37 views

Uploaded by mzansigirl

Business science

- Linking State Education Standards to Your School Garden
- Chapter1 Introduction to Statistics
- Chapter_08_revisesd0.ppt
- Probability and statistic slids
- Statistics
- Project Synopsis Template for Mba
- Chap007 Sampling aa
- BAJAJ
- Unit I - The Nature of Probability and Statistics
- Young Chinese Consumer´s Brand Perceptions towards Global Brands
- NTC ISO 9001 - 2015
- A CQMS 202 Course Outline W2012-1
- Project Management Quiz
- Fyp Present Nazre
- Prevalence Survey4[1]
- A03 - Index
- Research Paper
- Decision Making
- rs
- Example of Research CHAPTER 3

You are on page 1of 373

Applied Business

Statistics

METHODS AND EXCEL-BASED APPLICATIONS

4th Edition

TREVOR WEGNER

SOLUTIONS MANUAL

www.jutaacademic.co.za

TREVOR WEGNER

Applied Business Statistics: Methods and Excel-based Applications: Solutions Manual

Reprinted 2000 and 2003

Second Edition 2008

Third Edition 2012

Fourth edition 2015 (Web PDF)

First Floor

Sunclare Building

21 Dreyer Street

Claremont

7708

All rights reserved. No part of this publication may be reproduced or transmitted in any form or

by any means, electronic or mechanical, including photocopying, recording, or any information

storage or retrieval system, without prior permission in writing from the publisher. Subject to

any applicable licensing terms and conditions in the case of electronically supplied

publications, a person may engage in fair dealing with a copy of this publication for his or her

personal or private use, or his or her research or private study. See section 12(1)(a) of the

Copyright Act 98 of 1978.

The author and the publisher believe on the strength of due diligence exercised that this work

does not contain any material that is the subject of copyright held by another person. In the

alternative, they believe that any protected pre-existing material that may be comprised in it

has been used with appropriate authority or has been used in circumstances that make such

use permissible under the law.

CHAPTER 1

STATISTICS IN MANAGEMENT

1.1 It is a decision support tool. It generates evidence based information through analysis of

data to inform management decision making.

1.2 Descriptive statistics summarises (profiles) sample data; inferential statistics generalises

sample findings to a broader population (to estimate values or confirm relationships).

1.3 Statistical modelling is explores and quantifies relationships between variables for

estimation or prediction purposes.

1.4 Data quality is influenced by: (i) Data source; (ii) Data collection method; and (iii) Data type

1.5 Different statistical methods are valid for different data types.

1.6 In data preparation, consider (i) Data relevancy; (ii) Data cleaning; and (iii) Data enrichment.

(b) Population: All JSE companies

(c) Sample: The 68 HR managers surveyed

(d) Sampling unit: a JSE-listed company

(e) 46% is a sample statistic

(f) Random sampling is necessary to allow valid inferences to be drawn based on the sample

evidence.

(b) Population: All female magazine readers

(c) Sample: The 2000 randomly selected female magazine readers

(d) Sampling unit: a female reader of a female magazine

(e) 35% (700/2000) is a sample statistic

(f) Inferential statistics – as its purpose is to test the belief that market share = 38%

(i) weekly sales volume; (ii) number of ads placed per week; (iii) advertising media used.

(b) Dependent variable = weekly sales volume

(c) Independent variables = number of ads placed per week; advertising media used.

(d) Statistical model building (predict sales volume from ads placed and media used)

1.10 Scenario 1 Inferential statistics

Scenario 2 Descriptive statistics

Scenario 3 Descriptive statistics

Scenario 4 Inferential statistics

Scenario 5 Inferential statistics

Scenario 6 Inferential statistics

Scenario 7 Inferential statistics

(b) numeric, ratio-scaled, continuous {416,2m²; 3406,8m²}

(c) categorical, ordinal-scaled, discrete {matric; diploma}

(d) categorical, nominal-scaled, discrete {married; single}

(e) categorical, nominal-scaled, discrete {Boeing; Airbus}

(f) categorical, nominal-scaled, discrete {verbal; emotional}

(g) numeric, ratio-scaled, discrete {41 ; 62}

(h) categorical, ordinal-scaled, discrete {salary only; commission only}

(i) (i) categorical, ordinal-scaled, discrete {1 = apple; 2 = orange}

(ii) categorical, nominal-scaled, discrete {yes; no}

(iii) categorical, nominal-scaled, discrete {train; bus}

(iv) numeric, interval-scaled, discrete {2; 5}

(j) numeric, ratio-scaled, continuous {12,4kg; 7,234kg}

(k) categorical, nominal-scaled, discrete {Nescafe; Jacobs}

(l) numeric, ratio-scaled, continuous {26,4 min; 38,66 min}

(m) categorical, ordinal-scaled, discrete {Super; Standard}

(n) numeric, ratio-scaled, continuous {R85,47; R2315,22}

(o) numeric, ratio-scaled, discrete {75; 23}

(p) numeric, ratio-scaled, discrete {5; 38}

(q) numeric, ratio-scaled, continuous {9,54 hours; 10,12 hours}

(r) numeric, interval-scaled, discrete {2; 6}

(s) numeric, ratio-scaled, discrete {75; 238}

(t) categorical, nominal-scaled, discrete {Growth funds; Industrial funds}

(c) Illustration value

(b) Economic sector categorical, nominal-scaled, discrete {retail}

Head office region categorical, nominal-scaled, discrete {Gauteng}

Company size numeric, ratio-scaled, discrete {242}

Turnover numeric, ratio-scaled, continuous {R3 432 562}

Share price numeric, ratio-scaled, continuous {R18.48}

Earnings per share numeric, ratio-scaled, continuous {R2.16 / share}

Dividends per share numeric, ratio-scaled, continuous {R0.86 / share}

Number of shares numeric, ratio-scaled, discrete {12 045 622}

ROI (%) numeric, ratio-scaled, continuous {8.64%}

Inflation index (%) numeric, ratio-scaled, continuous {6.75%}

Year established numeric, ratio-scaled, discrete {1988}

1.13 (a) 13 random variables

(c) Illustration value

(b) Gender categorical, nominal-scaled, discrete {female}

Home language categorical, nominal-scaled, discrete {Xhosa}

Position categorical, ordinal-scaled, discrete {middle manager}

Join date numeric, ratio-scaled, discrete {1998}

Status categorical, ordinal-scaled, discrete {gold status}

Claimed categorical, nominal-scaled, discrete {yes}

Problems categorical, nominal-scaled, discrete {yes}

Yes problem categorical, nominal-scaled, discrete {online access difficult}

Services - airlines numeric, interval-scaled, discrete {2}

Services – car rentals numeric, interval-scaled, discrete {5}

Services - hotels numeric, interval-scaled, discrete {4}

Services – financial numeric, interval-scaled, discrete {2}

Services – telecomms numeric, interval-scaled, discrete {2}

Voyager Services Quality data: mainly categorical (qualitative); but when rating scales

are used, such as in Question 8, the data is numeric, but interval-scaled and discrete.

---ooOoo---

CHAPTER 2

(b) multiple (or stacked) bar chart

(c) histogram

(d) scatter plot

Exercise 2.3 Cross-tabulation table (or joint frequency table; or two-way pivot table).

(ii) categories can be displayed in any order

(iii) width of bars is arbitrary (but all of equal widths)

(ii) intervals must be continuous (and constant width) and sequential

(iii) width of bars is determined by interval width

Exercise 2.6 File: X2.6 - magazines.xlsx

Magazine Count %

True Love 95 19%

Seventeen 146 29%

Heat 118 24%

Drum 55 11%

You 86 17%

Total 500 100%

Seventeen

11%

Heat

Drum

29%

24% You

(b) Interpretation

Seventeen is the most popular teenager magazine (29% of female teenager prefer it).

Almost a quarter of the female readers surveyed prefer reading Heat (24%), while the least

preferred magazine is Drum with only 11% of female magazine readers preferring it.

Exercise 2.7 File: X2.6 - magazines.xlsx

Magazine Count %

True Love 95 19%

Seventeen 146 29%

Heat 118 24%

Drum 55 11%

You 86 17%

Total 500 100%

35%

29%

30%

24%

25%

19%

20% 17%

15% 11%

10%

5%

0%

Exercise 2.8 File: X2.8 - job grades.xlsx

A Count 14

% 35% Job Grade %

B Count 11 A 35

% 27.5% B 27.5

C Count 6 C 15

% 15% D 22.5

D Count 9 Total 100

% 22.5%

Total Count 40

Total % 100%

% of Employees per Job Grade

40

35

35

30 27.5

25 22.5

D A A

20 23% 35%

15 B

15 C C

15% B

10 D

27%

5

0

A B C D

Exercise 2.9 File: X2.9 - office rentals.xlsx

(a) and (b) Numerical Frequency Distribution and Cumulative Frequency Distribution

≤ 200 4 13.3 13.3%

201 - ≤250 8 26.7 40.0%

251 - ≤300 9 30 70.0%

301 - ≤350 6 20 90.0%

351 - ≤400 3 10 100.0%

More 0 0 100.0%

Total 30 100

(c) (i) 13.3% of all office space costs less than or equal to R200 / m2

(ii) 70% of all office space costs at most R300 / m²

(iii) 10% of all office space costs more than R350 / m²

(iv) 9 office buildings have rentals between R300 and R400 / m²

Exercise 2.10 File: X2.10 - storage dams.xlsx

Wemmershoek 158644 16.9

Steenbras 95284 10.2

Voelvlei 244122 26

Theewaterskloof 440255 46.9

Total capacity 938305 100

(in Million litre)

17% Wemmershoek

47% 10% Steenbras

Voelvlei

26% Theewaterskloof

(ii) Wemmershoek and Steenbras dams together provide 27.1% of Cape Town's water.

Exercise 2.11 File: X2.11 - taste test.xlsx

A Liqui Fruit 45 18.0

B Fruiti Drink 26 10.4

C Yum Yum 64 25.6

D Fruit Quencher 38 15.2

E Go Fruit 77 30.8

Total 250 100.0

40.0

35.0

30.8

30.0

25.6

% of consumers

25.0

20.0 18.0

15.2

15.0

10.4

10.0

5.0

0.0

Liqui Fruit Fruiti Drink Yum Yum Fruit Quencher Go Fruit

45, 18%

Liqui Fruit

77, 31%

Fruiti Drink

26, 10% Yum Yum

Fruit Quencher

64, 26%

(c) 56.4% of the sampled consumers prefer either Yum Yum or Go Fruit.

Exercise 2.12 File: X2.12 - annual car sales.xlsx

Toyota 96959 19.4

Nissan 63172 12.6

Volkswagen 88028 17.6

Delta 62796 12.6

Ford 74155 14.8

MBSA 37268 7.5

BMW 51724 10.4

MMI 25354 5.1

Total Sales 499456 100.0

120000

96959

100000

88028

80000 74155

63172 62796

60000 51724

37268

40000

25354

20000

MMI, 5.1

BMW, Toyota

10.4 Toyota, 19.4 Nissan

MBSA, 7.5

Volkswagen

Delta

Nissan, 12.6

Ford, 14.8 Ford

MBSA

Volkswagen,

BMW

Delta, 12.6 17.6

MMI

(c) Total % held by top three manufacturers - Toyota (19.4%), Volkswagen (17.6%)

and Ford (14.8%) - represents 51.8% of the total passenger car market.

Exercise 2.13 File: X2.13 - half-yearly car sales.xlsx

Toyota 42661 54298 27.3

Nissan 35376 27796 -21.4

VW 45774 42254 -7.7

Delta 26751 36045 34.7

Ford 32628 41527 27.3

MBSA 19975 17293 -13.4

BMW 24206 27518 13.7

MMI 14307 11047 -22.8

60000

50000

40000

First half

30000

Second half

20000

10000

0

Toyota Nissan VW Delta Ford MBSA BMW MMI

First half 42661 35376 45774 26751 32628 19975 24206 14307

Second half 54298 27796 42254 36045 41527 17293 27518 11047

(b) First half-year best performers: Nissan; Volkswagen; MBSA and MMI

(c) Delta showed the largest % increase from the first half to the second half of 34.7%.

Refer to the above Table for the % Change between First and Second Half-Year Sales.

Exercise 2.14 File: X2.14 - television brands.xlsx

Count of Brands

Brands Total

Daewoo 16%

LG 30.4%

Philips 10.4%

Sansui 24%

Sony 19.2%

Grand Total 100%

% of TV Brands Owned

35%

30.4%

30%

Brands Total 24%

25% Daewood 16%

20% LG 30.4% 19.2%

Philips

16% 10.4%

15% Sansui 24%

10.4%

10%

Sony 19.2%

Total 100%

5%

0%

Daewood LG Philips Sansui Sony

(c) Philips is the least preferred brand (preferred by only 10.4% of households surveyed).

(d) The most popular brand is LG that is owned by 30.4% of the households surveyed.

Exercise 2.15 File: X2.15 - estate agents.xlsx

Count of House sales

House sales Total

3 12

4 15

5 6

6 7

7 5

8 3

Grand Total 48

16

14

12

10

8

15

6 12

4

6 7

2 5

3

0

3 4 5 6 7 8

Total 12 15 6 7 5 3

(c) The most frequently sold number of properties per estate agent was 4.

4 properties each were sold by (15/48) 31.3% of all estate agents

(d) The same frequency count table (a) and histogram (b) is produced.

Exercise 2.16 File: X2.16 - fast foods.xlsx

KFC 56 17.2 KFC 17.2

St Elmo's 58 17.8 St Elmo's 17.8

Steers 45 13.8 Steers 13.8

Nandos 64 19.7 Nandos 19.7

Ocean B 24 7.4 Ocean B 7.4

Butler's 78 24.0 Butler's 24

Total 325 100

Percentage of Consumers

30.0

24

25.0

19.7

20.0 17.2 17.8

13.8

15.0

%

10.0 7.4

5.0

0.0

KFC St Elmo's Steers Nandos Ocean B Butler's

Firstly produce the categorical frequency table of Food Type Preferences. Sum

the frequency counts of the different food types (e.g. Chicken = 56 + 64 = 120) and

express the total count as a % of the total number of customers (e.g. 120/325 = 36.9%).

Food type %

Chicken 36.9

Pizza 41.8

Burger 13.8

Fish 7.4

Consumer preference (%) by Food Type

Fish, 7.4

Burger, 13.8

Chicken

Chicken,

36.9 Pizza

Burger

Pizza, 41.8

Fish

Pizza (42%) and Chicken (37%) dominate almost 80% of

the fast food market with Pizzas being slightly more favoured by fast food consumers.

Exercise 2.17 File: X2.17 - airlines.xlsx

(a) Two-way Pivot Table - Counts, Row % (by Airline), Column % (by Passenger)

Passenger

Airline Data Business Tourist Grand Total

Comair Count of Passenger 12 8 20

Row % 60.0% 40.0% 100%

Column % 33.3% 23.5% 28.6%

Total % 17.1% 11.4% 28.6%

Kulula Count of Passenger 4 16 20

Row % 20.0% 80.0% 100.0%

Column % 11.1% 47.1% 28.6%

Total % 5.7% 22.9% 28.6%

SAA Count of Passenger 20 10 30

Row % 66.7% 33.3% 100.0%

Column % 55.6% 29.4% 42.9%

Total % 28.6% 14.3% 42.9%

Total Count of Passenger 36 34 70

Total Row % 51.4% 48.6% 100%

Total Column % 100% 100% 100%

Total Total % 51.4% 48.6% 100%

Airline Business Tourist Grand Total

Comair 60.0% 40.0% 100.0%

Kulula 20.0% 80.0% 100.0%

SAA 66.7% 33.3% 100.0%

Grand Total 51.4% 48.6% 100.0%

90%

0.8

80%

% of passenger per airline

70% 0.67

0.6

60%

50%

0.4

40% 0.33 Business

30% Tourist

0.2

20%

10%

0%

Comair Kulula SAA

Business 0.6 0.2 0.666666667

Tourist 0.4 0.8 0.333333333

(e) Kulula is most preferred by tourists (47.1% of tourists prefer Kulula).

More than half (55.6%) of all business travellers prefer SAA.

Exercise 2.18 File: X2.18 - car occupants.xlsx

Data Type - Numerical, discrete, ratio-scaled

1 23 38.3 23 38.3

2 15 25.0 38 63.3

3 10 16.7 48 80.0

4 5 8.3 53 88.3

5 7 11.7 60 100.0

Total 60 100.0

30

25 23

20

No. of Cars

15

15

10

10 7

5

5

0

1 2 3 4 5

Occupants

(b) (iii) "Less-Than" Ogive (see (a) above) and Cumulative Frequency Polygon

70

60 60

53

50 48

No. of Cars

40 38

30

23

20

10

0

1 2 3 4 5

No. of occupants

(c) (i) 38.3% of motorists travel alone.

(ii) 36.7% of vehicles have 3 or more occupants

(iii) 63.3% of vehicles have no more than 2 occupants.

Exercise 2.19 File: X2.19 - courier trips.xlsx

(a) Random variable - distance travelled (in kms) per courier trip

Data type - numerical, continuous, ratio-scaled

≤10 4 8 4 8

11 - ≤15 7 14 11 22

16 - ≤20 15 30 26 52

21 - ≤25 12 24 38 76

26 - ≤30 9 18 47 94

31 - ≤35 3 6 50 100

Total 50 100

18

16 15

14

12

12

No. of trips

10 9

8 7

6

4

4 3

2

0

≤10 11 - ≤15 16 - ≤20 21 - ≤25 26 - ≤30 31 - ≤35

Distances Travelled per Trip

Distance Cum %

100 100

10 8 94

80 15 22

76

Percent of trips

20 52

60 25 76

52

30 94

40

35 100

20 22

8

0

10 15 20 25 30 35

(c) (i) 18% of deliveries (9 trips) were between 25 km and 30 km.

(ii) 76% of deliveries (38 trips) were within a 25 km radius.

(iii) 48% of deliveries (24 trips) were beyond a 20 km radius.

(iv) Reading off the Cumulative % Frequency Table (or Polygon) above

52% of trips were no more than 20 km from the depot.

(v) The longest 24% of trips were above 25 km.

Hence there is adherance to the company policy.

Exercise 2.20 File: X2.20 - fuel bills.xlsx

Data type - numerical, continuous, ratio-scaled

(b) (i) (ii) Numeric % Frequency Distribution (and (d) the Ogive)

≤300 7 14 7 14

301 - 400 15 30 22 44

401 - 500 13 26 35 70

501 - 600 7 14 42 84

601 - 700 5 10 47 94

701 - 800 2 4 49 98

800+ 1 2 50 100

Total 50 100

16 15

14 13

12

no. of motorists

10

8 7 7

6 5

4

2

2 1

0

≤300 301 - 400 401 - 500 501 - 600 601 - 700 701 - 800 800+

Fuel bill (in Rands)

(c) 14% (7 motorists) spend between R500 and R600 per month on fuel.

(See the Ogive in (b) for Cumulative Frequencies)

Cumulative % Frequency Polygon - Fuel Bills per Month

110

100 98 100

90 94

80 84

70 70

60

50

40 44

30

20

10 14

0

≤300 301 - 400 401 - 500 501 - 600 601 - 700 701 - 800 800+

(e) From (d), approx. 77% of motorists spend less than R550 per month on fuel.

(f) From (a) (and (d)), 30% (15 motorists) spend more than R500 per month on fuel.

Exercise 2.21 File: X2.21 - car sales.xlsx

1 37

2 25

3 41

4 29

5 31

6 28

7 30

8 36

9 38

10 62

11 53

12 63

13 43

14 39

15 52

16 61

17 58

18 65

19 73

20 52

21 61

22 46

23 49

24 54

80

70

60

50

units sold

40

30

20

10

0

1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24

Quarters

Exercise 2.22 File: X2.22 - market shares.xlsx

1 13.4 9.9

2 11.6 9.6

3 9.8 11.2

4 14.4 12.0

5 17.4 11.6

6 18.8 13.1

7 21.3 11.7

8 19.4 14.2

9 19.6 16.0

10 19.2 16.9

(a) Trend Line graphs of Market Shares (%) per car type (VW, Toyota)

25

20

% market share

15

10

0

1 2 3 4 5 6 7 8 9 10

year

Toyota shows a lower sales level, but at a rising growth rate.

(c) Choice of franchise is not clearcut, but suggest choosing Toyota because

of its more consistent (steady) growth rate.

Exercise 2.23 File: X2.23 - defects.xlsx

20

18

no. of defects found per batch

16

14

12

10

8

6

4

2

0

20 30 40 50 60 70

between the inspection time of a batch and the no. of defects found per batch.

