.3301 13. Page 9 of 20 . Justification: Need only look at the change only – as seen in the last two columns: Factor Intercept Empathy of staff Expertise of medical staff Administrative efficiency Trustworthiness of staff Cleanliness Quality of facilities Coefficients 2.05.7331 1. administration (2). hoping to obtain ratings of 5 and 6 respectively. (b) The index will improve by 6 points. (d) The index will improve by 8 points.0010 -0.FINAL PRACTICE EXAM P (SOURCE: PRE AUT 07) Last updated 22nd October 2010 Question P18: The regression output below represents the perceptions of private hospitals.0000 0.1794 0. ONLY Income is significant since . all else constant. All other variables are insignificant at the a=.0000 0.9316 7. measured on a scale of 1 to 7.4685 Standard Error 0.8181 t Stat 53. all else constant.05. rej Ho that bi=0.2917 0.1816 0. cleanliness (2) and quality (5). hence we cannot rewind.4829 2. Intercept Age Income Number of Children Sole Parent (dummy coded) Coefficients 49.0871 6. deflating the standard error.3912 -0. Answer: C = The index will improve by 7 points.6606 change 0 2 0 0 0 4 0 7.1421 14.0000 0.7431 -0.2917 0. hence.0000 0.0172 P-value 0.1193 Examining this regression output above.0886 0.5176 2.2205 0.2058 0. (c) The index will improve by 7 points.0004 0.8014 0. Suppose the Edyr Hospital aims to improve itself on perceptions regarding empathy and cleanliness.8014 0.8416 -0.4829 2.0172 < .5432 2.6029 2. all else constant. Justification: While income has the lowest coefficient this is likely b/c of the way in which it measured.

77% (d) 69.P(M|A) / [ P(A).FINAL PRACTICE EXAM P (SOURCE: PRE AUT 07) Last updated 22nd October 2010 Question P20: An investment company. ABI.05*. The survey revealed that only 5% of people invest in mining. the firm wishes to determine the likelihood a person will choose ABI Investments.P(M|A) + P(AC). Students in some semesters will not be exposed to Bayes Theorem so would not be expected to see such a difficult question on their final exams.10)(.90 A recent industry survey reveals that 20% of people invest in mining. P(A) = . P(M | AC) = .20) / [(. P(A|M) = P(A and M) / P(M) = 0.20) + (.05) ] = . the firm wishes to determine the likelihood a person will choose ABI Investments.065 = .10)(.02 + . given they choose ABI to manage this portfolio. rearranging gives P(M and A) = P(M|A)*P(A) = . To assess the potential attraction of people to ABI because of its successful portfolio management in mining investments.10 = 0.20 The survey revealed that only 5% of people invest in mining. P(A | M) = ? This probability is found to be: Using Bayes Theorem … P(A | M) = P(A). given they used an alternative firm to manage their portfolio. ABI.5% (c) 30.02 + .065 Finally. Whilst one can avoid using Bayes Theorem to come up with an answer it is useful to see its role. An investment company. rearranging gives P(M and AC) = P(M|AC)*P(AC) = .02 P(M | AC) = P(M and AC)/P(AC).7692 Justification: NOTE: The following solution utilises Bayes theorem.045 = 0.30792 Page 10 of 20 .P(M | AC) ] = (.90 = 0. given they have chosen to invest in mining.307692 Even without knowledge of Bayes Theorem. given they have chosen to invest in mining. determines from its market share that 10% of investors choose ABI to manage their portfolios.90)(.05 To assess the potential attraction of people to ABI because of its successful portfolio management in mining investments.23% Answer: C = 30. P(M | A) = . given they choose ABI to manage this portfolio. one can see that: P(M | A) = P(M and A)/P(A). determines from its market share that 10% of investors choose ABI to manage their portfolios. A recent industry survey reveals that 20% of people invest in mining.02 / (. You determine this probability to be: (a) 2% (b) 4. P(AC) = .10 . given they used an alternative firm to manage their portfolio.02 / 0.20*.045 P(M) = P(M | A) + P(M | AC) = .045) = . Please do not ask about this on the discussion board.

