Вы находитесь на странице: 1из 4

Section B: [Total marks: 60] (Use a separate booklet for your answers to Section B.

Question 1 [14 marks]

A pharmaceutical company manufactures tablets which, on the packaging, claims that each tablet contains 25
mg of the active ingredient. A random sample of 22 tablets was analysed. The weights (in mg) of the active
ingredient in each tablet were as follows:

24.1 27.2 26.7 23.6 26.4 25.2 25.8 27.3 23.2 28.5 26.9
27.1 26.7 22.7 26.9 24.8 24.0 23.4 25.0 24.5 28.6 26.1

with x = 26.1 mg and s = 1.72 mg and a stem and leaf plot as follows:

22 7
23 2 4 6
24 0 1 5 8
25 0 2 8
26 1 4 7 7 9 9
27 1 2 3
28 5 6

i. What is the value of the leaf units? (2 marks)


ii. Find a 95% confidence interval for the population mean of the active ingredient content. State any
assumptions required to produce this interval, and state also if they seem to be satisfied by the data set.
(4 marks)
iii. Does the claimed population mean lie within the 95% confidence interval? What conclusion can you
draw from this? (2 marks)
iv. If the active ingredient is normally distributed with population mean 25 mg and population standard
deviation 1.2 mg, find the probability that, for a randomly chosen tablet, the amount of active ingredient
exceeds 27.2 mg. (3 marks)
v. Using the assumptions in (iii), find the probability that, for a randomly chosen sample of 4 tablets, the
sample mean lies between 24.5 mg and 25.5 mg. (3 marks)

Question 2 [10 marks]

The number of errors per page of engineering calculations can be considered to be independent from page to
page. Also, from historical data analysis, they are assumed to be modelled by a Poisson distribution with
mean 2 errors per page.
i. What is the probability of obtaining at least one error in a single randomly chosen page? (3 marks)
Four pages are then chosen randomly and four assessors are chosen, each to examine a single page.
ii. Explain why the number of people who find at least one error is a Binomial random variable. (4 marks)
iii. What is the probability that exactly three out of the four assessors find at least one error? (3 marks)

3
Question 3 [5 marks]

For certain ore samples, the proportion, Y, of impurities per sample is given by:

3 y2 + y 0 y 1
f Y ( y) = 2
0 elsewhere

Find the mean and variance of Y.

Question 4 [15 marks]

Mechanical engineers are studying the efficiency of car engines. The following data table shows 1990 values
for fuel economy (in litres per 100km) of 12 engines and their lifetime carbon dioxide emission (in tons of
CO2).

economy 14.79 15.60 14.79 14.04 12.77 12.77 11.70 10.80 11.23 10.41 10.03 8.78
emissions 61.99 59.52 57.93 56.42 53.62 52.04 49.65 46.23 43.74 42.04 38.54 34.57

The data were entered into Excel and a regression analysis performed at the 5% level. The output included
the following table:

SUMMARY OUTPUT
Regression Statistics
Multiple R 0.97338
R Square 0.94747
Adjusted R Square 0.94221
Standard Error 2.10170
Observations 12

ANOVA
Df SS MS F Significance F
Regression 1 796.740 796.740 180.374 1.00645E-07
Residual 10 44.171 4.417
Total 11 840.911

Coefficients Stand Error t Stat P-value Lower 95% Upper 95%


Intercept 1.33591 3.651 0.3658 0.722 -6.7994 9.4712
fuel economy 3.92836 0.292 13.4303 1.00645E-07 3.2766 4.5800

4
The output included the following graphics:

Scatterplot. Histogram of residuals.


1990 engine efficiency and CO2 emission 7
65 6
60 5

Frequency
55
4
50

45
3
40 2
35 1
30
0
25
5 7 9 11 13 15 17 -4 -2 0 2 4
economy (L/100km) Residual Midpoint

i. If Y represents emissions, and X represents fuel economy, write down the equation for the line of best fit
for the data with X as the explanatory variable. (3 marks)
ii. Use the output and graphs above to give two reasons why is it reasonable to fit a line to this data.
(2 marks)
iii. What percentage of variation in emissions is explained by the fit of the line? (1 mark)
iv. Use the line of best fit to predict the amount of emissions released by an engine with economy 9.5
L/100km. (2 marks)
v. Is it reasonable to use this line to predict the amount of emissions produced by an engine with economy
7 L/100km? Justify your answer. (2 marks)
vi. What does the histogram of residuals suggest for the regression model? (1 mark)
vii. Using the output from the regression analysis, test H 0 : 1 = 3 vs H 1 : 1 3 at the 5% level of
significance. Explain all the steps in your testing. (4 marks)

Question 5 [16 marks]


A design engineer is planning for a new chemical process. The construction requires selection of a metal
which is resistant to corrosion i.e., the engineer wishes to minimise corrosion. Three different metal types
were immersed in a corrosive solution and the corrosion rates (%) were measured. Random samples of 7
aluminium strips, 8 stainless steel strips and 8 chromium-vanadium alloy strips were tested and the results
are shown in the following table:

stainless
Aluminium (Al) Steel (SS) Alloy (Cr-Va)
75 74 73
77 76 74
76 75 72
79 78 74
74 74 70
77 77 73
75 75 74
77 71
i. Write down the null and alternate hypotheses that the engineer would test in order to determine whether
there is a significant difference in mean corrosion rates between the three metal types. You must state the
hypotheses first in symbolic form (defining the symbols you use), and, secondly, with words. (3 marks)

5
The data were entered into Excel, and a one-way analysis of variance (ANOVA) was carried out at the 5%
level. The output follows:

Anova: Single Factor


SUMMARY
Groups Count Sum Average Variance
aluminium (Al) 7 533 76.1429 2.8095
stainless steel (SS) 8 606 75.7500 2.2143
alloy (Cr-Va) 8 581 72.6250 2.2679

ANOVA
Source of Variation SS df MS F P-value F crit
Between Groups 57.6809 2 28.8405 11.9590 0.0004 3.4928
Within Groups 48.2321 20 2.4116
Total 105.9130 22

ii. Perform a test of H 0 vs H 1 at the 5% level of significance. Include a diagram (which shows all
relevant values) to help explain the conclusion you make. (3 marks)
iii. The engineer wishes to construct 95% confidence intervals for the difference of (population) means
between the metal types, using the formula

1 1
CI y y2 = ( y1 y 2 ) t ,df s 2 + .
n1 n 2
1 2

Identify which value would be chosen from the output to estimate s2 in the formula, and explain why.
(2 marks)
iv. Three 95% confidence intervals for the pairwise difference of (population) means between the metal
types were obtained:
CI (mean Al meanSS ) = (1.22, 2.11)
CI (mean Al meanCr Va ) = (1.85, 5.18)
CI (meanSS meanCr Va ) = (1.46, 5.69)
In about 10 lines of writing, use these confidence intervals to determine which of the metals are
significantly different from each other, or are similar to each other in corrosion rate. In the discussion,
you should briefly comment on how confidence intervals and hypothesis tests relate to each other.
(6 marks)
v. Finally, in about 3 lines of writing, explain which metal the design engineer will choose for the new
chemical process. (2 marks)

End of Section B.

***************

Вам также может понравиться