Data Consignment Time Defects

AA 48 17

AB 50 9

AC 43 12

AD 36 7

AE 45 8

AF 49 10

AG 55 14

AH 63 18

AI 55 19

AJ 36 6

AK 40 8

AL 46 14

AM 32 10

AN 50 15

AO 42 14

AP 36 8

AQ 48 12

AR 38 8

AS 45 10

AT 30 6

AU 34 9

AV 43 11

AW 53 16

AX 48 16

AY 56 15

AZ 40 12

BA 33 7

BB 35 10

BC 50 16

BD 48 18

Exercise 2.24 File: X2.24 - leverage.xlsx

Scatter Graph

Profit Growth (y) and Leverage Ratio (x)

160

140

120

profit growth

100

80

60

40

20

0

30 32 34 36 38 40 42 44 46

leverage ratio

(b) Yes, there is a clear moderate to strong positive linear relationship between

the leverage ratio of a company and its profit growth.

40.8 111

42.3 116

43.2 132

37.9 105

36.2 69

35.6 40

36.4 58

39.5 118

42.6 104

42.1 125

37.8 97

34.4 76

36.5 98

38.3 100

39.3 75

36.4 88

33.5 20

32.4 25

35.4 78

35.4 65

35.7 84

35.2 88

35.3 86

44.9 115

35.9 50

38.0 92

36.7 110

39.2 72

41.1 128

38.7 86

Exercise 2.25 File: X2.25 - roi%.xlsx

Mining 9.87 4.58

Services 11.33 2.99

Grand Total 10.70 3.78

(b) Service companies have a higher average ROI% (11.33%) than mining companies (9.87%).

The volatility of ROI% amongst service companies (2.99%) is far lower than amongst mining companies (4.58%)

By inspection, there is a high overlap of ROI% between the two sectors (based on a two-standard deviation

interval around each sample mean). Thus it is likely that there is no statistically significant difference in

mean ROI% between the two sectors.

Aisle Middle Top Total

Front Average 6.08 5.08 5.58

Std dev 0.890 0.622 0.895

Middle Average 4.24 2.74 3.49

Std dev 1.387 0.297 1.232

Back Average 4.66 4.1 4.38

Std dev 1.193 0.648 0.952

Average 4.99 3.97 4.48

Std dev 1.359 1.114 1.327

Based on shelf position alone, middle shelf positions generate higher average sales (R4.99) than top shelf

(b)

positions (R3.97).

Based on aisle location alone, a front-of-store aisle location generates the highest average sales (R5.58), followed

by a back-of-store aisle location (R4.38).

The lowest average sales occur when the product is displayed in a middle-of-store aisle location (R3.49).

In combination, a front-of-store aisle location on a middle shelf position generates the highest average sales

(R6.08), while a top shelf position in a middle-of-store aisle is the least desirable product location with an average

sales volume of only R2.74.

Sales variability is relatively consistent across aisle locations (0.895 to 1.232) as well as between shelf positions

(1.114 to 1.359).

In combination however, sales volumes show highest variability when the product is positioned in a middle shelf

position in a middle-of-store location (1.387) while the lowest variability in sales volumes occur when positioned in

a top shelf position in a middle-of-store aisle location (0.297).

The large differences in average sales volumes (ranging from R6.08 to R2.74) is evidence of a likley statistically

significant difference in sales volumes due to choice of aisle location and shelf positioning.

Recommendation: A middle shelf position in a front-of-store ailse location is the most preferred product

(c)

display location.

Exercise 2.27 File: X2.27 - property portfolio.xlsx

-5 0 0.0%

-2.5 4 1.2%

0 7 3.4%

2.5 4 4.6%

5 22 11.4%

7.5 87 38.3%

10 109 71.9%

12.5 46 86.1%

15 31 95.7%

17.5 12 99.4%

20 1 100%

More 1 100%

324 100%

120 109 120.0%

100 87 100.0%

80 80.0%

Frequency

60 60.0%

46

40 31 40.0%

22

20 12 20.0%

4 7 4

0 1 1

0 0.0%

-5 -2.5 0 2.5 5 7.5 10 12.5 15 17.5 20 More

Bin NP%

Region Commercial Industrial Retail Total

A Average 7.5 4.9 10.2 7.9

Std dev 2.5 3.1 3.6 3.5

Minimum -2.4 -4.2 0.8 -4.2

Maximum 14.6 8.1 18.4 18.4

Count 104 40 70 214

B Average 12.3 6.8 8.5 9.8

Std dev 3.1 4.2 1.8 3.5

Minimum 2.7 -3.4 4.4 -3.4

Maximum 20.3 10.2 13.2 20.3

Count 46 16 48 110

Average 9.0 5.4 9.5 8.6

Std dev 3.5 3.5 3.1 3.6

Minimum -2.4 -4.2 0.8 -4.2

Maximum 20.3 10.2 18.4 20.3

Count 150 56 118 324

The company has almost twice as many properties in region A (214 or 66%) compared to region B (110 or 34%).

Almost half of their properties are commercial (46%) followed by retail (37%) and then industrial (17%).

Of all the prpoperties in the portfolio, the majority are commercial properties in region A (104 or 32%).

followed by retail properties in region A (70 or 22%).

The smallest component of their property portfolio consists of industrial properties in region B (only 16 or 5%).

(e) Portfolio performance :

Net profit % across the entire portfolio is normally distributed (histogram) with an average return of 8.6% and a

standard deviation of 3/6%. NP% ranged from the lowest of -4.2% to the highest of 20.3%.

From the cumulative % distribution, 75% of all properties (cumulative 86.1% - cumulative 11.4%) earned a NP% of

between 5% and 12.5% p.a.

Region B (9.8%) has outperformed region A (7.9%) by almost 2% on average, while commercial (9.0%) and retail (9.5%)

have significantly outperformed industrial properties (5.4%).

The worst performing segment is industrial properties in region A (4.9%) while the best performing properties are

commercial properties in region B (12.3%).

There are 15 (4.6%) properties that are under-performing (with less than a 5% net profit % p.a.).(see histogram).

Volatility of NP% is fairly consistent across the segments (approx. 3.5%), except for higher variability noted in the

industrial properties of region B (4.2%).

Growth potential (high NP% p.a. segments) is mainly in commercial properties in region B which represents only 14% of the current portfolio

retail properties in region A (currently only constitute 22% of the current portfolio).

(f) Recommendations: Dispose of the worst performing industrial properties in both regions A and B and

purchase more commercial properties in region B followed by retail properties in region A.

CHAPTER 3

Exercise 3.3 Statements (c) and (f). The mode would be more appropriate (both are categorical)

Exercise 3.4 (a) False (b) False (c) True (d) False (e) False

The new median mass will depend only on the masses of parcels in the 3rd and 4th

ordered positions out of the 6 positions (after adding the extra parcel).

Also, the rank order position of this extra parcel's mass is unknown. It could be the

4th, 5th or 6th ranked mass, but this depends on the masses of parcels that are heavier

than the current median mass.

If the 4th ranked mass is also equal to 6.5 kg, then the new median will not increase.

If, on the other hand, the 4th ranked mass is greater than 6.5 kg, then the median will

increase. Therefore the only statement that can be made with complete certainty is (c),

(i.e. that it is impossible for the new median mass to be less than it was.)

Exercise 3.5 Correct method is (b). Use the formula for the arithmetic mean (Formula 3.1)

By definition, Mean = Σx / n

compute Σx (total number of persons) = Mean x n .

i.e. total no. persons in Mossel Bay = 4.1 x 9245 = 37905 (rounded)

Exercise 3.6 File: X3.6 - equity returns.xls

General Equity Unit Trust % Returns

9.2 -0.1 0.01

8.4 -0.9 0.81

10.2 0.9 0.81

9.6 0.3 0.09

8.9 -0.4 0.16

10.5 1.2 1.44

8.3 -1 1

Sum 65.1 4.32

n 7

Using Excel

Exercise 3.7 File: X3.7 - luggage weights.xls

11 0.43 0.1849

12 1.43 2.0449

8 -2.57 6.6049

10 -0.57 0.3249

13 2.43 5.9049

11 0.43 0.1849

9 -1.57 2.4649

Sum 74 17.7143

n 7

Using Excel

68.3% of all hand luggage is likely to weigh between 8.85 kg and 12.29 kg.

(This corresponds to one standard deviation limits about the mean).

Exercise 3.8 File: X3.8 - bicycle sales.xls

25 1.4 1.96 16

18 -5.6 31.36 18

30 6.4 40.96 18

36 12.4 153.76 19

18 -5.6 31.36 20

20 -3.6 12.96 24

16 -7.6 57.76 25

24 0.4 0.16 30

30 6.4 40.96 30

19 -4.6 21.16 36

Sum 236 392.4

n 10

On average, 23.6 bicycles are sold each month.

For half of the months (i.e. 5 months), bicycles sales were

less than 22 bicycles per month.

(b) Range = 36 - 16 = 20

The range of sales between the worst and best months was 20 bicycles.

(i.e. In the worst sales month, 16 were sold; in the best month, 36 were sold.

68.3% of all monthly bicycle sales are likely to lie between 17 and 30.2.

th

(c) Lower Quartile (Q1) = 18 (2.5 position)

25% of monthly bicycle sales were less than or equal to 18.

Or: No more than 18 bicycles per month were sold in 25% of the months.

Note: Using Excel : QUARTILE(data values,1) = 18.25

25% of monthly bicycle sales were above 27.5.

Or: More than 28 (27.5) bicycles per month were sold in 25% of the months.

Note: Using Excel : QUARTILE(data values,3) = 28.75

There is moderate positive skewness in monthly bicycle sales.

(i.e. There are one / two months with relative high bicycle sales)

(e) Box plot of monthly bicycle sales

Interpretation

Monthly bicycle sales range between 16 and 36.

The median monthly sales is 22. The positive skewness shows a wider spread

of monthly sales toward the months of high sales.

(f) Opening monthly stock level = 23.6 + 6.603 = 30.2 bicycles in stock

If orders = 30, then the dealer will have sufficient bicycle stock to meet demand.

Exercise 3.9 File: X3.9 - setting times.xls

27 3 9

18 -6 36

21 -3 9

22 -2 4

20 -4 16

28 4 16

31 7 49

25 1 1

24 0 0

Sum 216 140

n 9

(c) No, since the consistency index is greater than 10%, this consignment

will not be approved for dispatch to the clients.

Exercise 3.10 File: X3.10 - wage increases.xls

5.6 -0.83 0.6910 3.4

7.3 0.87 0.7547 4.8

4.8 -1.63 2.6610 5.3

6.3 -0.13 0.0172 5.6

8.4 1.97 3.8760 5.8

3.4 -3.03 9.1885 5.8

7.2 0.77 0.5910 5.8

5.8 -0.63 0.3985 6.2

8.8 2.37 5.6110 6.3

6.2 -0.23 0.0535 7.2

7.2 0.77 0.5910 7.2

5.8 -0.63 0.3985 7.3

7.6 1.17 1.3660 7.4

7.4 0.97 0.9385 7.6

5.3 -1.13 1.2797 8.4

5.8 -0.63 0.3985 8.8

Sum 102.9 28.8144

n 16

Median = (6.2+6.3)/2 = 6.25 % (lies in the 8.5th position)

Std dev = √28.8144/(16-1) = 1.386 %

Upper limit = 6.43+2*(1.386) 9.20

95.5% of all agreed wage increases lie between 3.66% and 9.2%.

Agreed wage increases are only moderately consistent.

Using Excel

Excel 's Data Analysis option Excel 's Function Keys

Wage increases

(a) Mean 6.43 '=AVERAGE(data values)

Standard Error 0.346

Median 6.25 '=MEDIAN(data values)

Mode 5.8

(b) Standard Deviation 1.386 '=STDEV(data values)

Sample Variance 1.921 '=VAR(data values)

Kurtosis 0.196

Skewness -0.286

Range 5.4

Minimum 3.4

Maximum 8.8

Sum 102.9

Count 16

Exercise 3.11 Bank Trainee Exam Performance

Group 1 Group 2

Mean 76 64

Variance 110 88

Sample size 34 26

10.488 9.381

Group 1 Group 2

Coefficient (10.488/76)% (9.381/64)%

of Variation = 13.8 14.66

Interpretation

Both groups performed consistently well. The difference in CV% measures is marginal.

However, group 1 trainees performed marginally more consistently than group 2 trainees.

Exercise 3.12 File: X3.12 - meal values.xls

Data type - numerical, continuous, ratio-scaled (Using Data Analysis in Excel )

Meal values

(b) Mean 1119/20 = = R55.95 Mean 55.95

Standard Error 3.20

∑(deviation)² = 3902.95 Median 53

n= 20 Mode 44

Standard Deviation 14.33

Variance = 3902.95/(20-1) = 205.42 Sample Variance 205.42

Std deviation = √205.42 = 14.33 Kurtosis 0.26

Skewness 0.74

Range 55

(c) Median Minimum 35

th th

Average the Rand values in 10 and 11 positions. Maximum 90

= (51+55)/2 = R53 Sum 1119

Half off the meals were valued at R53 or less. Count 20

Ranked Meal

Position Value

1 35

2 36

3 44

4 44

5 44

6 47

7 48

8 48

9 50

10 51 Median is midpoint between

11 55 these two middle values (R51 and R55)

12 56

13 58

14 62

15 65

16 65

17 69

18 72

19 80

20 90

(d) Mode R44 occurs 3 times (see grouped ranked values in (c) above).

(e) There is moderate positive skewness caused by two high meal values (i.e. R80 and R90).

Hence recommend the median as the most representative central location meal value.

Exercise 3.13 File: X3.13 - days absent.xls

(a) Mean = 237/23 10.3 days absent (Using Data Analysis in Excel )

i.e. 12th position Mean 10.30

Standard Error 1.33

Median = 9 days absent Median 9

Mode 5

Position Days absent Standard Deviation 6.36

1 2 Sample Variance 40.49

2 4 Skewness 1.38

3 4 Range 28

4 5 Minimum 2

5 5 Q1 position Maximum 30

6 5 and value Sum 237

7 6 Count 23

8 6

9 6 Q1 5.5

10 8 Q3 15

11 9

12 9 Median position and value

13 10

14 10

15 10

16 12

17 15 Q3 position

18 15 and value

19 15

20 16

21 17

22 18

23 30

All occur with a frequency of 4.

Note: Only the first modal value is reported in Excel .

The mode is an unreliable measure of central location in this study.

Interpretation

On average, an employee is absent for 10.3 days over this 9-month period.

Half the employees were absent for up to 9 days.

The most common number of days absent was 5 (or 6, or 10 or 15) days

Q1 position = (23/4) = 5.75th position

Q1 value in this position is = 5+(0.75*(5 - 5) = 5 days absent

Using Excel 5.5 days absent =QUARTILE(data range,1)

25% of employees were absent for no more than 5 (or 5.5) days altogether.

Upper Quartile (approximated manually)

Q3 position = (23*3/4) = 17.25th position

Q3 value in this position is = 15+(0.25*(15 - 15) = 15 days absent

Using Excel 15 days absent =QUARTILE(data range,3)

25% of employees were absent for more than 15 days over this 9-month period.

Average per 1-month = 1.14 days absent per month = (10.3 / 9) = ave per month

Since the monthly average is above 1 day (actually 1.14 days), the company is

not succesfully managing its absenteeism levels.

Exercise 3.14 File: X3.14 - bad debts.xls

n= 17 Mean 4.665

Standard Error 0.45

Variance = 54.85882/(17-1) = 3.43 Median 5.4

Std deviation = √(3.43) = 1.85 Mode 2.2

Standard Deviation 1.85

Sample Variance 3.43

th

(b) Median Median is found in the (17+1)/2 position Kurtosis -1.49

th

i.e. 9 position Skewness -0.35

Median = 5.40% Range 5.4

Minimum 1.8

Rank Ordered Maximum 7.2

Position bad debts % Sum 79.3

1 1.8 Count 17

2 2.2

3 2.2 Q1 2.6

4 2.4 Lower Quartile position Q3 6.1

5 2.6 and value

6 3.4

7 4.4

8 4.7

9 5.4 Median position and value

10 5.7

11 5.7

12 5.8 Upper Quartile position

13 6.1 and value

14 6.3

15 6.6

16 6.8

17 7.2

(c) Average: On average, each furniture retailer has a bad debt % of 4.665%.

Median: 50% of furniture retailers have a bad debt % of 5.4% or less.

Since the mean < median, there is evidence of negative skewness.

Hence propose the use of the median as the representative central value.

(d) There are two modal values (2.2% and 5.7%) both occurring with frequency of 2.

This makes the mode an unreliable measure of central location.

(Formula 3.14) ∑(x - x(bar))3 = -31.2227

n= 17

s (std dev) = 1.85

3

Then Skp = (17*(-31.2227))/((17-1)*(17-2)*1.85 ) = -0.35

There is evidence of moderate negative skewness in the data on bad debts %

(i.e. only a few fairly low bad debt % values - the majority are higher).

(f) Lower Quartile (approximated manually)

Q1 position = (17/4) = 4.25th position

Q1 value in this position is = (2.4*+(0.25*(2.6 - 2.4)) = 2.45%

Using Excel = 2.6% =QUARTILE(data range,1)

25% of furniture retailers have a bad debts % of no more than 2.45% (or 2.6%).

Q3 position = (17*3/4) = 12.75th position

Q3 value in this position is = (5.8*+(0.75*(6.1 - 5.8)) = 6.025%

Using Excel = 6.1% =QUARTILE(data range,3)

25% of furniture retailers have a bad debts % of more than 6.025% (or 6.1%).

(g) The average % bad debts is 4.665% while the median % bad debts is 5.4%. Since

there is moderate negative skewness (Skp = -0.35), the median should be used

as the representative central value. Thus, since the median % of bad debts,

is above 5% (median = 5.4%), an advisory note should be sent out.

Exercise 3.15 File: X3.15 - fish shop.xls

(or Mean) 500 - 750 625 15 9375

750 - 1000 875 23 20125

1000 - 1250 1125 55 61875

1250 - 1500 1375 92 126500

1500 - 1750 1625 65 105625

1750 - 2000 1875 50 93750

Total 300 417250

Interpretation The average daily turnover for the fish shop is R1390.80 per day.

500 - 750 15 15

750 - 1000 23 38

1000 - 1250 55 93 Q1 interval

1250 - 1500 92 185 Median Interval

1500 - 1750 65 250 Q3 interval

1750 - 2000 50 300

Total 300

Median value therefore lies in the 4th interval [i.e. Between R1250 and R1500]

(based on Formula 3.2)

as it has the highest frequency count of 92 days

(based on Formula 3.3)

Q1 position = (300/4) = 75th position which lies in the interval [1000 - 1250]

Q1 value in this position is = (1000+250*(75-38)/(93-38)) = R 1,168.18

The maximum turnover for the slowest 25% of trading days was R1168.18.

Q3 position = (300*3/4) = 225th position which lies in the interval [1500 - 1750]

Q3 value in this position is = (1500+250*(225-185)/(250-185)) = R 1,653.85

A minimum turnover of R1653.85 was generated on the 25% of busiest trading days.

Exercise 3.16 File: X3.16 - grocery spend.xls

(Mean) 10 - 20 15 6 90

20 - 30 25 14 350

30 - 40 35 16 560

40 - 50 45 10 450

50 - 60 55 4 220

Total 50 1670

10 - 20 6 6

20 - 30 14 20 Q1 interval

30 - 40 16 36 Median Interval

40 - 50 10 46 Q3 interval

50 - 60 4 50

Total 50

Median value therefore lies in the 3rd interval [30 - 40]

Interpretation 50% of families spend no more than 33.125% of their income on groceries.

Q1 position = (50/4) = 12.5th position which lies in the interval [20 - 30]

Q1 value in this position is = 20+(10*(12.5-6)/(20-6)) = 24.64%

Interpretation 25% of families spend no more than 24.64% of their income on groceries.

Q3 position = (50*3/4) = 37.5th position which lies in the interval [40 - 50]

Q3 value in this position is = 40+(10*(37.5-36)/(46-36)) = 41.5%

Interpretation 25% of families spend more than 41.5% of their income on groceries.

Exercise 3.17 File: X3.17 - equity portfolio.xls

40 15 600

10 20 200

5 40 200

50 10 500

105 Total 1500

(This is a weighted average measure)

(Using Formula 3.5)

Exercise 3.18 File: X3.18 - car sales.xls

5 25000 125000

12 34000 408000

3 55000 165000

20 Total 698000

(This is a weighted average measure)

(Using Formula 3.5)

Exercise 3.19 File: X3.19 - rental increases.xls

4

Geometric mean = √(1.16*1.14*1.1*1.08) - 1 = 0.11955

Interpretation

The average annual % escalation rate in office rentals is 11.955%

Exercise 3.20 File: X3.20 - sugar increases.xls

6

Geometric mean = √(1,05*1,12*1,06*1,04*1,09*1,03) - 1 = 0.06456

Interpretation

The average annual % increase in the sugar price (per kg) has been 6.456%

(b) The geometric mean is more appropriate since the base value of each

percentage change is different .

Each year's percentage change is based on the previous year's sugar price.