the average (expected) number of visits to a National Park that a visitor will make will be? (a) No visits per year (b) 0.4 seconds.58659746** SAMPLE MEAN Var 4171.33 visits per year.09 Total sum x. respectively.428571 Stdevp 60. 70.33 visits per year (d) 1 visit per year Answer: C = .1 . 1. Hence.09 = .6 seconds. 220 The average and standard deviation of check-out waiting time for this sample is: (a) 85 seconds and 60. 40.6 Justification: Must be sample mean and standard deviation N 8 Average 85** Stdev 64.4 seconds.41522987 Varp 3650 It relies on the numerical value or quantitative properties of the data b/c it sums the data.8 0 1 .1 2 . 120.14 3 .14 + .03 .2. Justification: Looking only at the none.1 + . 80. (d) 85 seconds and 3650 seconds. respectively. respectively. the probabilities are: X p(x) x. 20. 100.p(x) 0 .p(x) = 0 + .14 visits per year (c) 0. (b) 85 seconds and 64. respectively. (c) 85 seconds and 4171. 30. Answer: B = 64. Page 11 of 20 .07 .3 only examines 1000 visitors.33 Question P22: A sub-set of responses from an observational survey of people at a supermarket reveals that each person spent the following amount of time (in seconds) waiting at the check-out.FINAL PRACTICE EXAM P (SOURCE: PRE AUT 07) Last updated 22nd October 2010 Question P21: A survey reveals a breakdown for the number of visits households had made to a National Park in the previous year: Number of Visits # Respondents None 800 1 only 100 2 only 70 3 only 30 4 only 0 Ignoring outcomes relating to those who visit five times a year or more.

FINAL PRACTICE EXAM P (SOURCE: PRE AUT 07) Last updated 22nd October 2010 Question P23: Based on historical attendance records.932338 0.. to run a regression with your weekly departmental expenditure for several years against various tasks that have been completed but with no direct expenditure amount. Caterer rates increase substantially when numbers exceed 60 people. Page 12 of 20 . The adjusted R-squared can either go up or down.285714.932054 Adjusted R Square 0.286329 Observations 495 495 You are immediately suspicious of Cindy’s report because: (a) Multiple R-squared went down (b) The adjusted R-squared went down (c) Both (a) and (b) (d) The number of observations did not change (e) Cindy is a psychopath and shouldn’t be trusted Answer: A = Multiple R-squared went down.931778 Standard Error 0. Regression 1 Regression 2 Multiple R 0. Question P25: You have asked your administrative assistant. Cindy runs a second regression in which she included some more variables that she had initially forgotten to include.965577 0.8413 = . Using last months stock listings. the probability that a stereo will be sold this upcoming month is closest too: (a) 29% (b) 40% (c) 60% (d) 71% Answer: A = 29% Justification: 20 stereos + 50 DVD players = 70 stock items. Hence. P (stereo ) = 20 / 70 = . What is the probability that this will occur? (a) 16% (b) 20% (c) 34% (d) 84% Answer: A = 16% Justification: P(x>60) = P(z > (60-50)/10) = P(z>10/10) = P(z>1) = 1 – P(z<=1) = 1 .931925 0.286021 0. Cindy runs regression 1 and reports the first part of the results in the table below.96543 R Square 0. Cindy. Justification: If you add additional items to a regression the R-squared should go up not down. an assessment of the number of people attending an IKG business seminar is normally distributed with a mean of 50 people and standard deviation of 10 people.1587 Question P24: A listing of stock reveals that twenty stereos were sold last month while a total of fifty DVD players were sold.

Which method would be the most appropriate summary technique to describe the data and achieve the manager's objectives? (a) Correlation Coefficient (b) Frequency Histogram (c) Ogive (d) Scatter-plot Answer: Ogive Justification: Correlation coefficient – numerical measure examines relationship between TWO QUANTITATIVE variables. Answer: B = positive. Hence. negative and positive. and positive. She is interested in predicting the movement in general inflation indicators such as the cost of a loaf of bread or a litre of milk. If Y is the average price in the consumer market Lisa is examining and she estimates the following regression function: Y = b0 + b1X1 + b2X2 + b3X3 + error.2 P(x<=7) = (7-5)*. b1 is positive. X3 = importation taxes. negative and negative. Also. She argues that any increase in importation taxes will see the supply of produce decrease and so prices impacted by such taxation increases will skyrocket. A survey asks employees how long they have taken that day and their answers are recorded in minutes. positive. (d) negative. the supply of goods will decrease. Question P28: Brett. negative. Lisa formalises her theory into a regression framework. b=10. 1 / (b-a) = 1 / (10-5) = 1/5 = . (c) positive. cost per barrel. b2 and b3 respectively are: (a) positive.FINAL PRACTICE EXAM P (SOURCE: PRE AUT 07) Last updated 22nd October 2010 Question P26: Lisa is a University researcher examining the impact of several variables and their effects on average prices in consumer markets. forcing prices up. negative effect. a small business owner of Drywall Plumbing has been told that his investment funds will take between 5 and 10 days to transfer to his day-to-day account. Question P27: A manager wishes to quickly assess the percentage of employees who take up to a certain amount of time to travel to work. negative and positive. b2 should be negative. As importation taxes increase.2 = .examines the cumulative percentage over and above a given level of x Scatterplot – graphical measure examines relationship between TWO QUANTITATIVE variables. He has a bill due in 7 days. interest rate rises will see average prices fall. Frequency histogram – shows frequency of qual or quant data – is a TABLE. What is the probability that the funds will be available in time to pay the bill? (a) 30% (b) 40% (c) 60% (d) 70% Answer: B = 40% Justification: Using uniform since we have no other information a = 5. Ogive .2 = 2*. b3 is positive. and positive (b) positive. Hence. negative. positive. Answer in summary is positive. then the expected signs of b1. Justification: higher [oil] prices will drive average prices up: hence. Lisa hypothesises that average prices are determined by oil prices: higher oil prices will drive average prices up. X2 = interest rates. she believes that interest rate rises will see average prices fall in most consumer markets. examining the following variables: X1 = oil.4 Page 13 of 20 .

(a) 24 kg (b) 26 kg (c) 31 kg (d) 37 kg Answer: B = 26 kg Justification: Confidence Interval = point estimate +/. (d) It cannot be determined because we don't know how many cars she observed were actually speeding.(1. with a standard deviation of 6kg determined from the same sample.071068 = .28 (c) 2. Question P31: A sample consisting of fifty observations is taken to examine the quality of a production line. = 1.(critical value)(standard error) where point estimate = sample mean = 25 kg.2828 Justification: n = 50. She over predicts on Thursday the number of cars speeding by 2 cars.41 the upper limit. critical value = sourced from t-distribution as σ unknown: 90% implies 0. N = infinite (unknown) Hence.41 kg. s(xbar) = s / sqrt(n) = 2 / sqrt(50) = 2 / 7. = 6/7.676 (based on t-tables so some element of inaccuracy).07 Answer: B = .59 and 26.840168) = 25 +/. s = 2.07 = .676)(.408122 = Approx 23. standard error = standard error of sample mean = s / sqrt(n) = 6 / sqrt(51). On Friday she under predicts the number of speeding cars by 4 cars. Using the philosophy of regression. With 26.04 (b) 0. What is the standard error of the sample mean quality rating? (a) 0. The population of quality is known to follow a normal distribution. (e) Karen is always speeding and should slow down Answer: B = Friday Justification: Friday – it doesn't matter whether we are over or underpredicting … what matters is by how much which is the residual squared or residual squared error! We don't need to know what the value was observed nor predicted to work this out. Constructing a 90% confidence interval estimate of the mean 12-month weight gain for all patients. the upper limit of your confidence interval will be approximately.840168 Hence 90% confidence interval is: 25 +/.2828 Page 14 of 20 . although not without some error. The sample reveals a mean quality rating of 95 and standard deviation of 2. Question P30: Karen found that she could predict whether a car was speeding from the make and model of the vehicle.FINAL PRACTICE EXAM P (SOURCE: PRE AUT 07) Last updated 22nd October 2010 Question P29: Suppose the weight gain over a 12 month period that is induced as a side-effect of a new drug for treating patients is normally distributed. a sample of 51 patients was drawn and the sample mean was found to be 25 kilograms gained over 12 months.00 (d) 7. which day has been associated with more error in making predictions? (a) Thursday (b) Friday (c) It cannot be determined because we don't know how many cars she predicted would be speeding.05 in upper tail area with n1 = 51-1 = 50 df.1. To estimate the mean weight gain.