Exercise 3.21 File: X3.21 - water usage.xls

water usage

(a) Mean 21.2

Standard Error 1.751

Median 19.5

Mode 25

(b) Standard Deviation 9.590

Sample Variance 91.959

Kurtosis 1.744

Skewness 1.165

Range 42

Minimum 8

Maximum 50

Sum 636

Count 30

Upper Quartile - Q3 25.75 '=QUARTILE(data range,3)

The average monthly water consumption per household is 21.2 kl.

50% of households consume no more than 19.5 kl per month.

The most frequently occuring monthly consumption is 25 kl (Note that this modal

(value is a misleading and misrepresentative measure at it only occurs 3 times).

Based on the mean and standard deviation values, approximately 95.5% of

households consume between 2.02 kl and 40.38 kl per month.

Note, that the data is heavily skewed to the right (see skewness = 1.165), implying

that there are a few households that consume high volumes of water per month.

This skewness makes the interpretation given above unreliable as both the mean

and the standard deviation are likely to be inflated (or over-estimated) due to the

presence of a few high-valued outliers).

Interpretation - quartiles

25% of households consume at most 15 kl of water per month.

The 25% of heavy water consumers use at least 25.75 kl per month.

- expected total usage per month 21.2 x 750 = 15 900 kl

- expected total usage per year 15900 x 12 = 190 800 kl

Exercise 3.22 File: X3.22 - veal dishes.xls

(a) Random variable - cost of a veal cordon bleu meal at a Durban restaurant

Data type - numerical, continuous, ratio-scaled.

Mean 61.25

Standard Error 1.99

Median 59

Mode 48

Standard Deviation 10.54

Sample Variance 111.08

Kurtosis 0.62

Skewness 0.78

Range 45

Minimum 45

Maximum 90

Sum 1715

Count 28

Interpretation

On average, a patron can expect to pay R61.25 for a veal cordon bleu meal

at a Durban restaurant.

50% of Durban restaurants charge no more than R59 for a veal cordon bleu meal.

(c) There are two modal values (R48 and R55). Both occur with a frequency of 3.

It is a misleading value because of its low frequency of occurrence.

Note: Excel only shows the first occurrence of a modal value (i.e. R48) in its output.

This means that 68.3% of Durban restaurants are likely to charge between

R50.71 (61.25 - 10.54) and R71.79 (61.25 + 10.54) for a veal cordon bleu meal.

Skewness coefficient (approx) = 0.64 Formula 3.15

The relative high positive skewness is caused by two Durban restaurants

charging high prices (R80 and R90) for a veal cordon bleu meal.

(f) Both measures indicate that there is moderate-to-high positive skewness, hence

the median cost of R59 would be a more representative measure of central location.

25% of Durban restaurants charge at least R68 for a veal cordon bleu meal.

25% of Durban restaurants charge no more than R54.75 for a veal cordon bleu meal.

Exercise 3.23 File: X3.23 - fuel bills.xls

fuel bill

(i) Mean Mean 418

Standard Error 13.128

(ii) Median Median 398

Mode 350

(iv) Standard deviation Standard Deviation 113.693

(iii) Variance Sample Variance 12926.054

Kurtosis -0.503

(v) Skewness Skewness 0.601

Range 420

Minimum 256

Maximum 676

Sum 31350

Count 75

The average monthly fuel bill per motorist is R418; while half of the motorists

spend no more than R398 on their monthly fuel.

68.3% of monthly fuel bills range between R304.31 and R531.69 (one std dev from mean)

95,5% of monthly fuel bills range between R190.61 and R645.39 (within 2 std devs of mean)

There is moderate positive skewness (Skp = 0.601) caused by 8 motorists who spend

more than R600 per month - well above the majority of motorists' fuel spend.

There is only moderate consistency across monthly fuel bills of motorists

This greater relative spread could be caused by different size vehicles and varying distances travelled.

Upper Quartile - Q3 R502.5 =QUARTILE(data range,3)

Inter Quartile Range = (Q3 - Q1) R170

The middle 50% of monthly fuel bills span R170 from a low of R332.50 to a high of R502.50.

Lower Quartile Q1 R332.5 =QUARTILE(data range,1)

Median R398 =MEDIAN(data range) or =QUARTILE(data range,2)

Upper Quartile Q3 R502.5 =QUARTILE(data range,3)

Maximum R676 =MAX(data range) or =QUARTILE(data range,4)

(g) Box plot

There is clear evidence of moderate positive skewness (skewed-to-the-right) in fuel bills.

There are a few motorists who spend a large amount on fuel a month.

(i) Total amount of fuel consumed monthly by Paarl motorists for commuting to work:

(average bill / motorist x no.motorists)

Expected Total fuel consumed (in litres) = R10450000/10 = 1 045 000 litres

(Expected total monthly cost / cost per litre)

Exercise 3.24 File: X3.24 - service periods.xls

service periods

(i) Mean Mean 7

Standard Error 0.384

(ii) Median Median 6

Mode 6

(iii) Std Dev Standard Deviation 3.838

Sample Variance 14.727

Kurtosis -0.437

(iv) Skewness Skewness 0.553

Range 15

Minimum 1

Maximum 16

Sum 700

Count 100

On average, each engineer has 7 years of service with a company

Half of all engineers spend no more than 6 years with a company

68.3% of engineers spend between 3.16 and 10.84 years with a company.

Similar intervals can be computed for 2 and 3 std devs from the mean.

Note: when the lower limit is computed to be negative, in practice it is zero.

Interpretation - skewness

There is a moderate positive skewness (Skp = 0.553) in length of service periods,

meaning that a few engineers have long service periods with their company.

0-3 21

3-6 30

6-9 24

9 - 12 15

12 - 15 8

15 - 18 2

(d) Lower limit = 7 - 3.838 = 3,16 years

Upper limit = 7 + 3.838 = 10,84 years

68.3% of engineers spend between 3.16 years and 10.84 years with a company.

Cumulative % up to 3 years = 21/100 = 21%

Cumulative % above 12 years = 10/100 = 10%

They meet the guidelines for "new blood" (21,2%) ,but have

far fewer "experienced" members (only 10%) than their guidelines require.

Exercise 3.25 File: X3.25 - dividend yields.xls

Data type numeric, ratio-scaled

Dividend yields

(i) Mean Mean 4.273

Standard Error 0.218

(ii) Median Median 4.1

(iii) Mode Mode 2.8

(iv) Std Dev Standard Deviation 1.444

Sample Variance 2.084

Kurtosis -0.359

(v) Skewness Skewness 0.259

Range 6.1

Minimum 1.5

Maximum 7.6

Sum 188

Count 44

(viz. 2,8, 3,6; 4,1; 5,1) of equal modal frequency (of 3)

Std dev 68.3% of all dividend yields lie between 2.83% and 5.72%

Similarly for 2 and 3 std devs from the mean

The histogram can be assumed to be normally distributed.

(d) the Mean (Average) as there is minimal skewness present in the data.

Below 2% 3

2 - 3.5% 11

3.5 - 5% 16

5 - 6.5% 11

6.5 - 8% 3

(f) Five-Number Summary Table

Minimum 1.5 =QUARTILE(data range,0)

Lower Quartile Q1 3.175 =QUARTILE(data range,1)

Median 4.1 =QUARTILE(data range,2)

Upper Quartile Q3 5.15 =QUARTILE(data range,3)

Maximum 7.6 =QUARTILE(data range,4)

The very slight positive skewness can be seen from the longer tail

on the right of the box plot. It implies a few companies have achieved

significantly higher dividend yields than the majority of the sample.

(h) Minimum dividend yield achieved by the top performing 10% of companies

6.17% =PERCENTILE(data range,0.90)

The top 10% of companies achieved at least a dividend yield of 6.17%

(i) % of companies who did not declare more than a 3.5% dividend yield

Cumulative frequency up to upper limit of 3.5 = (3 + 11) = 14.

% Cumulative = 14 / 44 % = 31.8%

Almost 1/3 (31.8%) of companies did not declare more than a 3.5% dividend yield.

Exercise 3.26 File: rosebuds.xls

Data type - numerical, continuous, ratio-scaled

selling price

(i) Mean 60.312

Standard Error 0.319

(iii) Median 59.95

Mode 60.6

(ii) Standard Deviation 3.188

Sample Variance 10.160

Kurtosis 1.770

(iv) Skewness 0.932

Range 18.3

Minimum 55.2

Maximum 73.5

Sum 6031.2

Count 100

Interpretation

Mean: The average selling price of rosebuds is 60.312 cents.

Median: 50% of rosebuds sold for no more than 59.95c.

Std dev: 68.3% of rosebud unit prices are likely to lie between 57.12c and 63.5c.

Skewness: There is excessive positive skewness (one very high unit price)

There is very low variability amongst unit selling prices of rose buds

Upper Quartile Q3 62.45 cents =QUARTILE(data range,3)

(e) The highest unit selling price of the cheapest 25% of sales is 57.7 cents (i.e. Q1)

(f) The minimum unit selling price of the highest-priced 25% of sales is 62.45 cents (i.e. Q3)

Overall interpretation of (a), (d), (e) and (f)

Unit selling prices ranged from 55.2c to 73.5c where 50% of the selling prices

were above 59.95c. 25% of sales were below 57.7c while the most expensive 25%

of unit selling prices were above 62.45c.

Lower Quartile Q1 57.7 =QUARTILE(data range,1)

Median 59.95 =QUARTILE(data range,2)

Upper Quartile Q3 62.45 =QUARTILE(data range,3)

Maximum 73.5 =QUARTILE(data range,4)

Boxplot

Interpretation

The distribution of unit selling prices of rosebuds is skewed to the right.

There is also one extremely high unit selling price of 73.5c which can be

considered an outlier.

Overall, unit selling prices of rosebuds ranged between 55.2c and 73.5c.

25% of unit selling prices lay below 57.7 c (i.e. lower quartile); while

the middle 50% of unit selling prices ranged from 57.7c to 62.45c.

The "best" 25% of unit selling prices achieved were above 62.45c (Q3).

Finally, the median (middle) unit selling price achieved was 59.95c.

Mini Case Study 3.27 File: X3.27 - savings balances.xlsx

0 - ≤ 200 21

80

201 - ≤ 400 71 71

401 - ≤ 600 56 70

601 - ≤ 800 19 60 56

Number of Clients

801 - ≤1000 4

More 50

4

40

30

21 19

20

10 4 4

0

0 - ≤ 200 201 - ≤ 400 401 - ≤ 600 601 - ≤ 800 801 - ≤1000 More

Savings Intervals

Mean 421.86 Lower Quartile 266.5

Standard Error 16.247 Upper Quartile 520.5

Median 385

Mode 326

Standard Deviation 214.93

Sample Variance 46195.1765

Kurtosis 4.727

Skewness 1.634

Range 1383

Minimum 85

Maximum 1468

Sum 73826

Count 175

Female 23 41 64

Male 83 28 111

Grand Total 106 69 175

Female 512.6 563.7 545.3

Male 368.0 299.3 350.7

Grand Total 399.4 456.4 421.9

Marital status Categorical Nominal-scaled

Month-end balances Numeric Ratio-scaled

50% of these clients have month-end balances of less than R385 (median).

From the histogram, 21 clients have month-end balances of less than R200 and 8 are above R800.

The 8 banking clients with relatively high month-end savings balances (skewness = 1.643 > 1.0) skew the average

towards these higher balances and distorts the overall picture.

Therefore the median month-end balance of R385 is a more representative indicator of savings balances.

25% of their clients have month-end balances of less than R266.6, while the top 25% of savers

have month-end balances in excess R520.5 with 4 clients above R1000 (maximum = R1468).

Amongst the female clients, the majority (64%) are single (41/64), while amongst the male clients,

the majority (75%) are married.

Thus they are attracting mainly single females; and married males.

Refer to 1 (d) (Breakdown Table of Average Savings by Gender and Marital Status)

Single females are saving the most (average of R563,7) while single males are the worst savers with

an average savings balance of only R299.3 (compared to the overall average of R421.9).

Also females save more, on average (R545) than males (R350).

Single bank clients save more, on average (R456.4) than married bank clients (R399.4).

2 (c) Plan of Action by Bank

Bank should target females in general (as they comprise only 37%) of the current client base

but have the largest average balances of all clients.

Single females (23%) in particular should be targetted to attract more high savers to the bank.

They should encourage their main client base - married males - to save more.

Mini Case Study 3.28 File: X3.28 - medical claims.xlsx

Below 0.1 6 4.0

40 35 36

0.1 - 0.5 35 23.3

35 30

0.5 - 0.9 26 17.3

No. of Members

30 26

0.9 - 1.3 36 24.0

25

1.3 - 1.7 30 20.0

20

1.7 - 2.1 14 9.3 14

15

2.1 - 2.5 2 1.3 10 6

Above 2.5 1 0.7 5 2 1

150 100 0

Claims Ratio

Mean 0.974 Lower Quartile 0.460

Median 0.996 Upper Quartile 1.421

Mode #N/A

Standard Deviation 0.594

Sample Variance 0.353

Skewness 0.242

Range 2.672

Minimum 0.013

Maximum 2.685

Sum 146.1

Count 150

Age

Marital 26 - 35 36 - 45 46 - 55 Row totals

Married average 1.326 0.915 1.123 1.112

std dev 0.552 0.589 0.629 0.610

minimum 0.147 0.051 0.013 0.013

maximum 2.685 1.989 2.408 2.685

count 23 27 35 85

Single average 0.919 0.474 0.947 0.793

std dev 0.543 0.356 0.500 0.524

minimum 0.014 0.024 0.218 0.014

maximum 1.905 1.262 1.814 1.905

count 36 19 10 65

average 1.078 0.733 1.084 0.974

Column totals std dev 0.577 0.547 0.602 0.594

minimum 0.014 0.024 0.013 0.013

maximum 2.685 1.989 2.408 2.685

count 59 46 45 150

Age Group (bands) Qualitative / Categoric Ordinal

Claims Ratio Quantitative / Numeric Ratio, Continuous

The average claims ratio is 0.974. This means that, on average, members are claiming

as much as they contribute. The median claims ratio is a similar value (0.996).

This means that at least half the members are claiming more than they contribute.

The distribution (see histogram) appears to be bi-modal showing a low claims ratio (i.e. less than half their contributions)

by at least a quarter of members (Q1 = 0.46) and a high claims ratio by at least half of the

members (50% greater than median claims ratio of 0.996). At least 25% of members are

claiming significantly more than they contribute (Q3 = 1.42).

Only 4% of members claim less than 10% of their contributions (first interval of histogram).

Also at least 11% (top three intervals of histogram) are claiming at least twice as much as they contribute.

It would appear that married members claim significantly more (1.112) than single members (0.793).

Married members also make up the majority of members (85/150 = 57%).

Also, the younger members (26 - 35) and older members (46 - 55) claim significantly more

(1.078 and 1.084 respectively) than the middle-aged members (36 - 45) (0.733), on average.

These two 'high-claiming' age segments of members represent 69% (104/150) of all members.

The highest claiming segments are the married younger (26-35) (1.326) and married older members (46 - 55) (1.123).

Members from these two segments also have the maximum claims ratio of all members (2.685 and 2.408 respectively).

Collectively these members represent 39% (58/150) of all members of the medical scheme.

The lowest claiming segment is the middle-aged single members (0.474) but they only

comprise 13% (19/150) of all scheme members.

The range of claims ratios is highest amongst married (25-35 years) (0.147 - 2.685) and married (46 - 55 years) (0.013 - 2.408).

It is lowest amongst single (36 - 45 year) members (0.024 - 1.262).

In conclusion:

Younger and Older married members claim more on average - and above their contribution levels in most cases).

They also make up the bulk of membership (57%).

The single members tend to claim less than the married members, but only make up 43% of the membership base.

2 (c) There is cause for concern about the financial viability of the Scheme because:

The claims ratio of married members exceed 1 on average

Within the married group, the age groups 26 -35 and 46 - 55 are claiming more than they contribute

The married group represent the majority of members (57%)

The two age groups within the married group represent 69% of all married (and are claiming more than they contribute).

Overall the scheme is operating close to its breakeven of financial non-viability (overall mean claims ratio = 0.974)

There is significant cross-subsidization of the married younger and older members by mostly

the middle-aged single members.

The mangement of this medical scheme needs to review member contributions from the married younger and older members.

CHAPTER 4

Exercise 4.1 P(A) = 0.2 means that an event has a 20% chance of occurring.

Exercise 4.3 The outcome of one event does not influence / nor is influenced

by / the outcome of the other event.

Exercise 4.4 P(A or B) = P(A) + P(B) - P(A ∩ B) = 0.26 + 0.35 - 0.14 = 0.47

P(Y / X) = P(Y ∩ X) / P(X) = 0.27 / 0.54 = 0.50

Thus P(X / Y) ≠ P(Y / X)

Exercise 4.6 File: X4.6 - economic sectors.xls

Mining 45 18.0

Financial 72 28.8

IT 32 12.8

Production 101 40.4

Total 250 100

In (c), used the Complementary probability rule

In (d), used the Addition rule for mutually exclusive events.

Exercise 4.7 File: X4.7 - apple grades.xls

A 795 53.0

B 410 27.3

C 106 7.1

D 189 12.6

Total 1500 100

In (c), used the Addition rule for mutually exclusive events.

In (d), used the Complementary probability rule.

Exercise 4.8 File: X4.8 - employment sectors.xls

Formal Business 6678 53.0

Commercial Agriculture 1492 11.8

Subsistence Agriculture 653 5.2

Informal Business 2865 22.7

Domestic Service 914 7.3

Total 12602 100

In (c), the Addition rule for mutually exclusive events.

In (d), the Conditional Probability rule.

Exercise 4.9 File: X4.9 - qualification levels.xls

Random variable 2 Qualification level (categorical, ordinal-scaled and discrete)

Qualification Section Head Dept Head Division Head Total

Matric 28 14 8 50

Diploma 20 24 6 50

Degree 5 10 14 29

Total 53 48 28 129

(c)

(i) P(Matric) = 50/129 = 38.76%

(ii) P(Section head ∩ Degree) = 5/129 = 3.88%

(iii) P(Dept head / Diploma) = 24/50 = 48%

(iv) P(Division head) = 28/129 = 21.71%

(v) P(Division head U Section head) = (53+28)/129 = 62.79%

(vi) P(Matric U Diploma U Degree) = (50+50+29)/129 = 100%

(vii) P(Degree / Dept head) = 10/48 = 20.83%

(viii) P(Division head U Diploma U both) = (28+50-6)/129 = 55.81%

(i) Marginal probability

(ii) Joint probability and Multiplication rule

(iii) Conditional probability

(iv) Marginal probability

(v) Addition rule for mutually exclusive events

(vi) Collectively exhaustive set of events and Addition rule for mutually exclusive events

(vii) Conditional probability

(viii) Addition rule for non-mutually exclusive events

Exercise 4.10 File: X4.10 - bonus options.xls

Admin 28 44 68 140

Production 56 75 29 160

Total 84 119 97 300

(b) P(Share option) = 97/300 = 32.33% Marginal probability

(c) P(Production ∩ Cash bonus) = 56/300 = 18.67% Joint probability

(d) P(Share option / Admin) = 68/140 = 48.57% Conditional probability

(e) P(Production / Cash bonus) = 56/84 = 66.67% Conditional probability

P(A) = 97/300 = 32.33%

Since P(A/B) ≠ P(A), the two events are statistically dependent.

Conclusion: The choice of bonus option and employee work function are associated.

Exercise 4.11 File: X4.11 - age profile.xls

Department

Age Production Sales Administration Total

<30 60 25 18 103

30-50 70 29 25 124

>50 30 8 35 73

Total 160 62 78 300

(ii) P(Production) = 160/300 = 53.33% Marginal probability

(iii) P(Sales ∩ (30-50) years) = 29/300 = 9.67% Joint probability

(iv) P(>50 / Admin) = 35/78 = 44.87% Conditional probability

(v) P(Production U <30 U both) = (160+103-60)/300 = 67.67% Conditional probability

(b) No, since outcomes of these two events can occur simultaneously.

Is P(A/B) = P(A) ? P(A/B) = 35/78 = 44.87%

P(A) = 73/300 = 24.33%

Since P(A/B) ≠ P(A), the two events are statistically dependent.

Conclusion: There is an association between the Age of an employee and the Department

in which they are employed (e.g. younger employees tend to be in Production).

Exercise 4.12 File: X4.12 - digital cameras.xls

Usage Canon Nikon Pentax Total

Professional 48 15 27 90

Personal 30 95 65 190

Total 78 110 92 280

(b) P(Nikon User) = 110/280 = 39.29% Marginal probability

(c) P(Pentax / Personal) = 65/190 = 34.21% Conditional probability

P(A/B) = P(A) ? P(A/B) = 48/78 = 61.54%

P(A) = 90/280 = 32.14%

Since P(A/B) ≠ P(A), the two events are statistically dependent.

Conclusion: Type of usage and choice of brand are associated.

(e.g. Professionals may prefer Canon while Nikon is favoured by Personal users).