Your lecture notes will guide you as to whether this topic is assessable or not. there is only a need to run a one off survey. Each email has a response asking them if they would come to the seminar series if organised. The solicitor sends an email to a random selection of thirty executive clients to gauge interest in the potential seminar series.FINAL PRACTICE EXAM P (SOURCE: PRE AUT 07) Last updated 22nd October 2010 Question P32: A solicitor wishes to assess the likely interest in running a series of seminars on property law in the upcoming month for the set of executive clientele listed who used the solicitor's services in the last year. The research stops people at random as they walk through the centre. on average. It is not a cluster sample b/c this is a type of random sampling. Only a random selection of clients was used: hence this represents a sample Question P33: A concreting company wishes to compare additives used in “batches” of concreting. Hence. (c) the difference between two population proportions using independent samples (d) the difference between two population proportions using a matched samples approach Answer: A = the difference between means of two populations.9 hours and 1.5 hours. The sampling method being used is best described as: (a) cluster sampling (b) convenience sampling (c) random sampling (d) justified sampling Answer: B = Convenience Sampling Justification: Please note that in previous semesters students have been assessed on this topic. and assume independent samples (b) the difference between means of two populations using matched samples approach. The student is simply obtaining responses from the residents who may be most conveniently located in the vicinity. a researcher now must advise which additive to use. This is an example of a: (a) Cross-sectional survey using the population (b) Cross-sectional survey using a sample (c) Time-series survey using the population (d) Time-series survey using a sample Answer: B = Cross-sectional survey using a sample Justification: Cross-sectional … while the seminars may be a series. Using a new style additive in 140 batches also. The sample is NOT random b/c we do not have a list of residents and drawing randomly from this. The old style additive is used in 140 batches and reveals that. the concrete takes hold with an average and standard deviation setting time of 15. respectively. Using only this information given. There are two sets of machinery each running different batches.8 hours. and which additive “is best”. on average. one with an old style additive and one with a new style additive. the company wishes to assess how long each additive sets. The researcher is best advised to consider testing: (a) the difference between means of two populations. There is no such thing as "justified sampling". Page 15 of 20 . concrete would set in 17. No clear information on how one would do this. Cannot create “pairs” of observations even when have same number of observations. forced to use independent samples Question P34: A student researching attitudes of residents to a new building proposal for a local shopping centre decides to visit the existing shopping centre.2 hours with a standard deviation of 2. so please do not ask on the UTSOnline discussion board if it is or is not in a given semester. assume independent samples Justification: Comparing average settting time – continuous variable – looking at MEAN. In particular.