Addition rule for non-mutually exclusive events

(g) No, as outcomes of the two events can occur simultaneously as illustrated in (f) above.

Exercise 4.13

P(F1) = 0.20

(Fail)

(Not fail) P(S2) = 0.85 P(F1 ∩ S2) = 0.20 x 0.85 = 0.17

(Not fail)

P(S1) = 0.80

1.00

Exercise 4.13

P(A ∩ B) = P(A) x P(B) = 0.2 * 0.15 = 0.03 (3% chance)

P(A U B) = P(A) + P(B) - P(A∩B) = 0.2+0.15-0.03 0.32

Hence P(not Fail) = 1 - P(A U B) = 1 - 0.32 = 0.68 (68% chance)

Exercise 4.14

P(Y) = 0.30

(Attend)

(Fail) P(F) = 0.20 P(Y ∩ F) = 0.30 x 0.20 = 0.06

(Not attend)

P(N) = 0.70

1.00

Exercise 4.14

(b) Refer to the end nodes of the Probability Tree to answer question 4.14 (b)

(i) P(Pass and Attended workshop) = P(Y ∩ S) = 0.30 x 0.8 = 0.24 24%

P(S ∩ N) = 0.70 x 0.60 = 0.42

Exercise 4.15

Using Excel

(ii) 3! 5! = (3.2.1)(5.4.3.2.1) = 720 =FACT(3)*FACT(5)

(iii) 4! 2! 3! = (4.3.2.1)(2.1)(3.2.1) = 288 =FACT(4)*FACT(2)*FACT(3)

(iv) 7C4 = 7!/(4!*(7-4)!) = 35 =FACT(7)/(FACT(4)*FACT(3))

(v) 9C6 = 9!/(6!*(9-6)!) = 84 =FACT(9)/(FACT(6)*FACT(3))

(vi) 8P3 = 8!/(8-3)!) = 336 =FACT(8)/FACT(5)

(vii) 5P2 = 5!/(5-2)!) = 20 =FACT(5)/FACT(3)

(viii) 7C7 = 7!/(7!*(7-7)!) = 1 =FACT(7)/(FACT(7)*FACT(0))

(ix) 7P4 = 7!/(7-4)!) = 840 =FACT(7)/FACT(3)

Likely scenarios

(ii) Number of ways of arranging the seating plan of 8 persons,

consisting of 3 males and 5 females.

(iii) Number of sequences of visiting 9 stores, consisting of 4 clothing stores,

2 home décor stores and 3 coffee shops.

(iv) Selecting all combinations of 4 holiday destinations from a possible 7 destinations.

Similar for (v) and (viii)

(vi) Selecting a committee of 3 persons from 8 candidates, where the first person

selected is the chairman, the second is the secretary and the third is a member.

Similar for (vii) and (ix)

Exercise 4.16

Exercise 4.17

Exercise 4.18

=FACT(9)/(FACT(4)*FACT(5)) (Using Excel )

Exercise 4.19

1/120 = 0.00833 (0,833% chance)

Exercise 4.20

of 3 tourist attractions from 10 options.

=FACT(10)/(FACT(3)*FACT(7)) (Using Excel )

(b) P(a given combination of 3 out of 10) = 1/120 = 0.0083 (0,833% chance)

Exercise 4.21

=FACT(4)/(FACT(2)*FACT(2))*FACT(7)/(FACT(4)*FACT(3))

(Using Excel )

=2*FACT(4)/(FACT(2)*FACT(2))*FACT(7)/(FACT(4)*FACT(3))

(Using Excel )

Exercise 4.22 Project Scoping Study Bayes Theorem

L Late NS No scope change

P(T) = 0.7

On time P(NS/T) = 0.6 P(NS and T) = 0.42

P(L) 0.3

P(NS/L) = 0.2 P(NS and L) = 0.06

Bayes Application Using the Joint Probabilities from the Probability Tree

Find P(T/S) = P(T and S)/P(S) = Posterior Probability

(ii) P(T and S) = 0.28

---ooOoo---

Additional Information

S NS

T 0.28 0.42 0.7 P(T/S) = P(T and S)/P(S)

L 0.24 0.06 0.3 =0.28/(0.28+0.24)

0.52 0.48 1 0.5385

---ooOoo---

Exercise 4.23 Married Couples Sporting Habits Study Bayes Theorem

HN Husband does not play sport WNS Wife does not play sport

P(HS) = 0.6

Husband P(WNS/HS) = 0.6 P(HS and WNS) = 0.36

plays sport

Husband does

not play sport P(WNS/HNS) = 0.7 P(HNS and WNS) = 0.28

Bayes Application Using the Joint Probabilities from the Probability Tree

Find P(HS/WS) = P(HS and WS)/P(WS) = Posterior Probability

(ii) P(HS and WS) = 0.24

There is a 66,67% chance that a husband plays sport if the wife also plays sport.

---ooOoo---

Additional Information

WS WNS

HS 0.24 0.36 0.6 P(HS/WS) = P(HS and WS)/P(WS)

HNS 0.12 0.28 0.4 =0.24/(0.24+0.12)

0.36 0.64 1 0.6667

---ooOoo---

Exercise 4.24 Airline Departure Times Study Bayes Theorem

B Airline B L Leaves Late

P(A) = 0.6

A P(L/A) = 0.2 P(A and L) = 0.12

P(B) = 0.4

P(L/B) = 0.35 P(B and L) = 0.14

Bayes Application Using the Joint Probabilities from the Probability Tree

Find P(A/T) = P(A and T)/P(T) = Posterior Probability

(ii) P(A and T) = 0.48

There is a 64.86% chance that it is Airline A, if the aircraft that has just left, left on time.

---ooOoo---

Additional Information

T L P(A/T) = P(A and T)/P(T)

A 0.48 0.12 0.6 =0.48/(0.48+0.26)

B 0.26 0.14 0.4 0.6486

0.74 0.26 1

---ooOoo---

Exercise 4.25 New Business Venture Study Bayes Theorem

NG NBV started by non-Graduate F Failure

P(G) = 0.6

Graduate P(F/G) = 0.2 P(G and F) = 0.12

Non

graduate P(S/NG) = 0.65 P(NG and S) = 0.26

P(NG) = 0.4

P(F/NG) = 0.35 P(NG and F) = 0.14

Bayes Application Using the Joint Probabilities from the Probability Tree

Find P(G/F) = P(G and F)/P(F) = Posterior Probability

(ii) P(G and F) = 0.12

There is a 46.15% chance that a NBV was started by a Graduate given that it has failed.

---ooOoo---

Additional Information

S F

G 0.48 0.12 0.6 P(G/F) = P(G and F)/P(F)

NG 0.26 0.14 0.4 =0.12/(0.12+0.14)

0.74 0.26 1 0.4615

---ooOoo---

Exercise 4.26 On-line Airline Tickets Purchase Study Bayes Theorem

NE non e-Ticket purchase NB non-Business Traveller

P(E) = 0.4

e-Ticket P(NB/E) = 0.2 P(E and NB) = 0.08

non-

e-Ticket P(B/NE) = 0.45 P(NE and B) = 0.27

P(NE) = 0.6

P(NB/NE) = 0.55 P(NB and NB) = 0.33

1

Bayes Application Using the Joint Probabilities from the Probability Tree

Find P(E/B) = P(E and B)/P(B) = Posterior Probability

(ii) P(E and B) = 0.32

There is a 54.24% chance that an e-ticket was bought given that it was bought by a business traveller.

---ooOoo---

Additional Information

B NB

E 0.32 0.08 0.4 P(E/B) = P(E and B)/P(B)

NE 0.27 0.33 0.6 =0.32/(0.32+0.27)

0.59 0.41 1 0.5424

---ooOoo---

CHAPTER 5

PROBABILITY DISTRIBUTIONS

Poisson probability distribution

(b) discrete (e.g. 132 employees)

(c) discrete (e.g. 7046 households)

(d) continuous (e.g. 514.68 km)

Exercise 5.3

3 (7-3)

(a) (i) n = 7; p = 0,2; x = 3 P(x = 3) = 7C3 (0.2) (0.8) =

0.1147 11.47%

4 (10-4)

(a) (ii) n = 10; p = 0,2; x = 4 P(x = 4) = 10C4 (0.2) (0.8) =

0.0881 8.81%

0.01384+0.07118+0.16779+0.2397+0.23114 = 0.7237 72.37%

0.07464+0.01048 = 0.0851 8.51%

1 - (0.10011+0.26697+0.31146) = 0.3215 32.15%

(b) (ii) =BINOMDIST(4,10,0.2,0) 0.0881

(b) (iii) =BINOMDIST(4,12,0.3,1) 0.7237

(b) (iv) =BINOMDIST(2,10,0.05,0)+BINOMDIST(3,10,0.05,0) 0.0851

(b) (v) =1-BINOMDIST(2,8,0.25,1) 0.3215

Exercise 5.4

There are only two possible outcomes (in-stock; out-of-stock)

This outcome is observed 6 times (n = 6 stores)

The probability of observing the "out-of-stock" outcome, p = 0.20, is constant.

The stores (trials) are independent of each other

0.2621 + 0.3932 + 0.2458 = 0.9011 90.11% =BINOMDIST(2,6,0.2,1)

On average, 1,2 stores out the 6 stores surveyed can be expected to be out of stock in a week.

Exercise 5.5

n = 12 p = 0.15

P(x = 0) = 12C0 (0.15)0 (0.85)(12-0) = 0.1422 =BINOMDIST(0,12,0.15,0)

(b) n = 15 p = 0.15

P(x < 3) = P(x = 0) + P(x = 1) + P(x = 2) =

= 0.0874 + 0.2312 + 0.2857 = 0.6042 =BINOMDIST(2,15,0.15,1)

Exercise 5.6

n = 10 p = 0.30 (probability of preferring the deluxe model)

P(x = 3) = 10C3 (0.30)3 (0.70)(10-3) = 0.2668 =BINOMDIST(3,10,0.3,0)

P(x > 2) = 1 - P(x ≤ 2) = 1 - (P(x = 0) + P(x = 1) + P(x = 2)) =

= 1 - (0.0000059 + 0.000138 + 0.00145) = 0.9984 =1-BINOMDIST(2,10,0.7,1)

Exercise 5.7

n=8 p = 0.05 (probability of a defective Tata truck)

P(x = 1) = 8C1 (0.05)1 (0.95)(8-1) = 0.2793 =BINOMDIST(1,8,0.05,0)

P(x ≤ 2) = P(x = 0) + P(x = 1) + P(x = 2) =

= 0.66342 + 0.27934 + 0.05146 = 0.9942 =BINOMDIST(2,8,0.05,1)

P(x = 0) = 8C0 (0.05) (0.95)(8-0) = 0.6634 =BINOMDIST(0,8,0.05,0)

0

Based on 64 sales, the dealer can expect 3.2 trucks to be returned for assemby defective repairs

Exercise 5.8

n=6 p = 0.80 (probability of a UT out-performing the JSE Index)

P(x = 6) = 6C6 (0,80)6 (0,20)(6-6) = 0.2621 =BINOMDIST(6,6,0.8,0)

P(x = (2 U 3)) = P(x = 2) + P(x = 3) =

= 0.01536 + 0.08192 = 0.0973

=BINOMDIST(2,6,0.8,0)+BINOMDIST(3,6,0.8,0)

P(x ≤ 2) = P(x = 0) + P(x = 1) + P(x = 2) =

= 0.26214 + 0.39322 + 0.24576 = 0.9011 =BINOMDIST(2,6,0.2,1)

Exercise 5.9

n = 12 p = 0.20 (probability of a person participating in a focus group)

P(x = 2) = 12C2 (0.20)2 (0.80)(12-2) = 0.2835

=BINOMDIST(2,12,0.2,0)

P(x = 5) = 12C5 (0.20) (0.80)(12-5) = 0.5320

5

=BINOMDIST(5,12,0.2,0)

P(x ≥ 6) = 1 - P(x ≤ 5) = 1 - (P(x=0)+P(x=1)+P(x=2)+P(x=3)+P(x=4)+P(x=5)) =

= 1 - (0.0687 + 0.2062 + 0.2835 + 0.2362 + 0.1329 + 0.0532) = 0.0194

=1 - BINOMDIST(5,12,0.2,1)

Exercise 5.10

n = 10 p = 0.10 (probability of a general population person being a heavy reader )

P(x < 2) = P(x=0)+P(x=1) = 10C0(0.10)0(0.90)(10-0) + 10C1(0.10)1(0.90)(10-1)

= 0.34868 + 0.38742 = 0.7361

P(x < 2) = P(x=0)+P(x=1) = 10C0(0.35)0(0.65)(10-0) + 10C1(0.35)1(0.65)(10-1)

= 0.01346 + 0.07249 = 0.0860

(b) n = 280 p = 0.65 (probability of a pensioner person not being a heavy reader )

Expected number (mean) of "non-heavy" readers = 280 x 0.65 = 182

Exercise 5.11

P(x = 5 / a = 3) = e-3 35 / 5! = 0.1008 10.08% =POISSON(5,3,0)

-3 0 -3 1 -3 2 -3 3

= 1 - (e 3 / 0! +e 3 / 1! + e 3 / 2! + e 3 / 3!)

= 1 - (0.04979 + 0.14936 + 0.22404 + 0.22404)

= 1 - 0.64723 = 0.3528 35.28% =1-POISSON(3,3,1)

Exercise 5.12

P(x ≤ 2 / a = 4) = P(x=0)+P(x=1)+P(x=2)

-4 0 -4 1 -4 2

= e 4 / 0! +e 4 / 1! + e 4 / 2!

= 0,01832 + 0,07326 + 0,14653 = 0.2381 23.81% =POISSON(2,4,1)

= 1 - (0,2381 (from (i) above) + e-3 43 / 4!)

= 1 - (0,2381 + 0,19537)

= 1 - 0,43347 = 0.5665 56.65% =1-POISSON(3,4,1)

Exercise 5.13

P(x = 1 / a = 6) = e-6 61 / 1! = 0.0149 1.49% =POISSON(1,6,0)

-6 0 -6 1 -6 2 -6 3

= e 6 / 0! +e 6 / 1! + e 6 / 2! + e 6 / 3!

= 0,00248 + 0,01487 + 0,04462 + 0,08924 0.1512 15.12%

=POISSON(3,6,1)

(a) (iii) P(x ≥ 3 / a = 6) = 1 - P(x ≤ 2) = 1 - (P(x=0)+P(x=1)+P(x=2))

= 1 - (e-6 60 / 0! +e-6 61 / 1! + e-6 62/ 2!)

= 1 - (0,00248 + 0,01487 + 0,04462)

= 1 - 0,06197 = 0.938 93.80% =1-POISSON(2,6,1)

(b) Note: a = 3 since the mean orders must refer to a given half-day interval (i.e. 6/2 = 3)

-3 1

P(x = 1 / a = 3) = e 3 / 1! = 0.1494 14.94% =POISSON(1,3,0)

Exercise 5.14

P(x ≥ 3 / a = 1.8) = 1 - P(x ≤ 2) = 1 - (P(x = 0) + P(x = 1) + P(x = 2))

-1.8 0 -1.8 1 -1.8 2

= 1 - (e 1.8 / 0! + e 1.8 / 1! + e 1.8 / 2!)

= 1 - (0.1653 + 0.29754 + 0.26778)

= 1 - 0.7306 = 0.2694 26.94% =1-POISSON(2,1.8,1)

= e-1.8 1.80 / 0! + e-1.8 1.81 / 1! + e-1.8 1.82/ 2! + e-1.8 1.83/ 3!

= 0.1653 + 0.29754 + 0.26778 + 0.1607 = 0.8913 89.13%

=POISSON(3,1.8,1)

Exercise 5.15

P(x ≤ 5 / a = 7) = P(x = 0) + P(x = 1) + P(x = 2) + P(x = 3) + P(x = 4) + P(x = 5)

= e-7 70 / 0! + e-7 71 / 1! + e-7 72/ 2! + e-7 73/ 3! + e-7 74 / 4! + e-7 75 / 5!

= 0.00091 + 0.00638 + 0.02234 + 0.05213 + 0.09123 +0.12772 =

0.3007 30.07% =POISSON(5,7,1)

= e-7 76 / 6! + e-7 79 / 9! = 0.1490 + 0.1014 = 0.2504 25.04%

=POISSON(6,7,0)+POISSON(9,7,0)

(c) Note: a = 14 since the time interval is a given two-day period (i.e. 7 x 2 = 14)

P(x > 20 / a = 14) = 1 - P(x ≤ 20)

= 1 - (P(x = 0) + P(x = 1) + P(x = 2) + …… + P(x = 19) + P(x = 20))

Use Excel only.

=1 - POISSON(20,14,1) = = 1 - 0.9521 = 0.0479 4.79%

Exercise 5.16

(a)

(i) P(0 < z < 1.83) = 0.4664

(ii) P(z > -0.48) = 0.5 + P(0 < z < 0.48) = 0.5 + 0.1844 = 0.6844

(iii) P(-2.25 < z < 0) = P(0 < z < 2.25) = 0.4878

(iv) P(1.22 < z ) = 0.5 - P(0 < z < 1.22) = 0.5 - 0.3888 = 0.1112

(v) P(-2.08< z < 0.63) = P(0 < z < 2.08)+P(0 < z < 0.63) = 0,4812 + 0,2357 = 0.7169

(vi) P(z < -0.68) = 0.5 - P(0 < z < 0.68) = 0,5 - 0,2517 = 0.2483

(vii) P(0.33 < z < 1.5) = P(0 < z < 1.5) - P(0 < z < 0.33) = 0.4332 - 0.1293 = 0.3039

(i) =NORMSDIST(1.83)-0.5 0.4664

(ii) =1-NORMSDIST(-0.48) 0.6844

(iii) =0.5-NORMSDIST(-2.25) 0.4878

(iv) =1-NORMSDIST(1.22) 0.1112

(v) =NORMSDIST(0.63)-NORMSDIST(-2.08) 0.7169

(vi) =NORMSDIST(-0.68) 0.2483

(vii) =NORMSDIST(1.5)-NORMSDIST(0.33) 0.3039

Exercise 5.17

(i) P(z < ?) = 0.9147 0.9147 - 0.5 = 0.4147 1.37

(ii) P(z > ?) = 0.5319 0.5319 - 0.5 = 0.0319 -0.08

Since area to right of z is greater than 0.5, z must be negative.

(iii) P(0 < z < ?) = 0.4015 0.4015 1.29

(iv) P(? < z < 0) = 0.4803 0.4803 -2.06

(v) P(? < z ) = 0.0985 0.5 - 0.0985 = 0.4015 1.29

(vi) P(z < ?) = 0.2517 0.5 - 0.2517 = 0.2483 -0.67

Since area to left of z is less than 0.5, z must be negative.

(vii) P(? < z ) = 0.6331 0.6331 - 0.5 = 0.1331 -0.34

Since area to right of z is greater than 0.5, z must be negative.

(i) =NORMSINV(0.9147) 1.3703

(ii) =NORMSINV(1-0.5319) -0.0800

(iii) =NORMSINV(0.5+0.4015) 1.2901

(iv) =NORMSINV(0.5-0.4803) -2.0600

(v) =NORMSINV(1-0.0985) 1.2901

(vi) =NORMSINV(0.2517) -0.6691

(vii) =NORMSINV(1-0.6331) -0.3401

Exercise 5.18

(a)

(i) P(x < 62) = P(z < (62-64)/2.5) = P(z < -0.08) = 0.5 - 0.2881 = 0.2119

(ii) P(x > 67.4) = P(z > (67.4-64)/2.5) = P(z > 1.36) = 0.5 - 0.4131 = 0.0869

(iii) P(59.6 < x < 62.8) = P((59.6-64)/2.5 < z < (62.8-64)/2.5) = P(-1.76 < z < -0.48) =

0.4608 - 0.1844 = 0.2764

(iv) P(x > ?) = 0.1026 Look up area = 0.3974 giving z ≈ 1.267

then x = 64+1.267*2.5 = 67.168 min

(v) P(x > ?) = 0.9772 Look up area = 0.9772 - 0.5 = 0.4772 giving z = -2.00

then x = 64-2.00*2.5 = 59 min

(vi) P(60.2 < x < ?) = 0.6652 Find P(60.2 < x < 64) = P(-1.52 < z < 0)

giving area = 0.4357.

Look up area = 0.6652 - 0.4357 = 0.2295 giving z ≈ 0.611

then x = 64+0.611*2.5 = 65.528 min

(ii) =1-NORMDIST(67.4,64,2.5,1) 0.0869

(iii) =NORMDIST(62.8,64,2.5,1)-NORMDIST(59.6,64,2.5,1) 0.2764

(iv) =NORMINV(1-0.1026,64,2.5) 67.167

(v) =NORMINV(1-0.9772,64,2.5) 59.002

(vi) =NORMDIST(60.2,64,2.5,1)+0.6652 0.729455

Apply area = 0.729455 to '=NORMINV(0.729455,64,2.5) 65.528

Exercise 5.19 Gym Attendance Duration

(a) P(x > 120) = P(z > (120-80)/20) = P(z > 2) = 0.5 - 0.4772 = 0.0228 2.28%

(b) P(x < 60) = P(z (60-80)/20) = P(z < -1) = 0.5 - 0.3413 = 0.1587 15.87%

(c) P(x > k) = 0.60 Look up area = 0.60 - 0.50 = 0.10 giving z ≈ -0.253

Then x = 80 - 0.253*20 ≈ 74.94 min

Exercise 5.20 Automatic Washing Machine Lifespan

(a) P(x < 1) = P(z < (1-3.1)/1.1) = P(z < -1.9091) = 0.5 - 0.4719 = 0.0281 2.81%

(b) (i) P(x > 4) = P(z > (4-3.1)/1.1) = P(z > 0.819) = 0.5 - 0.2939 = 0.2061 20.61%

(b) (ii) P(x > 5.5) = P(z > (5.5-3.1)/1.1) = P(z > 2.182) = 0.5 - 0.4854 = 0.0146 1.46%

(c) P(x < k) = 0.05 Look up area = 0.50 - 0.05 = 0.45 giving z ≈ -1.645

Then x = 3.1 - 1.645*1.1 ≈ 1.29 years

Exercise 5.21 Household Daily Water Usage

(a)

(i) P(x > 300) = P(z > (300-220)/45) = P(z > 1.778) = 0.5 - 0.4625 = 0.0375 3.75%

(ii) P(x < 100) = P(z < (100-220)/45) = P(z < -2.667) = 0.5 - 0.4962 = 0.0038 0.38%

(iii) P(x < k) = 0.15 Look up area = 0.50 - 0.15 = 0.35 giving z ≈ -1.038 (approx)

Then x = 220 - 1.038*45 ≈ 173.29 litres

(iv) P(x > k) = 0,20 Look up area = 0.50 - 0.20 = 0.30 giving z ≈ 0.842 (approx)

Then x = 220 + 0.842*45 ≈ 257.89 litres

(a) (ii) =NORMDIST(100,220,45,1) 0.0038

(a) (iv) =NORMINV(1-0.2,220,45) 257.87 litres

Exercise 5.22 Long-distance Truck Drivers' Reaction Times

(a) Normal distribution x ≡ N(1.4; 0.25) Variance = 0.0625 Std dev = 0.25

(i) P(x > 2) = P(z > (2-1.4)/0.25) = P(z > 2.4) = 0.5 - 0.4918 =

0.0082 0.82%

(ii) P(1.2 < x < 1.4) = P((1.2-1.4)/0.25 < z < (1.4-1.4)/0.25) = P(-0.80 < z < 0) =

0.2881 28.81%

(iii) P(x < 0.9) = P(z < (0.9-1.4)/0.25) = P(z < -2.00) = 0.5 - 0.4772 =

0.0228 2.28%

(iv) P(0.5 < x < 1.0) = P((0.5-1.4)/0.25 < z < (1.0-1.4)/0.25) = P(-3.6

0.49984

< z <--1.6)

0.4452

= =

0.0546 5.46%

(b) P(x > 1.8) = P(z > (1.8-1.4)/0.25) = P(z > 1.6) = 0.5 - 0.4452 =

0.0548 5.48%

No. of truck drivers = 120 * 0,0548 = 6,576 drivers 7 drivers (approx)

(c) P(x > 1.7) = P(z > (1.7-1.4)/0.25) = P(z > 1.2) = 0.5 - 0.3849 =

0.1151 11.51%

No. of truck drivers = 360 * 0.1151 = 41.436 drivers 42 drivers (approx)

(ii) =0.5-NORMDIST(1.2,1.4,0.25,1) 0.2881 28.81%

(iii) =NORMDIST(0.9,1.4,0.25,1) 0.0228 2.28%

(iv) =NORMDIST(1,1.4,0.25,1)-NORMDIST(0.5,1.4,0.25,1) 0.0546 5.46%

truck drivers 6.576 7

truck drivers 41.425 42

Exercise 5.23 Hair dye Containers Mass

(a) P(x < 18) = P(z < (18-18.2)/0.7) = P(z < -0.2857) = 0.5 - 0.1124 (approx) = 0.3876

or 38.76%

(b) P(x > k) = 0.15 Look up area = 0.50 - 0.15 = 0.35 giving z ≈ 1.038 (approx)

Then x = 18.2 + 1.038 * 0.7 ≈ 18.927 gm

(b) =NORMINV(0.85,18.2,0.7) 18.926 gm

Exercise 5.24 Motor Vehicle Service Time

(a) P(x < 60) = P(z < (60-70)/9) = P(z < -1.11) = 0.5 - 0.3665 = 0.1335 13.35%

(b) P(x > 90) = P(z > (90-70)/9) = P(z > 2.22) = 0.5 - 0.4868 = 0.0132 1.32%

(c) P(50 < x < 60) = P((50-70)/9 < z < (60-70)/9) = P(-2.22 < z < -1.11) =

0.4868 - 0.3665 = 0.1203 12.03%

(d) P(x > 80) = P(z > (80-70)/9) = P(z > 1.11) = 0.5 - 0.3665 = 0.1335 13.35%

No. of customers = 0.1335 * 80 = 10.68 customers

(e) P(z > (80-μ)/9) = 0.05 Look up area = 0.50 - 0.05 = 0.45 giving z ≈ 1.645

Then 1.645 = (80 - μ)/9 giving μ = 80 - 1.645 * 9 = 65.195 min

(e) This answer cannot be computed from the Excel function NORMDIST.

Exercise 5.25 Coffee Dispensing Machine - Cup Fill

(a)

(i) P(x > 235) = P(z > (235-230)/10) = P(z > 0.5) = 0.5 - 0.1915 =

0.3085 30.85%

(ii) P(235 < x < 245) = P((235-230)/10 < z < (245-230)/10) = P(0.5 < z < 1.5) =

0.4332 - 0.1915 =

0.2417 24.17%

(iii) P(x < 220) = P(z < (220-230)/10) = P(z < -1.00) = 0.5 - 0.3413 =

0.1587 15.87%

(b) P(x > k) = 0.15 Look up area = 0.50 - 0.15 = 0.35 giving z ≈ 1.038 (approx)

Then x = 230 + 1.038 * 10 = 240.38 ml

(c) P(z < (220-μ)/10) = 0.10 Look up area = 0.50 - 0.10 = 0.40 giving z ≈ -1.282

Then -1.282 = (220 - μ)/10 giving μ = 220 + 1.282 * 10 = 232.82 ml

(a) (ii) =NORMDIST(245,230,10,1)-NORMDIST(235,230,10,1) 0.2417 24.17%

(a) (iii) =NORMDIST(220,230,10,1) 0.1587 15.87%

Exercise 5.26 Car Battery Lifespan

(a) P(30 < x < 34) = P((30-28)/4 < z < (34-28)/4) = P(0.50 < z < 1.50) =

0.4332 - 0.1915 = 0.2417

or 24.17%

(b) P(x < 24) = P(z < (24-28)/4) = P(z < -1.00) = 0.5 - 0.3413 = 0.1587

or 15.87%

(c) P(x > k) = 0.60 Look up area = 0.60 - 0.50 = 0.10 giving z ≈ -0.253 (approx)

Then x = 28 - 0.253 * 4 ≈ 26.988 months

(d) P(z < (x-28)/4) = 0.05 Look up area = 0.50 - 0.05 = 0.45 giving z ≈ -1.645

Then -1.645 = (x - 28)/4 giving x = 28 - 1.645 * 4 = 21.42 months

(b) =NORMDIST(24,28,4,1) 0.1587 15.87%

(c) =NORMINV(0.4,28,4) 26.987 26.987 mths

(d) =NORMINV(.05,28,4) 21.421 21.421 mths

CHAPTER 6

SAMPLING AND SAMPLING DISTRIBUTIONS

6.3 subset.

6.4 representative.

6.5 Non-probability sampling: sample members are chosen using non-random criteria meaning

that some members of the target population are excluded from being included in the

sample.

Probability sampling: sample members are chosen using random selection processes so that

each target population member stands a chance of being included in the sample.

6.6 Random (probability) sampling methods. Every member of the target population has a

chance of being included in the sample. This is likely to result in a more representative

sample than if a non-probability sampling method was used.

6.9 equal

6.18 95.5%

6.19 Normal

6.20 n = 30 or larger

6.22 Sampling error is the error made when estimating the true population parameter (e.g.

population mean) when using the sample statistic (e.g. sample mean).

CHAPTER 7

Exercise 7.1 To estimate a population parameter value by defining an Interval within which the true

population value is likely to fall at a stated level of confidence.

Exercise 7.2 Use the z-statistic since the population standard deviation, σ is known (Given σ = 8)

95% Confidence level z(0.95) = 1.96 Use NORMSINV(0.975)

Margin of error z x SE 1.96

Upper 95% confidence limit 85 + 1.96(1) 86.96

Exercise 7.4 Use the t -statistic since the population standard deviation, σ is unknown (Given s = 6)

Degrees of freedom = 25 - 1 24

90% Confidence level t (0.10,24) = 1.711 Use TINV(0.10,24)

Margin of error t x SE 2.91 Note: TINV requires tail probability

Upper 90% confidence limit 54 + 1.711(1.2) 56.05

Exercise 7.5

(a) Manually

x(bar) = 24.4 σ = 10.8 n = 144

z crit (95%) = 1.96

95% Confidence level = 1.96 *0.9 = 1.764

Upper 95% confidence limit 24.4 + 1.764 = 26.164

Interpretation There is a 95% chance that the interval (22.636 to 26.164) covers

the actual average number of employees per SME in Gauteng.

Exercise 7.6

(a) Manually

x(bar) = 131.6 σ = 25 n = 87

z crit (90%) = 1.645

90% Confidence level = 1.645*2.68 4.409

Upper 90% confidence limit 131.6 + 4.409 = 136.009

Interpretation

There is a 90% chance that the interval (127.191 to 136.009) covers the

actual average no. of palettes per order received by a sugar mill in Durban.

Upper 90% limit 136.009*720 = 97926.48 palettes

Interpretation

Total palettes shipped in a year, on average, is likely to be

between 91 577 and 97 926 palettes, with 90% confidence.

Exercise 7.7

(a) Manually

x(bar) = R356 σ = R44 n = 256

z crit (95%) = 1.96

95% Confidence level = 1.96 * 2.75 = 5.39

Upper 95% confidence limit 356 + 5.39 = R361.39

Interpretation

There is a 95% chance that the interval (R350.61 to R361.39) covers the

actual average monthly car insurance premium of medium-sized cars.

90% Confidence level = 1.645*2.75 = 4.524

Upper 90% confidence limit 356 + 4,524 = R360.52

Interpretation

There is a 90% chance that the interval (R351.48 to R360.52) covers the

actual average monthly car insurance premium of medium-sized cars.

the 95% confidence interval limits.

The lower confidence results in a more precise set of interval limits (narrower).

Upper 95% confidence limit 361.39*3000 R1 084 170

Interpretation

Total monthly premium income, on average, is likely to be between

R1 051 830 and R1 084 170 with 95% confidence.

Exercise 7.8

(a) Manually

x(bar) = 4.985 σ = 0.04 n = 50

z crit (99%) = 2.58

99% Confidence level = 2.58 * 0.005657 = 0.01460

Upper 99% confidence limit 4.985 + 0.0146 = 4.9996

Interpretation

There is a 99% chance that the interval (4.97 litres to 4.9996 litres)

covers the actual average volume of paint in all five-litre cans.

(b) Yes, the store owner has statistical evidence to confirm that the average

volume of paint in 5-litre cans is most likely to be below 5 litres.

Exercise 7.9

(a) Manually

x(bar) = 3.8 σ = 0.6 n = 24

z crit (90%) = 1.645

90% Confidence level = 0.20147

Upper 90% confidence limit 3.8 + 0.20147 = 4.001

Interpretation

There is a 90% chance that the interval (3.599 to 4.001) covers

the actual average inventory turnover rate of all convenience stores.

Exercise 7.10

(a) Manually

x(bar) = 166.2 s = 22.8 n = 21

t crit (95%,20) 2.086

95% Confidence level = 10.3786

Upper 95% confidence limit 166.2 + 10.3786 176.58

Interpretation There is a 95% chance that between 155.8 and 176.6 calls, on

average, will be received by the call centre daily.

(b) Manually

x(bar) = 166.2 s = 22.8 n = 21

t crit (99%,20) 2.845

99% Confidence level = 14.1549

Upper 99% confidence limit 166.2 + 14.1550 180.35

Interpretation There is a 99% chance that between 152.1 and 180.4 calls, on

average, will be received by the call centre daily.

(c) The 95% confidence interval is more precise (narrower), but less reliable than the

99% confidence interval which is less precise (wider), but more reliable.

(e) Over a 30 day period Lower 95% confidence limit 155.82*30 = 4674.60

Upper 95% confidence limit 176.58*30 = 5297.40

Interpretation There is a 95% chance that between 4675 and 5298 calls, on

average, will be received over a 30-day period.

Exercise 7.11

(a) Manually

x(bar) = 12.5 s = 3.4 n = 28

t crit (90%, 27) 1.703

90% Confidence level = 1.0943

Upper 90% confidence limit 12.5 + 1.0943 = 13.594

Interpretation There is a 90% chance that the actual mean dividend yield of

all JSE-listed companies lies between 11.41% and 13.59%.

Exercise 7.12

(a) Manually

x(bar) = 0.981 s = 0.052 n = 18

t crit (99%, 17) = 2.8982

99% Confidence level = 0.0355

Upper 99% confidence limit 0.981 + 0.0355 = 1.0165

Interpretation There is a 99% chance that the average fill of 1-litre cartons

of milk lies between 0.9455 litres and 1.0165 litres.

the cartons do contain one litre of milk on average.

(b) Manually

x(bar) = 0.981 s = 0.052 n = 18

t crit (95%, 17) = 2.1098

95% Confidence level = 0.0259

Upper 95% confidence limit 0.981 + 0.0259 = 1.0069

Interpretation There is a 95% chance that the average fill of 1-litre cartons

of milk lies between 0.9551 litres and 1.0069 litres.

The 95% confidence interval is more precise (narrower), but less reliable than the

99% confidence interval which is less precise (wider), but more reliable.

t-crit (0.95, 17) =TINV(0.05,17) 2.1098

Exercise 7.13

(a) Manually

x(bar) = 1420 s = 160 n = 50

t crit (90%, 49) = 1.676 (use df = 50 from Table 2, Appendix)

90% Confidence level = 37.9235

Upper 90% confidence limit 1421 + 37.9235 = 1457.92

union members lies between R1382.08 and R1457.92.

(b) Manually

x(bar) = 1420 s = 160 n = 50

t crit (99%, 49) = 2.678 (use df = 50 from Table 2, Appendix)

99% Confidence level = 60.5962

Upper 99% confidence limit 1420 + 60.5962 = 1480.6

union members lies between R1359.40 and R1480.60.

(c) The 90% confidence interval is more precise (narrower), but less reliable than the

99% confidence interval which is less precise (wider), but more reliable.

t-crit (0.99, 49) =TINV(0.01,49) 2.6800

Exercise 7.14

x = 84 n = 200

p = 84/200 = 0.42

std error = √[(0.42)(0.58)/200] 0.0349

z crit (95%) = 1.96

95% Confidence level = 0.0684

Upper 95% confidence limit 0.42 + 0.0684 = 0.4884

Interpretation

There is a 95% chance that the percentage of manufacturing firms that

meet the employment equity charter lies between 35.2% and 48.8%.

Exercise 7.15

x = 68 n = 160

p = 68/160 = 0.425

std error = √[(0.425)(0.575)/160] = 0.03908

z crit (95%) = 1.96

95% Confidence level = 0.0766

Upper 95% confidence limit 0.425 + 0.0766 = 0.5016

Interpretation

There is a 95% chance that the percentage of cash-paying customers

lies between 34.8% and 50.2%.

Exercise 7.16

p = 287/365 = 0.786

std error = √[(0.786)*(0.214)/365] = 0.021467

z crit (90%) = 1.645

90% Confidence level = 0.03532

Upper 90% confidence limit 0.786 + 0.0353 = 0.8213

Interpretation

There is a 90% chance that the percentage of non-overdrawn cheque accounts

at the Tshwane branch lies between 75.1% and 82.1%.

Exercise 7.17

x = 120 n = 300

p = 120/300 = 0.4

std error = √[(0.4)*(0.6)/300] = 0.02828

z crit (90%) = 1.645

90% Confidence level = 0.04652

Upper 90% confidence limit 0.4 + 0.04652 = 0.4465

Interpretation

There is a 90% chance that the percentage of shoppers who frequent a shopping

mall primarily because of its store mix lies between 35.4% and 44.7%.

Exercise 7.18 File: X7.18 - cashier absenteeism.xlsx

Absent Days

Mean 9.379

Standard Error 0.625

Median 9

Mode 7

Standard Deviation 3.364

Sample Variance 11.315

Kurtosis -0.919

Skewness -0.009

Range 12

Minimum 3

Maximum 15

Sum 272

Count 29

Confidence Level (95.0%) 1.280

68% of employees were absent between 6 and 12.7 days (within 1 std dev).

The data is symmetrical about the mean (skewness = -0.009).

Absenteeism per employee ranged between 3 (min) and 15 days (max) last year.

Upper 95% Confidence Limit 9.379 + 1.28 = 10.66

Interpretation

There is a 95% chance that the average number of days absent last year by

all cashiers in a supermarket was between 8.1 days and 10.7 days.

(c) Since the 95% confidence interval covers 10 days (8.1 < μ < 10.66), it is

possible that the mean number of days absent per employee does exceed 10 days.

Thus the company's policy is not being strictly adhered to.

Given standard error = 0.6250

Confidence level (95%) = 0.625 * 2.0484 = 1.2803

Exercise 7.19 File: X7.19 - parcel masses.xlsx

(a) Descriptive Statistics - Parcel Weights (kg) (Using Data Analysis in Excel )

Mean 2.829

Standard Error 0.091

Median 2.78

Mode 3.22

Standard Deviation 0.5998

Sample Variance 0.360

Kurtosis -0.972

Skewness -0.108

Range 2.17

Minimum 1.75

Maximum 3.92

Sum 121.63

Count 43

Confidence Level (90.0%) 0.154

68% of parcels weigh between 2.23 kg and 3.43 kg (within 1 std dev).

The data is symmetrical about the mean (skewness = -0.1076).

Parcel weights ranged between 1.75 kg (min) and 3.92 kg (max).

Upper 90% Confidence Limit 2.829 + 0.154 = 2.983

Interpretation

There is a 90% chance that the average parcel weight is between 2.68 kg and 2.98 kg.

(c) Since the 90% confidence interval lies below 3 kg (2.675 < μ < 2.983), it is

highly likely that parcel weights do not exceed 3 kg.

Thus Post-Net is adhering to its policy.

Given standard error = 0.0910

Confidence level (90%) = 0.091 * 1.682 = 0.154

Exercise 7.20 File: X7.20 - cost-to-income.xlsx

Cost-to-Income Ratio

Mean 71.240

Standard Error 1.997

Median 68

Mode 84

Standard Deviation 14.123

Sample Variance 199.451

Kurtosis -1.145

Skewness 0.139

Range 53

Minimum 44

Maximum 97

Sum 3562

Count 50

Confidence Level (95.0%) 4.014

68% of companies have a cost-to-income ratio of between 57.1% and 85.4% (within 1 std dev).

The data is reasonably symmetrical about the mean (skewness = 0.139).

Cost-to-income ratios ranged between 44% and 97% for the sample of public companies.

Upper 95% Confidence Limit 71.24 + 4.014 = 75.254

Interpretation

There is a 95% chance that the average cost-to-income ratio of public companies

is between 67.2% and 75.3%.

Given standard error = 1.9970

Confidence level (95%) = 1.997 * 2.0096 = 4.0132

(d) Find P(x > 75) using "μ" = 71.24 and "σ"= 14.123

Standardise x = 75 to t-stat t-stat = (75 - 71.24)/14.123 = 0.26623

=TDIST(0.26623,49,1) 0.39559

Interpretation

39.6% of all public companies are likely to have a cost-to-income ratio

in excess of the 75% rule of thumb.

Sample size determination

e 10

σ 50

Exercise 22

(a) z 2.58 n = 666

e 0.1

σ 1

e 0.15

σ 1

e 0.2

σ 1

p 0.5

e 0.03

CHAPTER 8

HYPOTHESIS TESTS

Exercise 8.1 To test whether a claim / statement made about a population parameter

value is probably true or false, based on sample evidence.

Exercise 8.2 The “closeness” of the sample statistics to the claimed population parameter value.

Step 1: Define the statistical hypotheses (the null and alternative hypotheses).

Step 2: Determine the region of acceptance of the null hypothesis.

Step 3: Compute the sample test statistic.