753. minimise prob Type I error. 50 panels are selected at random and precisely measured.747893 and . sx = 0.FINAL PRACTICE EXAM P (SOURCE: PRE AUT 07) Last updated 22nd October 2010 Question P35: A clothing manufacturer has had some bad experiences in the past by taking some actions at times that deviated from what they have normally done.75 centimetres. With 1000 panels.034cm. After 20 hours. the company should consider: (a) maximising critical values (b) minimising statistical risk (c) minimising Type I errors (d) minimising Type II errors ANSWER: minimising Type I errors. again lies in rejection region compared to critical value of 1.05. σ is unknown so assume σ = s. That is.025 upper tail and 999df = 1.75cm (claim) α is unknown so assume α=. you are more likely to stay with the Status Quo. Based on the sample data. a total of 1000 panels have been measured.96 Critical value in cms = μ +/.753 – 0.034.75cm.75cm (status quo) . SE = sx/sqrt(n) Critical value = t with 0. Question P36: A furniture manufacturer receives instructions that the commercial panels it produces must conform so as to be made to an average thickness of . That is.753cm lies in rejection region.75 +/. Ho: μ =. the thickness averages 0. Xbar = 0. For instance.753cm with a standard deviation of . Looking at their future decisions to deviate from current strategies. For instance. they decided to print t-shirts using a new dye given that the rate of defects from several experiments appeared to be significantly lower than defect rates using previous dyes. what should the company conclude about its product meeting the thickness specification? (a) The average thickness is not significantly different to the desired level (b) The average thickness is significantly different to the desired level (c) The sample mean is not normal so we cannot make any conclusion about meeting standards.two tailed Ha: μ ≠ . Upon implementing the new dye system.752107cm Evidence: xbar = 0. Justification: Adapted from Groebner p325.034/sqrt(1000) = .96*.1.75)/(0. Sample size is large (n=1000>30). Justification: The company wants to decrease the likelihood that they make a decision to reject the null hypothesis (status quo) given that the state of nature is that the null hypothesis is true.96 We reject Ho and adopt view suggested by alternative: hence. n=1000.034/sqrt(1000) = 2. it turned out to be a major mistake. Page 16 of 20 . Confirming via test-statistic: Test stat = (0. Minimising Type II is reducing the likelihood you make a decision to NOT reject the null hypothesis even though the null hypothesis is false.T*SE = . concluding Ha: mu>current defect rate when H0: mu<=current defect rate. but would instead use a binomial distribution (d) One would need to observe the entire population of panels to be confident that production meets the desired standard Answer: B = The average thickness is significantly different to the desired level.790245. the average thickness is significantly different from the desired level of . but use t-distribution with n-1 df. Each hour.

p = 400/1000 = . Statement d = If the variation decreases. this decreases the standard error and hence. Justification: Confidence Interval = point estimate +/. a random selection of six is made for a complete audit of the patent office's decision. hence increasing the CI length.66% (c) 31. Hence the length of the CI is unaffected.1% Justification: Using BINOMIAL since independent trials. There are 15 / 50 = 30% of employees in production. Statement b and c = if the actual mean number of visitor numbers decreases or increases.(critical value)(standard error) In this example. but not necessarily that the standard error changes. There are n=6 selections/trials.FINAL PRACTICE EXAM P (SOURCE: PRE AUT 07) Last updated 22nd October 2010 Question P37: The owner of Wild Club and Grill. Hence. Assuming each patent filing is independent of each other. The width of the confidence interval estimate will be narrower: (a) if one decreases the number of evenings that are monitored. Page 17 of 20 . (c) if the mean number of visitor numbers increases.6% (b) 18. its use is incorrect. Answer: D = if the actual variation in visitor numbers over all evenings decreases. What is the probability that exactly two being selected for the audit will have been made with an accelerated request? (a) 4.36% Answer: C = 31.4^2)(.1% (d) 54. This implies the point estimate changes.31104 Question P39: The following table summarises workers roles’ within a particular firm with 50 employees: Department Frequency (Number of Employees) Cumulative Frequency Management 10 10 Production 10 20 Marketing 15 35 Accounting 15 50 Which statement is correct? (a) 30% of employees are employed in production (b) 30% of employees are employed in production or less (c) 30% of employees are employed in marketing (d) 30% of employees are employed in marketing or less Answer: C = 30% of employees are employed in marketing Justification: Total freq = (5+10+15+20) = 50 There are 10 / 50 = 20% of employees in production. implies s or sigma decreasing (even though we may or may not be observing it).6^4) = .40 P(x=2) = 6C2. Hence.(za/2 or t dist value)(s or sigma/ sqrt(n)) Statement a = if one decreases the number of evenings that are monitored. 400 patents were filed with an accelerated request being made. The owner monitors numbers of several randomly selected evenings. length of the CI decreases. This implies that n decreases … this will increase the standard error. wishes to construct a confidence interval estimate for the mean number of customers coming into the club. Students should recognise that the cumulative frequency is meaningless. (d) if the variation in visitor numbers over all evenings decreases.(. CI = sample mean visitors +/. Question P38: Out of 1000 patents filed. (b) if the mean number of visitor numbers decreases. a nightclub.