Step 4: Compare the sample test statistic to the region of acceptance.

Step 5: Draw the statistical and management conclusions.

(and sample size when the population standard deviation is unknown)

Exercise 8.6

x (bar) = 577 σ = 86 n = 120 α = 0.05

Use z test statistic since σ is known.

=NORMSINV(0.95) [using Excel ]

(b) z-stat = (577-560)/(86/√(120) = 2.165

Since z-stat (2.165) > z-crit (1.645), there is sufficient sample evidence at the

5% level of significance to reject H0 in favour of H1. (i.e. Reject H0)

Conclude that the population mean value is significantly larger than 560.

(c) p -value = P(z > 2.165) = 0.5 - 0.4848 = 0.0152 From z-table; or

=1-NORMSDIST(2.165)

x = 216 n = 330 α = 0.10 Derive p = 216/330 = 0,6545

Use z test statistic for proportions

=NORMSINV(0.9) [using Excel ]

Since z-stat (-2.65) < z-crit (-1.28), there is sufficient sample evidence at the

10% level of significance to reject H0 in favour of H1. (i.e. Reject H0)

Conclude that the true population proportion is significantly less than 0.72.

(c) p -value = P(z < -2.65) = 0.5 - 0.496 = 0.0040 From z-table; or

=NORMSDIST(-2.65)

x (bar) = 9.6 s = 2.9 n = 30 α = 0.01

Use t test statistic since σ is unknown

(only given the sample standard deviation, s and n is small (n < 30))

or '=TINV(0.01,29) [using Excel ]

Since t-stat (2.644) falls within the area of acceptance, there is insufficient sample

evidence at the 1% level of significance to reject H0 in favour of H1. (i.e. Accept H0)

Conclude that the true mean value is equal to 8.2.

(iv) H0: µ ≥ 18 H1: µ < 18

x (bar) = 14.6 s = 3.4 n = 12 α = 0.01

Use t test statistic since σ is unknown

(only have the sample stardard deviation, s and n is small (n < 30))

or '=TINV(0.02,11) [using Excel ]

Since t-stat (-3.464) < t-crit (-2.718), there is sufficient sample evidence at the

1% level of significance to reject H0 in favour of H1. (Reject H0)

Conclude that the true mean value is significantly below 18.

x = 68 n = 250 α = 0.05 Derive p = 68/250 = 0.272

Use z test statistic for proportions

=NORMSINV(0.975) [Excel ]

Since z-stat (-1.627) falls within the area of acceptance, there is insufficient sample

evidence at the 5% level of significance to reject H0 in favour of H1. (i.e. Accept H0)

Conclude that the true population proportion is equal to 0.32.

Exercise 8.7

Area of Acceptance -1.96 ≤ z ≤ 1.96 Read off 0.475 from z-table; or

=NORMSINV(0.975) [using Excel]

Statistical conclusion

Since z-stat (-2.068) lies below -z-crit (-1.96), there is sufficient sample evidence at

the 5% level of significance to reject H0 in favour of H1. (i.e. Reject H0).

Management conclusion

Conclude that the population mean value is significantly different from 85 minutes.

Visitors to the Knysna shopping mall do not spend 85 minutes on average in the mall.

(e) Since p -value (0.0386) < α (0.05), there is strong sample evidence to

conclude that visitors to the Knysna shopping mall do not spend 85 minutes,

on average, in the mall.

shopping mall spend significantly less than 85 minutes in the mall.

Exercise 8.8

(a) H0: µ ≥ 30 min H1: µ < 30 min One-sided lower tailed test

(b) Use the z test statistic since σ is known (σ = 10.5 min (given))

Area of Acceptance z ≥ -2.33 Read off 0.49 from z-table; or

=NORMSINV(0.01) [using Excel]

Statistical conclusion

Since z-stat (-1.855) lies above z-crit (-2.33), there is insufficient sample evidence at

the 1% level of significance to reject H0 in favour of H1 (Do not reject H0).

Management conclusion

Conclude that the population mean value is at least 30 minutes.

The supermarket manager's belief is therefore valid.

Since p -value (0.0318) > α (0.01) the sample evidence does not refute H0.

The sample evidence is not strong enough to refute the belief that customers

spend 30 minutes or more, on average, doing their purchases at the supermarket.

Hence conclude that customers are likely to spend at least half-an-hour,

on average, in the supermarket doing their grocery shopping.

Exercise 8.9

(a) H0: µ ≤ 72 hours H1: µ > 72 hours One-sided upper tailed test

Use α = 0.10

Area of Acceptance z ≤ 1.28 Read off 0.40 from z-table; or

=NORMSINV(0.90) [using Excel]

Statistical conclusion

Since z-stat (1.47) > z-crit (1.28), there is sufficient sample evidence at

the 10% level of significance to reject H0 in favour of H1 (Reject H0).

Management conclusion

Conclude that the local importer's claim is valid.

Consignments are taking significantly longer than 72 hours to clear customs.

Since p -value (0.0708) < α (0.10) this confirms support for H1.

consignment clearance times, on average, are significantly longer than 72 hours.

Exercise 8.10

(b) H0: µ ≤ 40% H1: µ > 40% One-sided upper tailed test

Area of Acceptance z ≤ 2.33 Read off 0.49 from z-table; or

=NORMSINV(0.99) [using Excel]

Statistical conclusion

Since z-stat (2.431) > z-crit (2.33), there is sufficient sample evidence at

the 1% level of significance to reject H0 in favour of H1 (Reject H0).

Management conclusion

Conclude that the Department of Health's concern is justified.

The average markup is significantly greater than 40%.

Since p -value (0.0075) << α (0.01) it confirms H1. There is overwhelming sample

evidence to conclude that the average % markup is significantly greater than 40%.

Exercise 8.11

(a) Use the t test statistic since σ is unknown (only s = 21 gms is given)

Area of Acceptance -2.00 ≤ t ≤ 2.00 Read off t(0.025,63) from t-table; or

'=TINV(0.05,63) [using Excel]

Statistical conclusion

Since t-stat (-1,905) lies within the acceptance area, there is insufficient sample

evidence at the 5% level of significance to reject H0 in favour of H1 (i.e. Accept H0)

Management conclusion

Conclude that Ryeband Bakery is both legally compliant and not wasting ingredients.

The average weight of all white loaves produced by Ryeband Bakery is 700 gms.

Since p -value (0.0613) > α (0.05) H0 cannot be rejected as the sample evidence

is weak in favour of H1.

Hence it can be concluded that, on average, the weight of white bread

loaves produced by the Ryeband Bakery is 700 gm.

(c) H0: µ ≥ 700 gms H1: µ < 700 gms One-sided lower tailed test

Area of Acceptance t ≥ -1.671 Read off t(0.05,63) from t-table; or

=TINV(0.10,63) [using Excel]

Statistical conclusion

Since t-stat (-1.905) < t-crit (-1.671), there is sufficient sample evidence at

the 5% level of significance to reject H0 in favour of H1 (i.e. Reject H0).

Management conclusion

Conclude that Ryeband Bakery is not legally compliant.The average weight of all

white bread loaves produced by Ryeband Bakery is significantly below 700 gms.

Exercise 8.12

(a) Use the t test statistic since σ is unknown (only s = R788 is given)

Area of Acceptance t ≥ -1.33 Read off t(0.10,17) from t-table; or

=TINV(0.20,17) [using Excel]

Statistical conclusion

Since t-stat (-1.211) lies within the acceptance area, there is insufficient sample

evidence at the 10% level of significance to reject H0 in favour of H1 (i.e. Accept H0)

Management conclusion

Conclude that mean weekly sales of the new pudding flavour is not less than R5500.

The company should therefore not withdraw the product at this stage.

Since p -value (0.1212) > α (0.10), H0 cannot be rejected as the sample evidence

is weak in favour of H1.

Hence it can be concluded that, on average, the weekly sales of the new pudding

flavour is not less than R5500 and the product should not be withdrawn.

Exercise 8.13

(a) Use the t test statistic since σ is unknown (only s = 3.6 is given)

Area of Acceptance t ≤ 1.708 Read off t(0.05,25) from t-table; or

=TINV(0.10,25) [using Excel]

Statistical conclusion

Since t-stat (1.841) > t-crit (1.708), it lies in the region of rejection.

Therefore there is sufficient sample evidence at

the 5% level of significance to reject H0 in favour of H1 (Reject H0)

Management conclusion

Conclude that mean tensile strength of the consignment of wire is more than 80 kg.

Marathon Products should accept this consignment as it meets quality specs.

Since p -value (0.0388) < α (0.05) there is strong sample evidence to reject H0

in favour of H1.

Hence it can be concluded that, on average, the tensile strength of

the wire in the consignment exceeds 80 kg. Hence accept the consignment.

Exercise 8.14

(a) Use the t test statistic since σ is unknown (only s = 0.068 is given)

Area of Acceptance t ≥ -1.729 Read off t(0.05,19) from t-table; or

=TINV(0.10,19) [using Excel]

t-stat = (0.982 - 1)/(0.0152) = -1.1842

Statistical conclusion

Since t-stat (-1.1842) > t-crit (-1.729), it lies within the region of acceptance.

There is therefore insufficient sample evidence at the

5% level of significance to reject H0 in favour of H1 (i.e. Accept H0)

Management conclusion

Conclude that the mean fill of one-litre milk containers is not less than 1 litre.

The Consumer Council's claim that containers are being underfilled is not valid.

evidence to support H1.

Hence it can be concluded that, on average, the mean fill of milk containers

is at least one 1 lite. Thus there is no statistical support for the claim.

Exercise 8.15

H0: π ≥ 0.30 H1: π < 0.30 One sided lower tailed test

Use α = 0.05

Area of Acceptance z ≥ -1.645 Read off z(0.45) from z-table; or

=NORMSINV(0.05) [using Excel ]

n= 400 x= 106

p= 106/400 = 0,265

std error = √(0.3*0.7)/400 = 0,0229

z-stat = (0.265 - 0.3)/(0.0229) = -1.5284

Statistical conclusion

Since z-stat (-1.5284) > z-crit (-1.645) there is insufficient sample evidence at

the 5% level of significance to reject H0 in favour of H1 (i.e. Accept H0)

Management conclusion

Conclude that 30% or more of listeners tune into the news broadcast of this station.

The company should advertise in this radio station's news timeslots.

reject H0 in favour of H1. (i.e. Accept H0)

The company can accept the radio station's claim as valid.

Exercise 8.16

Use α = 0.05

Area of Acceptance z ≤ 1.645 Read off z(0.45) from z-table; or

=NORMSINV(0.95) [using Excel ]

n= 150 x= 54

p= (150-54)/150 = 0.64

std error = √(0.6*0.4)/150 = 0.04

z-stat = (0.64 - 0.6)/(0.04) = 1.000

Statistical conclusion

Since z-stat (1.00) < z-crit (1.645) there is insufficient sample evidence at

the 5% level of significance to reject H0 in favour of H1 (i.e. Accept H0)

Management conclusion

Conclude that not more than 60% of Cape Town motorists do not have vehicle insurance.

The motor vehicle advisor's claim is not valid.

Since p -value (0.1587) > α (0.05), the sample evidence does not support H1.

Hence the insurance advisor's claim has no statistical validity.

Exercise 8.17

Use α = 0.10

Area of Acceptance z ≤ 1.28 Read off z(0.40) from z-table; or

=NORMSINV(0.90) [using Excel]

n= 560 x= 96

p= 96/560 = 0.1714

std error = √(0,15*0,85)/560 = 0.0151

z-stat = (0,1714 - 0,15)/(0,0151) = 1.417

Statistical conclusion

Since z-stat (1.417) > z-crit (1.28) there is sufficient sample evidence at

the 10% level of significance to reject H0 in favour of H1 (i.e. Reject H0)

Management conclusion

Conclude that the churn rate in the telecommunications industry exceeds 15%.

Since p -value (0.0782) > α (0.10), there is moderate sample evidence to

reject H0 in favour of H1.

The same management conclusion applies as in (a) above.

Exercise 8.18

Use α = 0.01

Area of Acceptance z ≥ -2.33 Read off z(0.49) from z-table; or

=NORMSINV(0.01) [using Excel]

n= 300 x= 260

p= 260/300 = 0.8667

std error = √(0.9*0.1)/300 = 0.0173

z-stat = (0.8667 - 0.9)/(0.0173) = -1.9249

Statistical conclusion

Since z-stat (-1.9249) > z-crit (-2.33) there is insufficient sample evidence at

the 1% level of significance to reject H0 in favour of H1 (i.e. Do not reject H0)

Management conclusion

Conclude that the germination rate of the barley seed is at least 90%.

The cooperative can justify buying the barley seed from this merchant.

The cooperative can accept the seed merchant's claim.

Since p -value (0.0271) > α (0.01), there is weak sample evidence at

the 1% level of signficance to reject H0 in favour of H1. Hence do not reject H0.

The same management conclusion applies as in (a) above.

Exercise 8.19 File: X8.19 - cost-to-income.xlsx

Cost-to-income (%)

H0: µ ≥ 75% H1: µ < 75% Mean 71.24

Standard Error 1.9973

Use α = 0.05 and degrees of freedom = (n -1) = 49 Median 68

Read off t(0.05,49) from t-table; or Mode 84

=TINV(0.10,49) [using Excel] Standard Deviation 14.1227

Sample Variance 199.451

Area of Acceptance t ≥ -1.676 Kurtosis -1.1450

Skewness 0.1391

std error = 1.9973 Range 53

Minimum 44

t-stat = (71.24 - 75)/(1.9973) = -1.8825 Maximum 97

Sum 3562

Count 50

Statistical conclusion

Since t-stat (-1.8825) < t-crit (-1.676) there is sufficient sample evidence at

the 5% level of significance to reject H0 in favour of H1 (i.e. Reject H0)

Management conclusion

Conclude that the mean cost-to-income ratio of JSE companies is less than 75%.

JSE companies are therefore adhering to the rule of thumb.

Exercise 8.20 File: X8.20 - kitchenware.xlsx

≤ 75 5

76 - ≤ 100 6

101 - ≤ 125 8

126 - ≤ 150 11

151 - ≤ 175 9

176 - ≤ 200 6

201 - ≤ 225 4

> 225 1

(b) Use the t test statistic since σ is unknown Sales Transaction value

Mean 137.12

H0: µ ≥ R150 H1: µ < R150 Standard Error 6.657

Median 138.5

Use α = 0.05 and degrees of freedom = (n -1) = 49 Mode 184

Read off t(0.05,49) from t-table; or Standard Deviation 47.074

=TINV(0.10,49) [using Excel] Sample Variance 2215.944

Kurtosis -0.741

Area of Acceptance t ≥ -1.676 Skewness 0.022

Range 190

std error = 6.657 Minimum 52

Maximum 242

t-stat = (137.12 - 150)/(6.657) = -1.9348 Sum 6856

Count 50

Statistical conclusion

Since t-stat (-1.9348) < t-crit (-1.676) there is sufficient sample evidence at

the 5% level of significance to reject H0 in favour of H1 (Reject H0).

Management conclusion

Conclude that the mean transaction value at the Claremont branch is signficantly

below R150.

The management are advised to close the Claremont branch since it is unprofitable.

Exercise 8.21 File: X8.21 - flight delays.xlsx

delays Count

≤5 0

5 - ≤ 7.5 8

7.6 - ≤ 10 28

10.1 - ≤ 12.5 29

12.6 - ≤ 15 13

15.6 - ≤ 17.5 2

> 17.5 0

35

28 29

30

25

No. of Flights

20

15 13

10 8

5 2

0 0

0

5 7.5 10 12.5 15 17.5 More

Mean 10.324

H0: µ ≤ 10 min H1: µ > 10 min Standard Error 0.261

Median 10.3

Use α = 0.10 and degrees of freedom = (n -1) = 79 Mode 11

Read off t(0.10,79) from t-table; or Standard Deviation 2.333

=TINV(0.20,79) [using Excel] Sample Variance 5.444

Kurtosis -0.550

Area of Acceptance t ≥ 1.292 Skewness 0.084

Range 10.3

std error = 0.261 Minimum 5.3

Maximum 15.6

t-stat = (10.324 - 10)/(0.261) = 1.241 Sum 825.9

Count 80

Statistical conclusion

Since t-stat (1.241) < t-crit (1.292) there is insufficient sample evidence at

the 10% level of significance to reject H0 in favour of H1 (i.e. Accept H0).

Management conclusion

Conclude that flight delay times, on average, do not exceed 10 minutes.

ACASA management do not need to conduct an indepth investigation on flight delays.

Exercise 8.22 File: X8.22 - medical claims.xls

≤ 100 5

101 - ≤ 125 3

126 - ≤ 150 9

151 - ≤ 175 21

176 - ≤ 200 25

201 - ≤ 225 18

226 - ≤ 250 8

251 - ≤ 300 11

30

25

25

21

20 18

No. of Claims

15

11

9

10 8

5

5 3

0

0

100 125 150 175 200 225 250 300 More

Mean 190.39

H0: µ ≤ 180 claims H1: µ > 180 claims Standard Error 4.568

Median 190

Use α = 0.01 and degrees of freedom = (n -1) = 99 Mode 210

Read off t(0.01,99) from t-table; or Standard Deviation 45.680

=TINV(0.02,99) [using Excel] Sample Variance 2086.685

Kurtosis -0.092

Area of Acceptance t ≥ 2.365 Skewness 0.096

Range 199

std error = 4.568 Minimum 92

Maximum 291

t-stat = (190.39 - 180)/(4.568) = 2.275 Sum 19039

Count 100

Statistical conclusion

Since t-stat (2.275) < t-crit (2.365) there is insufficient sample evidence at

the 1% level of significance to reject H0 in favour of H1 (Thus accept H0).

Management conclusion

Conclude that the average number of claims processed daily does not exceed 180.

Thus the supervisor has no statistical grounds for requesting additional staff.

Exercise 8.23 File: X8.23 - newspaper readership.xlsx

Sun % 15.8%

Count 19

Guardian % 35.0% Tabloid %

Count 42 Sun 15.8

Mail % 25.8% Guardian 35

Count 31 Mail 25.8

Voice % 23.3% Voice 23.3

Count 28

Total % 100%

Total Count 120

40

35

35

30

25.8

23.3

% of Readers

25

20

15.8

15

10

5

0

Sun Guardian Mail Voice

Interpretation

Based on the sample fo 120 tabloid readers, the Guardian newspaper has the

largest share of 35%. The Sun has the lowest percentage of readers at 16%.

H0: π ≥ 0.40 H1: π < 0.40 Test claim that Guardian newspaper

has at least a 40% market share.

(c) Use α = 0.05

Area of Acceptance z ≥ -1.645 Read off z(0.45) from z-table; or

=NORMSINV(0.05) [using Excel]

n= 120 x= 42

p= 42/120 = 0.35

std error = √(0.4*0.6)/120 = 0.0447

z-stat = (0.35 - 0.4)/(0.0447) = -1.1186

Statistical conclusion

Since z-stat (-1.1186) > z-crit (-1.645) there is insufficient sample evidence at

the 5% level of significance to reject H0 in favour of H1 (Thus accept H0)

Management conclusion

Hence conclude that the Guardian's market share is at least 40%

The Guardian newspaper 's claim is justified statistically.

Exercise 8.24 File: X8.24 - citrus products.xlsx

High Count 34

% 20% Awareness %

Low Count 72 Low 42.35

% 42.35% Moderate 37.65

Moderate Count 64 High 20

% 37.65%

Total Count 170

Total % 100%

50

42.35

37.65

% of respondents

40

30

20

20

10

0

Low Moderate High

Awareness Levels

Interpretation

80% of sampled consumers have a low or moderate awareness. Only 20% of

the sampled consumers indicated a high awarness of the nutritional benefits of citrus.

Area of Acceptance z ≤ 2.326 Read off z(0.49) from z-table; or

=NORMSINV(0.99) [using Excel]

n= 170 x= 34

p= 34/170 = 0.2

std error = √(0.15*0.85)/170 = 0.0274

z-stat = (0.20 - 0.15)/(0.0274) = 1.825

Statistical conclusion

Since z-stat (1.825) < z-crit (2.326) there is insufficient sample evidence at

the 1% level of significance to reject H0 in favour of H1 (i.e. Accept H0)

Since the level of high consumer awareness does not exceed 15%, it is recommended that

Fruitco should launch a national awareness campaign.

Exercise 8.25 File: X8.25 - aluminium scrap.xlsx

≤ 2.8 6

2.8 - ≤ 3.2 12

3.2 - ≤ 3.6 13

3.6 - ≤ 4.0 11

4.0 - ≤ 4.4 5

4.4 - ≤ 4.8 3

Total 50

16

14 13

12

12 11

10

No. of days

8

6

6 5

4 3

Scrap % intervals

Interpretation

The assumption of normality is largely satisfied.

The histogram is only moderately skewed to the right.

Mean 3.483

Standard Error 0.0723

Median 3.41

Mode 2.97

Standard Deviation 0.511

Sample Variance 0.261

Kurtosis -0.851

Skewness 0.331

Range 1.83

Minimum 2.67

Maximum 4.5

Sum 174.14

Count 50

Confidence Level(95.0%) 0.145

Upper 95% confidence limit = 3.483+0.145 = 3.63

There is a 95% chance that the average daily % scrap produced by the machine

is likely to lie between 3.34% and 3.63%.

Use the t test statistic since σ is unknown (only s is given)

H1: μ < 3.75%

t-crit = t(0.05)(49) = =TINV(0.1,49) -1.677

-3.693

Statistical conclusion

Since t-stat (-3.693) lies well within the region of rejection, there is strong sample

evidence at the 5% significance level to reject H0 in favour of H1. (Thus reject H0).

Management conclusion

Conclude that the average daily % scrap produced by the machine is less than 3.75%.

The machine is not yet due for a full maintenance service.

Exercise 8.26 Hypothesis Test for a Single Population Variance, σ2

5m pipe length variability analysis

Problem Characteristics

Variable x = length of pipe (specification = 5m)

Data type numerical, ratio scaled, continuous.

Data n = 26 pipes

α= 0.05

s= 3.46 σ0 = 3

H1 : σ2 > 9 Management Question in H0

2

Upper Χ -crit = 37.652 Do not reject Ho if Χ2-stat ≤ 37.652

2

Step 3 Sample test statistic X -stat Formula 8.4

2

X -stat = 33.254 Also p -value = 0.12

Use CHISQ.DIST.RT(x, df)

Steps 4, 5 Statistical Conclusion

Since X2-stat < upper X2-crit (i.e. lies in region of non-rejection of H0),

do not reject H0 at 5% level of significance.

Management Conclusion

The production manager can be 95% confident that the variation

in pipe lengths is within the limits of the product specification.

Exercise 8.27 Hypothesis Test for a Single Population Variance, σ2

Problem Characteristics

Variable x = unknown - but numeric

Data type numerical, ratio scaled, continuous

Data n = 20

α= 0.1

s2 = 49.3 σ20 = 30

2

H1 : σ > 30

2

Upper Χ -crit = 27.204 2

Do not reject Ho if Χ -stat ≤ 27.204

2

Step 3 Sample test statistic X -stat Formula 8.4

X2-stat = 31.223 Also p -value = 0.0382

Use CHISQ.DIST.RT(x, df)

Steps 4, 5 Statistical Conclusion

Since X2-stat > X2-crit (i.e. lies in region of rejection of H0),

reject H0 at 10% level of significance.

Management Conclusion

We are 90% confident that the population variance is significantly greater

than the specified value of 30.

Exercise 8.28 Hypothesis Test for a Single Population Variance, σ

2

Problem Characteristics

Variable x = claim values (in Rand)

Data type numerical, ratio scaled - implies means and standard deviations

Data n = 32 claims

α= 0.05

2

s= 84 σ 0= 5625

2

Step 1 H0 : σ = 5625 Two-tailed test

2

H1 : σ ≠ 5625 Management Question in H0

2 2

Upper Χ -crit = 48.232 Lower Χ -crit = 17.539

2

Do not reject Ho if 17.539 ≤ Χ -stat ≤ 48.232

2

Step 3 Sample test statisticX -stat Use Formula 8.4

2

X -stat = 38.886 Also p -value = 0.1561 Use CHISQ.DIST.RT(x, df)

Since X2-stat < X2-crit (i.e. lies in region of non-rejection of H0),

do not reject H0 at 5% level of significance.

Management Conclusion

2

The insurance industry can be 95% confident that the variation in claim values is still σ = R5625.

---ooOoo---

Exercise 8.29 File: X8.29 - pain relief.xlsx

Problem Characteristics

Variable x = time to pain relief (in minutes)

Data type numerical, ratio scaled, continuous

Data n = 16 patients

α= 0.01

s= 0.9004 σ20 = 1.8

2

H1 : σ < 1.8 Management Question in H0

2

Lower Χ -crit = 5.229 2

Do not reject Ho if Χ -stat ≥ 5.299

2

Step 3 Sample test statistic X -stat Use Formula 8.4

2

X -stat = 6.756 Also p -value = 0.0359

Use CHISQ.DIST(x, df)

Steps 4, 5 Statistical Conclusion

Since X2-stat > lower X2-crit (i.e. lies in region of non-rejection of H0),

do not reject H0 at 1% level of significance.

the variation in time to pain relief from the new headache pill is not significantly less than

2

the current headache pill (of σ = 1.8 min).

(i.e. it does not significantly reduce variation in time to pain relief).

CHAPTER 9

HYPOTHESIS TESTS

Exercise 9.1 When the population standard deviations of the two populations are unknown.

z-stat = [(72 - 66) - 0]/√(202/40+102/50)

z-stat = 1.7321

(b) t-crit (0.05,90) = 1.987 Refer to Appendix 1 Table 3

H1: π1 < π2 One sided lower tailed test.

Exercise 9.6

H1 : σ²1 ≠ σ²2

Decision rule: Do not reject H0 if F-stat ≤ 2.50 2.515

Since F-stat (1.555) < F-crit (2.515), do not reject H0 in favour of H1.

Therefore, use the pooled-variances t -test to test for differences in means.

Let population 2 = Retailers

Let μi = population mean earnings yield (%) per sector i .

Use the t test statistic since σi's are unknown (only s1 and s2 are given)

H1 : μ1 ≠ μ2

t-crit = t(0.05)(41) = 2.021

14.454

1.16747

t-stat = ((8.45-10.22)-(0))/1.16747

-1.51609

Conclusion

Since t-stat (-1.51609) lies within the region of acceptance, there is

insufficient sample evidence at the 5% level of significance to

reject H0 in favour of H1. (i.e. Accept H0).

Conclude that there is no difference in the mean earnings yield (%) between

manufacturing companies and retail companies.

Exercise 9.7

Let population 2 = Non-DIY Consumers

Let μi = population mean age of each consumer group i .

Use the t test statistic since σi's are unknown (only s1 and s2 are given)

H1: μ1 < μ2

t-crit = t(0.10)(61) = -1.296

258.0197

4.060301

t-stat = ((41.8-47.4)-(0))/4.0603

-1.379

Since t-stat (-1.379) lies outside (below) the region of acceptance,

there is sufficient sample evidence at the 10% level of significance

to reject H0 in favour of H1. (i.e. Reject H0).

Management conclusion

Conclude that the mean age of DIY consumers is significantly lower than

the mean age of non-DIY consumers at the 10% level of significance.

t-crit = t(0.05)(61) = -1.671

Decision rule Do not reject H0 in favour of H1 if -1.671 ≤ t-stat

Statistical conclusion

t-stat (-1.379) now lies within the (new) region of acceptance.

Thus there is insufficient sample evidence at the 5% level of significance

to reject H0 in favour of H1. (i.e. Accept H0).

Management conclusion

Conclude that there is no significant difference in the mean ages between

DIY consumers and non-DIY consumers.

Exercise 9.8

Let population 2 = Train commuters

Let μi = population mean commuting time for each transport mode i .

Use the t test statistic since σi's are unknown (only s1 and s2 are given)

H1: μ1 > μ2

2 2

H0: σ 1 =σ 2 Two tailed test

H1: σ²1 ≠ σ²2

Decision rule: Do not reject H0 if F-stat ≤ 2.10 2.105

Therefore, use the unequal-variances t -test to test for differences in means.

(Use df formula 9.9) df (denom) =0.374049

Use α = 0.01 with df = 30

t-crit = t(0.01)(30) = 2.75

1.680545

t-stat = ((35.3-31.8)-(0))/1.6805

2.083

Statistical conclusion

Since t-stat (2,083) lies within the region of acceptance,

there is insufficient sample evidence at the 1% level of significance

to reject H0 in favour of H1. (i.e. Accept H0).

Conclude that there is no difference in the mean commuting times between

bus and train commuters.

Recommendation

Since there is no difference in mean commuting times, either mode of transport can be

prioritised for upgrading (or both can be upgraded simultaneously).

Exercise 9.9

Let population 2 = Visa Card users

Let μi = population mean month-end credit card balance (in Rands) for each card type i .

H1: μ1 ≠ μ2

z-crit = z(0.05) = 1.96

60.26065

z-stat = ((922-828)-(0))/60.26065

1.5599

Statistical conclusion

Since z-stat (1.5599) lies within the region of acceptance, there is insufficient sample

evidence at the 5% significance level to reject H0 in favour of H1. (i.e. Accept H0).

Management conclusion

Conclude that there is no difference in the mean month-end credit card balances

between Mastercard holders and Visa Card holders.

Exercise 9.10

Let population 2 = attendees of job enrichment workshops

Let μi = population mean job satisfaction rating for each employee category i .

H1: μ1 < μ2 ← non-attendees have lower job satisfaction than

attendees.

z-crit = z(0.05) = -1.645

0.283901

z-stat = ((6.9-7.5)-(0))/0.283901

-2.1134

Statistical conclusion

Since z-stat (-2.1134) lies outside (below) the region of acceptance, there is sufficient

sample evidence at the 5% significance level to reject H0 in favour of H1. (i.e. Reject H0).

Management conclusion

Conclude that the mean job satisfaction score for non-attendees is significantly

lower than the mean job satisfaction score for job enrichment attendees.

Thus, the statistical evidence supports the view that the job enrichment

workshops significantly increased job satisfaction levels of sales consultants.

Exercise 9.11

Std error = √(2.3²/15) = 0.5939

z (0.95) = 1.96

Upper 95% confidence limit = 12.4 + 1.96 (0.5939) = 13.56 days

There is a 95% chance that the true average time to settlement for

claims lodged against the Explorer Fund lies between 11.24 days and 13.56 days.

Let population 2 = Explorer Medical Fund

Let μi = population mean time to settlement of claims by each medical fund i .

H1: μ1 < μ2 ← Green-Aid Fund settles quicker than Explorer Fund

z-crit = z(0.05) = -1.645

1.041199

z-stat = ((10.8-12.4)-(0))/1.041199

-1.5367

Statistical conclusion

Since z-stat (-1.5367) lies within the region of acceptance, there is insufficient sample

(or p-value = 0.0622 > α = 0.05) (see (iii) below), there is insufficient sample evidence

evidence at the 5% significance level to reject H0 in favour of H1. (i.e. Accept H0).

Management conclusion

There is no difference in the mean claims settlement time between the two Funds.

Thus, the statistical evidence does not support the view that the Green-Aid Medical

Fund settles claims sooner, on average, than the Explorer Medical Fund.

Exercise 9.12

Let population 2 = Electric ovens

Let μi = population mean baking time for each oven type i .

Use the t test statistic since σi's are unknown (only s1 and s2 are given)

H1: μ1 < μ2 ← gas ovens bake faster than electric ovens

t-crit = t(0.05)(8) = -1.86

0.01685

0.08210

t-stat = ((0.75-0.89)-(0))/0.0821

-1.705

Statistical conclusion

Since t-stat (-1.705) lies within the region of acceptance, there is insufficient sample

evidence at the 5% significance level to reject H0 in favour of H1. (i.e. Accept H0).

Management conclusion

Conclude that there is no difference in the mean bread baking time between gas

and electric ovens, at the 5% significance level.

Gas ovens are therefore, not faster, on average, than electric ovens.

Exercise 9.13

H0: σ21 = σ22 H1: σ²1 ≠ σ²2

F-stat = = 152.2²/121.5² = 1.569

Since F-stat (1.569) < 3, do not reject H0 in favour of H1.

Conclude that the two population variances are equal.

Therefore, use the pooled-variances t -test to test for differences in means.

Let population 2 = Durban branch

Let μi = population mean size of orders received by each branch i .

Use the t test statistic since σi's are unknown (only s1 and s2 are given)

H1: μ1 > μ2 ← CT branch performing better than Durban branch

t-crit = t(0.10)(31) = 1.309

18556.97

47.6243

t-stat = ((335.2-265.6)-(0))/47.6243

1.4614

Statistical conclusion

Since t-stat (1.4614) lies outside (above) the region of acceptance, there is sufficient

sample evidence at the 10% significance level to reject H0 in favour of H1 (i.e. Reject H0).

Management conclusion

Conclude that the mean size of orders received by the Cape Town branch is

significantly larger than the mean size of orders received by the Durban branch.

Thus the Cape Town branch is performing better than the Durban branch in

terms of average order size.

t-crit = t(0.05)(31) = 1.696

Decision rule (new) Do not reject H0 in favour of H1 if t-stat ≤ 1.696

Statistical conclusion

t-stat (= 1.4614) now lies within the (new) region of acceptance.

Thus there is insufficient sample evidence at the 5% level of significance to

reject H0 in favour of H1. (i.e. Accept H0).

Management conclusion

Conclude that there is no significant difference in the mean order sizes between the

Cape Town branch and the Durban branch.

Thus there is no evidence to believe that the Cape Town branch is performing better

than the Durban branch in terms of average order size.

(d) Findings based on the 5% significance level are more meaningful than those based

on the 10% significance level because it requires stronger (more convincing) sample

evidence before tests conducted at 5% are prepared to reject the null hypothesis.

The operations manager can be more confident that there is no difference in mean

performance between the two branches (conclusion based on (b)).

Exercise 9.14 File: X9.14 - package designs.xlsx

H1: σ²1 ≠ σ²2

Decision rule: Do not reject H0 if F-stat ≤ 4.99 4.995

Since F-stat (1.377) < F-crit (4.99), do not reject H0 in favour of H1.

Therefore, use the pooled-variances t -test to test for differences in means.

Now conduct the t-test for equal means, using the pooled-variances t-test approach.

Let population 2 = Barrel-shaped carton

Let μi = population mean sales volume of one-litre cartons for each carton shape i .

Use the t test statistic since σi's are unknown (only s1 and s2 are given)

H1: μ1 < μ2 'Pyramid' sales are less than 'Barrel' sales

t-crit = t(0.05)(14) = -1.761

28.09874

2.650412

t-stat = ((23.75-27.375)-(0))/2.650412

-1.3677

Statistical conclusion

Since t-stat (-1.3677) lies within the region of acceptance, there is insufficient sample

evidence at the 5% significance level to reject H0 in favour of H1. (i.e. Accept H0).

Management conclusion

Conclude that there is no difference in the mean weekly sales of one-litre cartons

of apple juice between the pyramid-shaped and barrel-shaped carton designs.

Thus the marketer can choose either package design to achieve higher weekly sales.

Exercise 9.15

Let population 2 = Fruity Wheat consumers

Let πi = population proportion of consumers who prefer fruit-flavoured wheat cereal i .

H1: π1 > π2 Fruit Puffs is preferred by more consumers than

Fruity Wheat.

z-crit = z(0.05) = 1.645

n 175 150

x 54 36

pi 0.309 0.240

0.2769

0.04978949

z-stat = (0.309-0.24)/0.049789

1.3858

Statistical conclusion

Since z-stat (= 1.3858) lies within the region of acceptance, there is insufficient sample

evidence at the 5% significance level to reject H0 in favour of H1. (i.e. Accept H0).

Management conclusion

Conclude that there is no difference in the percentage of consumers who prefer

each type of fruit-flavoured wheat cereal.

The marketer's view that Fruit Puffs is more preferred than Fruity Wheat cannot be

validated based on the statistical evidence at the 5% level of significance.

The marketer can therefore choose to launch either fruit flavour of wheat cereal.

Exercise 9.16

Let population 2 = Female respondents

Let πi = population proportion who prefer jazz for each gender i .

H1: π1 ≠ π2

z-crit = 1.96

n 140 110

x 46 21

pi 0.329 0.191

0.268

0.056432928

z-stat = (0.329-0.191)/0.056433

2.445

Statistical conclusion

Since z-stat (2.445) lies outside (above) the region of acceptance, there is sufficient

sample evidence at the 5% significance level to reject H0 in favour of H1 (i.e. Reject H0)

Management conclusion

Conclude that there is a difference in the proportion of males compared to the

proportion of females who enjoy listening to jazz.

Since the p-value (=0.0145) < α = 0.05, there is strong sample evidence in

support of H1. Hence conclude that there is a difference in the proportion of males

compared to the proportion of females who enjoy listening to jazz.

Exercise 9.17

Let population 2 = Elite Cheque Account clients

Let πi = population proportion of clients for each account type i who are overdrawn.

H1: π1 < π2 'Status' proportion less than 'Elite' proportion

z-crit = -1.645

n 300 250

x 48 55

pi 0.16 0.22

0.1873

0.0334

z-stat = (0.16-0.22)/0.0334

-1.796

Statistical conclusion

Since z-stat (= -1.796) lies outside (below) the region of acceptance, there is sufficient

sample evidence at the 5% significance level to reject H0 in favour of H1. (i.e. Reject H0).

Management conclusion

Conclude that proportionately more Elite cheque account clients are overdrawn

compared to Status cheque account clients.

Since the p-value (=0.0363) < α = 0.05, there is strong sample evidence in

support of H1. Hence conclude that proportionately more Elite cheque account clients

are overdrawn compared to Status cheque account clients.

Exercise 9.18 File: X9.18 - aluminium scrap.xlsx

2 2

H0: σ 1 =σ2 Two tailed test

H1: σ²1 ≠ σ²2

Since F-stat (1.702) < F-crit (1.77), do not reject H0 in favour of H1.

Use the pooled-variances t -test to test for differences in means.

Machine 1 Machine 2

Mean 3.483 3.668

Variance 0.261 0.153

Observations 50 30

df 49 29

F-stat 1.702

P(F<=f) one-tail 0.0639

F Critical one-tail 1.777

Let population 2 = machine 2 daily % scrap

Let μi = population average daily % scrap produced by each machine i .

(a) (iii) Use the pooled-variances t test statistic since σi's are unknown (only s1 and s2 are given)

H1: μ1 < μ2 Machine 1 produces lower % scrap than machine 2

t-crit = t(0.05)(78) = TINV(0.1,78) -1.664

Machine 1 Machine 2

Mean 3.483 3.668

Standard Error 0.072 0.072

Median 3.41 3.705

Mode 2.97 3.77

Standard Deviation 0.511 0.392

Sample Variance 0.261 0.153

Skewness 0.331 0.014

Range 1.83 1.59

Minimum 2.67 2.88

Maximum 4.5 4.47

Sum 174.14 110.03

Count 50 30

t-stat s² (pooled variance) = ((50-1)*0.511²+(30-1)*0.392²)/(50+30-2)

0.221169038

std error = √(0.221169*(1/50+1/30))

0.108607919

t-stat = ((3.483-3.668)-(0))/0.10861

-1.7032

(a) (iii) Test performed using t-Test Assuming Equal Variances in Data Analysis

t-Test: Two-Sample Assuming Equal Variances

Machine 1 Machine 2

Mean 3.483 3.668

Variance 0.261 0.153

Observations 50 30

Pooled Variance 0.221

Hypothesized Mean Difference 0

df 78

t Stat -1.703

P(T<=t) one-tail 0.046

t Critical one-tail -1.665

P(T<=t) two-tail 0.093

t Critical two-tail 1.991

Since t-stat (= -1.7032) lies outside (below) the region of acceptance, there is sufficient

or p-value = 0.0463 < α = 0.05 (see below), there is sufficient sample evidence

sample evidence at the 5% significance level to reject H0 in favour of H1. (i.e. Reject H0).

Management conclusion

Conclude that machine 1 has a significantly lower average daily % scrap than

machine 2.

Since the p-value (=0.046) < α = 0.05, there is strong sample evidence in

support of H1. Conclude that machine 1 produces scrap at a significantly lower

average daily rate than machine 2. This conclusion is valid at the 5% significance level.

Exercise 9.19 File: X9.19 - water purification.xlsx

2 2

H0: σ 1 =σ 2 Two tailed test

H1: σ²1 ≠ σ²2

Since F-stat (1.063) < F-crit (1.924), do not reject H0 in favour of H1.

Use the pooled-variances t -test to test for differences in means.

Free State KZN

Mean 27.458 26.448

Variance 1.563 1.470

Observations 24 29

df 23 28

F-stat 1.063

P(F<=f) one-tail 0.4342

F Critical one-tail 1.924

(a) (ii) Let population 1 = Free State plant daily impurities levels

Let population 2 = KZN plant daily impurities levels

Let μi = population average daily impurities level per plant i .