FINAL PRACTICE EXAM P (SOURCE: PRE AUT 07) Last updated 22nd October 2010 Question P40: A new technology for cars requires that a specially developed vehicle will require a special charging station. The current proposed number of charging stations for a 50km journey follows a Poisson distribution../(x!) P(x=0) = (4^0)(exp(-4)).981684 THIS COMPLETES ALL THE QUESTIONS IN YOUR FINAL EXAM CONGRATULATIONS ☺ Page 18 of 20 . P(x=0) = (mu^x)(exp(-mu)). on average./(0!) P(x=0) = 1*exp(-4)/1 = exp(-4) = .018316 = . P(x>=1) = 1 – P(x=0).. What is the probability at least one station will be encountered within the maximum range journey? (a) 2% (b) 63% (c) 47% (d) 98% Answer: D = 98% Justification Poisson distribution has average mu = 1 stations / 50 km mu = 4 stations / 200 km (multiplying both by 4) We desire at least one i.018316 P(x>=1) = 1 . with only one station encountered. The new vehicle has a maximum range of 200km. similar to the concept of a petrol station.e.

9616 0.1 2.8980 0. z 0.9932 0.9826 0. z=1.9842 0.9922 0.9 1.9850 0.5040 0.9 2.9977 0.6554 0.9370 0.9890 0.8365 0.9115 0.6331 0.9984 0.9177 0.9744 0.5714 0.8389 0.9767 0.8729 0.9968 0.3 2.8554 0.9772 0.8770 0.6915 0.07 0.9985 0.7580 0.9943 0.9066 0.7517 0.9986 0.8212 0.8749 0.9864 0.7764 0.6808 0.9881 0.9082 0.5120 0.9911 0.05 0.8577 0.5832 0.9756 0.03 0.9989 0.3 0.5753 0.9525 0.8413 0.9798 0.9893 0.8665 0.9 3.6 2.9032 0.9279 0.5948 0.6103 0.9945 0.9793 0.6736 0.9382 0.9625 0.9505 0.9868 0.9951 0.6700 0.7123 0.7704 0.9357 0.9015 0.4 1.8907 0.7 1.6664 0.5557 0.9564 0.9957 0.9049 0.7995 0.4 2.5199 0.0 2.2 1.7486 0.9989 0.9956 0.7910 0.6 0.5359 0.9750 0.9934 0.7422 0.7357 0.7054 0.9664 0.9963 0.9633 0.8106 0.5 0.9887 0.9222 0.8 2.9960 0.7257 0.9966 0.00 0.9973 0.8686 0.7019 0.9406 0.9988 0.9808 0.06 0.4 0.5398 0.7967 0.1 1.8340 0.8461 0.6443 0.8599 0.9920 0.9306 0.8944 0.5675 0.9961 0.9394 0.9236 0.9554 0.5279 0.8830 0.04 0.9693 0.9495 0.9319 0.02 0.9649 0.9904 0.7291 0.9671 0.8925 0.9974 0.25 the cumulative probability is .1 0.6368 0.9641 0.8888 0.9982 0.7324 0.9979 0.9452 0.9474 0.5000 0.9738 0.0 0.FINAL PRACTICE EXAM P (SOURCE: PRE AUT 07) Last updated 22nd October 2010 Cumulative Probabilities for the Standard Normal Distribution Entries in the table given the area under the normal probability distribution to the left of the z value.8238 0.9463 0.9989 0.08 0.9916 0.7881 0.7 0.9980 0.9884 0.8531 0.5319 0.7454 0.6255 0.9906 0.9964 0.9515 0.8944.2 0.8315 0.9834 0.0 1.9345 0.8078 0.2 2.8023 0.9686 0.9192 0.9147 0.9441 0.3 1.7549 0.5596 0.8810 0.9732 0.7157 0.6179 0.9982 0.6480 0.5 1.8133 0.7734 0. For example.9959 0.7088 0.6950 0.9783 0.9484 0.9967 0.9985 0.5636 0.9656 0.9854 0.9955 0.9699 0.6141 0.6217 0.7190 0.9162 0.8438 0.9949 0.9861 0.7939 0.09 0.5 2.6406 0.7852 0.7823 0.9986 0.7611 0.6879 0.9761 0.9821 0.9878 0.9927 0.9719 0.6772 0.9987 0.8 0.9292 0.9812 0.9901 0.9918 0.6293 0.5871 0.9857 0.6517 0.5080 0.9979 0.7642 0.8849 0.5517 0.7389 0.9418 0.8186 0.9573 0.9131 0.9608 0.6591 0.8643 0.9931 0.9726 0.9990 Page 19 of 20 .9535 0.5239 0.5160 0.9830 0.9099 0.9332 0.9972 0.8962 0.9946 0.7673 0.9896 0.9678 0.9591 0.8 1.9803 0.9990 0.9898 0.9962 0.9974 0.9817 0.7 2.9953 0.7794 0.9929 0.5438 0.9938 0.9987 0.9976 0.9952 0.9948 0.01 0.9913 0.9925 0.6985 0.9778 0.9706 0.9599 0.6 1.9265 0.6628 0.5910 0.9838 0.8485 0.9987 0.8289 0.7224 0.9871 0.8264 0.5793 0.6026 0.5478 0.9545 0.8051 0.9969 0.9846 0.9582 0.9936 0.8997 0.9978 0.8869 0.9941 0.9940 0.8790 0.9988 0.9909 0.9251 0.9977 0.8621 0.5987 0.8159 0.9984 0.9981 0.9788 0.9429 0.9965 0.0 0.9875 0.9975 0.9981 0.9207 0.9983 0.6064 0.9713 0.6844 0.9970 0.8508 0.9971 0.8708 0.

500 2.289 1.078 1.012 2.681 2.896 2.718 2.306 2.476 1.920 2.604 4.541 3.FINAL PRACTICE EXAM P (SOURCE: PRE AUT 07) Last updated 22nd October 2010 t Distribution Entries in the table give t values for an area or probability in the upper tail of the t distribution.771 1.861 2.363 1.895 1.358 2.086 2.145 2.812.656 9.132 2.311 1.101 2.821 6.747 3.819 2.771 2.841 4.624 2.699 1.353 2.321 1.485 2.787 2.750 2.650 2.845 2.350 1.821 2.415 1.898 2.473 2.052 2.645 0.552 2.341 1.518 2.05 = 1.833 1.055 3.262 2.479 2.706 4.315 1. With 10 degrees of freedom and .372 1.383 1.325 1.228 2.314 2.533 1.714 1.106 3.697 1.492 2.462 2.064 2.943 1.734 1.355 3.998 2.005 63.638 1.746 1.960 0.684 1.313 1.980 1.921 2.365 3.05 area in the upper tail.160 2.333 1.706 1.671 1.05 6.093 2.707 3.042 2.571 2.756 2.499 3.776 2.032 3.337 1.356 1.397 1.080 2.528 2.000 1.10 3.314 1.182 2.056 2.390 2.725 1.201 2.797 2.753 1.282 0.977 2.740 1.048 2.763 2.708 1.539 2. Area in Upper tail Degrees of Freedom 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 40 60 120 ∞ 0.169 3.310 1.764 2.01 31.729 1.179 2.110 2.567 2.711 1.326 0.467 2.796 1.965 4.457 2.658 1.250 3.120 2.069 2.423 2.947 2.925 5.025 12.021 2.782 1.440 1.060 2.812 1. t.330 1.447 2.721 1.143 2.319 1.015 1.508 2.617 2.318 1.328 1.660 2.860 1.878 2.807 2.296 1.583 2.602 2.316 1.303 1.365 2.323 1.576 Page 20 of 14 .345 1.779 2.131 2.717 1.704 2.701 1.074 2.303 3.045 2.831 2.886 1.703 1.761 1.

Sign up to vote on this title
UsefulNot useful

Master Your Semester with Scribd & The New York Times

Special offer for students: Only $4.99/month.

Master Your Semester with a Special Offer from Scribd & The New York Times

Cancel anytime.