(a) (iii) Use the pooled-variances t test statistic since σi's are unknown (only s1 and s2 are given)

H1: μ1 > μ2 FS treatment plant has higher level of impurities

than the KZN water treatment plant.

t-crit = t(0.01)(51) = TINV(0.02,51) 2.402

Mean 27.458 26.448

Standard Error 0.255 0.225

Median 27.5 27

Mode 28 27

Standard Deviation 1.250 1.213

Sample Variance 1.563 1.470

Skewness 0.030 -0.193

Range 5 5

Minimum 25 24

Maximum 30 29

Sum 659 767

Count 24 29

Confidence Level(99.0%) 0.717 0.622

1.51247

std error = √(1.51247*(1/24+1/29))

0.33937

t-stat = ((27.458-26.448)-(0))/0.33937

2.9764

Test performed using t-Test Assuming Equal Variances in Data Analysis

t-Test: Two-Sample Assuming Equal Variances

Mean 27.458 26.448

Variance 1.563 1.470

Observations 24 29

Pooled Variance 1.512

Hypothesized Mean Difference 0

df 51

t Stat 2.9764

P(T<=t) one-tail 0.00223

t Critical one-tail 2.402

P(T<=t) two-tail 0.004

t Critical two-tail 2.676

Since t-stat (2.9764) lies outside (above) the region of acceptance, there is sufficient

sample evidence at the 1% significance level to reject H 0 in favour of H1. (i.e. Reject H0).

Management conclusion

Conclude that the KZN plant produces water of a higher quality (fewer impurities, on average)

than the Free State plant. Thus the KZN plant manager's claim can be supported

statistically at the 1% level of significance.

Since the p-value (=0.00223) << α = 0.01, there is overwhelming sample evidence in

support of H1. Same conclusion as in (a) applies.

Exercise 9.20 File: X9.20 - herbal tea.xlsx

H1: σ²1 ≠ σ²2

Since F-stat (1.227) < F-crit (2.098), do not reject H0 in favour of H1.

Use the pooled-variances t -test to test for differences in means.

Freshpak Yellow Label

Mean 7.689 6.917

Variance 2.170 1.768

Observations 19 23

df 18 22

F-stat 1.227

P(F<=f) one-tail 0.3205

F Critical one-tail 2.098

Let population 2 = Yellow Label brand

Let μi = population mean level of quercetin in mg/kg in each brand i .

Use the t test statistic since σi's are unknown (only s1 and s2 are given)

H1: μ1 ≠ μ2

t-crit = t(0.05)(40) = 2.021

Mean 7.689 6.917

Standard Error 0.3379 0.2772

Median 7.9 7.1

Mode 7.9 5.7

Standard Deviation 1.4731 1.3296

Sample Variance 2.1699 1.7679

Skewness -0.1752 0.1030

Range 5.1 5.3

Minimum 5 4.5

Maximum 10.1 9.8

Sum 146.1 159.1

Count 19 23

t-stat s² (pooled variance) = ((19-1)*1.473²+(23-1)*1.3296²)/(19+23-2)

1.948688

0.43277

t-stat = ((7.689-6.917)-(0))/0.43277

1.784

t-Test: Two-Sample Assuming Equal Variances

Mean 7.6895 6.9174

Variance 2.1699 1.7679

Observations 19 23

Pooled Variance 1.9488

Hypothesized Mean Difference 0

df 40

t Stat 1.7840

P(T<=t) one-tail 0.0410

t Critical one-tail 1.6839

P(T<=t) two-tail 0.0820

t Critical two-tail 2.0211

Statistical conclusion

Since t-stat (1.784) lies within the region of acceptance, there is insufficient sample

evidence at the 5% significance level to reject H0 in favour of H1. (i.e. Accept H0).

Management conclusion

Conclude that there is no difference in the mean quercetin content (in mg/kg)

between the Freshpak and Yellow Labels brands of rooibos tea.

(b) (ii) Let population 1 = Freshpak brand (FP)

Let population 2 = Yellow Label brand (YL)

H1: μ1 > μ2 FP contains more quercetin than YL

t-Test: Two-Sample Assuming Equal Variances

Mean 7.6895 6.9174

Variance 2.1699 1.7679

Observations 19 23

Pooled Variance 1.9488

Hypothesized Mean Difference 0

df 40

t Stat 1.7840

P(T<=t) one-tail 0.0410

t Critical one-tail 1.6839

P(T<=t) two-tail 0.0820

t Critical two-tail 2.0211

Statistical conclusion

Since t-stat (= 1.784) lies outside (above) the region of acceptance (t-stat ≤ 1.6839),

(see table above), there is sufficient sample evidence at the 5% level of

significance to reject H0 in favour of H1. (i.e. Reject H0).

Management conclusion

Conclude that Freshpak's claim that their brand contains more quercetin, on average,

than the Yellow Label brand, can be supported statistically at the 5% significance level.

Exercise 9.21 File: X9.21 - meat fat.xlsx

H1: σ²1 ≠ σ²2

Since F-stat (1.261) < F-crit (2.066), do not reject H0 in favour of H1.

Use the pooled-variances t -test to test for differences in means.

Namibia Little Karoo

Mean 27.074 30.333

Variance 33.840 26.833

Observations 27 21

df 26 20

F 1.261

P(F<=f) one-tail 0.3002

F Critical one-tail 2.066

Let population 2 = Little Karoo meat producer

Let μi = population average fat content of meat supplied by each producer i .

Use the t test statistic since σi's are unknown (only s1 and s2 are given)

H1: μ1 < μ2

(a) (iii) Test performed manually using Descriptive Statistics from Data Analysis

t-crit = t(0.01)(46) = =TINV(0.02,46) 2.412

30.794

1.61459

t-stat = ((27.0741-30.3333)-(0))/1.61459

-2.0186

Test performed using t-Test Assuming Equal Variances in Data Analysis

t-Test: Two-Sample Assuming Equal Variances

Mean 27.0741 30.3333

Variance 33.8405 26.8333

Observations 27 21

Pooled Variance 30.7939

Hypothesized Mean Difference 0

df 46

t Stat -2.0186

P(T<=t) one-tail 0.0247

t Critical one-tail -2.4102

P(T<=t) two-tail 0.0494

t Critical two-tail 2.6870

Since t-stat (-2.0186) lies within the region of acceptance, there is insufficient sample

evidence at the 1% significance level to reject H0 in favour of H1. (i.e. Accept H0).

Management conclusion

Conclude that the mean fat content of meat between the Namibian producer and the

Little Karoo producer is the same .

exclusive agreement with the Namibian producer.

This is the same p-value as shown in the Data Analysis output for a one-tailed test.

Since p-value = 0.0247 > α = 0.01 (see Table above), there is insufficient sample evidence

to reject H0 in favour of H1 at the 1% level of signficance.

Same management conclusion as in (a) (iii) above.

Exercise 9.22 File: X9.22 - disinfectant sales.xlsx

The same retail outlets were surveyed both before and after the

promotional campaign. Thus the two samples are not independent.

x2 = Sales per outlet after the promotional campaign

Let μd = population mean difference in sales from before to after the promotional campaign.

H1: μd < 0 'before' sales are lower than the 'after' sales

t-crit = t(0.05)(11) = -1.796

n 12

x(bar)d -0.667

sd 1.231

-1.877

Statistical conclusion

Since t-stat (-1.877) lies outside (below) the region of acceptance, there is sufficient

sample evidence at the 5% significance level to reject H0 in favour of H1. (i.e. Reject H0).

Management conclusion

Conclude that there has been a significant increase in mean sales of 500ml bottles

of disinfectant liquid from before to after the promotional campaign.

increasing mean sales volume of the product - at the 5% significant level.

(d) t-Test: Paired Two Sample for Means using Data Analysis

Before After

Mean 11.5 12.17

Variance 4.455 3.606

Observations 12 12

Pearson Correlation 0.817

Hypothesized Mean Difference 0

df 11

t Stat -1.876

P(T<=t) one-tail 0.0437

t Critical one-tail 1.7959

P(T<=t) two-tail 0.0874

t Critical two-tail 2.2010

Statistical conclusion

Since p -value = 0.0437 < α = 0.05, there is moderately strong sample

evidence to reject H0 in favour of H1 and conclude that the promotional

campaign has been effective.

Exercise 9.23 File: X9.23 - performance ratings.xlsx

(a) The samples are dependent as the same employee is tested both before and

after the training sessions.

x1 = Performance rating before the training sessions

x2 = Performance rating after the training sessions

Let μd = population mean difference in performance rating scores from

before to after the training sessions.

H1: μd < 0 'before rating scores lower than 'after' rating scores.

t-crit = t(0.05)(17) = -1.74

n 18

x(bar)d -0.356

sd 0.7114

-2.123

Before After

Mean 11.067 11.422

Variance 6.024 5.835

Observations 18 18

Pearson Correlation 0.95743865

Hypothesized Mean Difference 0

df 17

t Stat -2.120

P(T<=t) one-tail 0.0245

t Critical one-tail 1.740

P(T<=t) two-tail 0.04898861

t Critical two-tail 2.10981556

Statistical conclusion

Since t-stat (-2.123) lies outside (below) the region of acceptance, there is sufficient

sample evidence at the 5% significance level to reject H0 in favour of H1 (i.e. Reject H0)

Management conclusion

Conclude that there has been a significant increase in mean performance ratings

scores of employees who attended the series of workshops ans seminars.

motivation and productivity - at the 5% level of significance.

Since p-value = 0.0244 < α = 0.05, there is strong sample evidence to reject H0

in favour of H1 and conclude that the workshops have significantly increased

employee motivation and productivity.

Exercise 9.24 File: X9.24 - household debt.xlsx

(a) The samples are dependent as the same household is tested both a year ago

and at the current time period .

x1 = Household debt level a year ago.

x2 = Household debt level currently.

Let μd = population mean difference in household debt levels from a year ago

to the current period.

H1: μd > 0 Debt higher a year ago than today (current period)

t-crit = t(0.05)(9) = 1.833

n 10

x(bar)d 1.2

sd 1.8135

2.0925

Mean 40.5 39.3

Variance 35.611 36.011

Observations 10 10

Pearson Correlation 0.9541

Hypothesized Mean Difference 0

df 9

t Stat 2.0925

P(T<=t) one-tail 0.0330

t Critical one-tail 1.8331

P(T<=t) two-tail 0.0659

t Critical two-tail 2.2622

Statistical conclusion

Since t-stat (2.0925) lies outside (above) the region of acceptance, there is sufficient

sample evidence at the 5% significance level to reject H0 in favour of H1. (i.e. Reject H0).

Management conclusion

Conclude that there has been a significant decrease in the average level of

household debt from a year ago.

The increase in prime interest rate (from 6% to 11%) has lead to a significant decline

in the average level of household debt from a year ago - at the 5% significance level.

Since the p -value (0.033) < α = 0,05, there is strong sample evidence to reject H0 in

favour of H1 at the 5% level of significance.

Same statistical and management conclusions as (b) above.

Exercise 9.25 Process output variability study

unit of measure: units produced per hour

2 2

H0 σ (1) = σ (2) Management question in H0

H1: 2

σ (1) ≠ σ (2) 2

Note: Set up the F-test as an upper tailed test (F-stat = larger s2/smaller s2)

Decision rule: Do not reject H0 if F-stat (the sample evidence) ≤ 1.939

s12 = 14.6 s22 = 23.2

n1 = 25 n2 = 31

p -value = 0.1240

Statistical conclusion

Since F-stat = 1.589 < F-crit = 1.939, there is insufficient sample evidence at the

5% level of significance to reject H0 in favour of H1.

Management conclusion

Therefore conclude, with 95% confidence, that the variability of hourly outputs

between the two production processes is the same.

Exercise 9.26 File: X9.26 - milk yield.xlsx

Random variable: milk yield (in litres per week) per cow

H1: σ2(f) > σ2(c) Management question in H1

Region of Acceptance: Use α = 0.05 with df1 = 16-1 = 15 and df2 = 16-1 = 15

F-crit = F(0.05,15,15) = 2.403 (See Excel output)

Decision rule: Do not reject H0 if F-stat ≤ 2.403

Free grazing Controlled Feed

Mean 31.306 35.375

Variance 72.138 43.110

Observations 16 16

df 15 15

F-stat 1.673

P(F<=f) one-tail 0.1647

F Critical one-tail 2.403

Statistical conclusion

Since F-stat = 1.673 < F-crit = 2.403, there is insufficient sample evidence at the

5% level of significance to reject H0 in favour of H1.

Management conclusion

Therefore conclude, with 95% confidence, that there is no signficant difference

in the variability in milk yields of cows between the two feeding practices.

Exercise 9.27 File: X9.27 - employee wellness.xlsx

H1: σ2(over 40) > σ2(under 40) Management question in H1

Region of Acceptance: Use α = 0.05 with df1 = 23-1 = 22 and df2 = 21-1 = 20

F-crit = F(0.05,22,20) = 2.102 (See Excel output)

Decision rule: Do not reject H0 if F-stat ≤ 2.102

Over 40 Under 40

Mean 2.461 3.086

Variance 0.836 0.373

Observations 23 21

df 22 20

F-stat 2.240

P(F<=f) one-tail 0.0373

F Critical one-tail 2.102

Statistical conclusion

Since F-stat = 2.24 > F-crit = 2.102, there is sufficient sample evidence at the

5% level of significance to reject H0 in favour of H1.

Management conclusion

Therefore conclude, with 95% confidence, that 'over 40' employees do indeed

exercise more 'erractically' (i.e. show significantly greater variability in exercise times)

than 'under 40' employees.

Exercise 9.28 File: X9.28 - attrition rate.xlsx

Random variable Attrition rates (%) per call center per month

2 2

H0: σ (fin) = σ (health) Management question in H0

2 2

H1: σ (fin) ≠ σ (health)

This depends on whether the smaller or the larger sample variance is in the numerator of F-stat

F-stat = Larger variance / Smaller variance F-stat = Smaller variance / Larger variance

Health Financial Financial Health

Mean 5.46 6.13 Mean 6.13 5.46

Variance 1.1413 0.6653 Variance 0.6653 1.1413

Observations 17 21 Observations 21 17

df 16 20 df 20 16

F-stat 1.715 F-stat 0.583

P(F<=f) one-tail 0.1263 P(F<=f) one-tail 0.1263

F-crit (upper) 2.184 F-crit (lower) 0.458

OR F-crit (lower) = 1/F(0.05,16,20) =0.6653/1.1413 0.458

(or F(0.95,20,16))

Decision rule: For upper tailed test: Do not reject H0 if F-stat ≤ 2.184

OR For lower tailed test: Do not reject H0 if F-stat ≥ 0.458

Now F-stat = 1.715 (for an upper tailed test); or F-stat = 0.583 (for a lower tailed test)

Since F-stat = 1.715 > F-crit (upper) = 2.184 (and hence lies within the acceptance region), there

is insufficient sample evidence at the 5% level of significance to reject H0 in favour of H1.

Same Conclusion (based on a lower tailed F-test)

Since F-stat = 0.583 > F-crit (lower) = 0.458 (and hence lies within the acceptance region), there

is insufficient sample evidence at the 5% level of significance to reject H0 in favour of H1.

Management conclusion

With 95% confidence, it can be concluded that there is no significant difference in the variability

in attrition rates between the two sectors (financial and health).

CHAPTER 10

Exercise 10.1 Purpose: To test whether there is any statistically significant association between

the outcomes of two categorical variables. Stated differently, are the outcomes

associated with two categorical variables independent of each other or not?

Exercise 10.3 H0: There is no statistical association between the two categorical variables

Exercise 10.4 Expected frequencies represent the null hypothesis of no association (or statistical

independence ) between the two categorical variables.

Exercise 10.6 File: X10.6 - motivation status.xlsx

Motivation level

High Moderate Low Total

Male 26.7 26.7 46.7 100

Female 47.5 30.0 22.5 100

Total 38.6 28.6 32.9 100

When compared to the general population profile, males tend to have low levels of

motivation, while females tend to be more highly motivated. It appears

therefore that a statistical association exists between gender and motivation level.

H1: There is an association between Gender and Motivation level

χ²-crit = χ²(0.10)(2) = 4.605

Decision rule

Reject H0 in favour of H1 if χ²-stat ≥ 4.605

High Moderate Low Total

Male 8 8 14 30

Female 19 12 9 40

Total 27 20 23 70

High Moderate Low Total

Male 11.57 8.57 9.86 30

Female 15.43 11.43 13.14 40

Total 27 20 23 70

Chi-Squared components

High Moderate Low

Male 1.102 0.038 1.741

Female 0.827 0.029 1.306

χ²-stat = 5.0428

Conclusion

Since χ²-stat = 5.0428 > χ²-crit = 4.605, there is sufficient sample evidence at the

10% significance level to reject H0 in favour of H1. Therefore conclude that there is a

statistical association between the gender of an employee and their level of motivation.

The nature of the relationship is described in (a) above.

Exercise 10.7 File: X10.7 - internet shopping.xlsx

Internet shopping

Yes No Total

full-time 24.3 75.7 100

at-home 18.5 81.5 100

Total 20.8 79.2 100

Since the row percentage profiles (between full-time employed and at-home customers) are

very similar to each other and to the general population profile, it can be concluded, by

observation, that the two attributes are not associated (i.e. they are statistically independent).

(b) H0: There is no association between Employment Status and Use of Internet Shopping

H1: There is an association between Employment Status and Use of Internet Shopping

χ²-crit = χ²(0.05)(1) = 3.843

Decision rule

Reject H0 in favour of H1 if χ²-stat ≥ 3.843

Yes No Total

full-time 35 109 144

at-home 40 176 216

Total 75 285 360

Yes No Total

full-time 30 114 144

at-home 45 171 216

Total 75 285 360

Chi-Squared components

Yes No

full-time 0.8333 0.2193

at-home 0.5556 0.1462

χ²-stat = 1.7544

Conclusion

Since χ²-stat = 1.7544 < χ²-crit = 3.843, there is insufficient sample evidence at the

5% significance level to reject H0 in favour of H1. Therefore conclude that there is

no statistical association between the employment status of a customer and their

use of the internet for shopping purposes.

These two events are statistically independent.

Exercise 10.8 File: X10.8 - car size.xlsx

Small Medium Large Total

Under 30 15.2 33.3 51.5 100

30 - 45 21.1 36.8 42.1 100

Over 45 37.5 29.2 33.3 100

Total 26.3 33.0 40.7 100

With reference to the general population profile (Total row %),

under 30's tend to prefer larger cars; 30-45 year age car buyers marginally tend

towards medium to large cars, while over 45's strongly tend to prefer smaller cars.

(b) H0: There is no association between Age of car buyer and Car Size bought.

H1: There is an association between Age of car buyer and Car Size bought.

χ²-crit = χ²(0.01)(4) = 13.277

Decision rule

Reject H0 in favour of H1 if χ²-stat ≥ 13.277

Small Medium Large Total

Under 30 10 22 34 66

30 - 45 24 42 48 114

Over 45 45 35 40 120

Total 79 99 122 300

Small Medium Large Total

Under 30 17.38 21.78 26.84 66

30 - 45 30.02 37.62 46.36 114

Over 45 31.6 39.6 48.8 120

Total 79 99 122 300

Chi-Squared components

Small Medium Large

Under 30 3.134 0.002 1.910

30 - 45 1.207 0.510 0.058

Over 45 5.682 0.534 1.587

χ²-stat = 14.6247

Conclusion

Since χ²-stat = 14.6247 > χ²-crit = 13.277, there is sufficient sample evidence at the

1% significance level to reject H0 in favour of H1. Therefore conclude that there is a

statistical association between the age of a car buyer and the size of car bought.

(c) Management Conclusion

The nature of the statistical relationship found in (b) is described in (a) above.

Under 30's tend to prefer larger cars; 30-45 year age car buyers marginally tend

towards medium to large cars, while over 45's strongly tend to prefer smaller cars.

Recommendation

Target larger cars to the younger market and smaller cars to the older market.

Exercise 10.9 File: X10.9 - sports readership.xlsx

(a) Let πi = proportion of people who read Sports News in each of the i regions.

H0: π1 = π2 = π3

H1: At least one πi is different (i = 1,2,3)

0.160 0.104 0.250

χ²-crit = χ²(0.01)(2) = 9.21

Decision rule

Reject H0 in favour of H1 if χ²-stat ≥ 9.21

E Cape W Cape KZN Total

No 84 86 78 248

Yes 16 10 26 52

Total 100 96 104 300

E Cape W Cape KZN Total

No 82.67 79.36 85.97 248

Yes 17.33 16.64 18.03 52

Total 100 96 104 300

Chi-Squared components

E Cape W Cape KZN

No 0.0215 0.5556 0.7395

Yes 0.1026 2.6496 3.5267

χ²-stat = 7.5954

Conclusion

Since χ²-stat = 7.5954 < χ²-crit = 9.21, there is insufficient sample evidence at the

1% significance level to reject H0 in favour of H1. Therefore conclude that the

proportion of people who read Sports News is the same in each Geographical Region.

These two events are statistically independent.

(c) H0: There is no association between the propensity to read Sports News and Region

H1: There is an association between the propensity to read Sports News and Region

Exercise 10.10 File: X10.10 - gym activity.xlsx

Gym activity

Spinning Swimming Circuit Total

Male 42.35 22.35 35.29 100

Female 52.73 29.09 18.18 100

Total 46.43 25.00 28.57 100

When compared to the general population of gym goers, more females than males

tend to prefer spinning and swimming; while more males relative to females tend to

prefer mainly doing the circuit.

The evidence is however not strongly convincing.

(b) H0: There is no association between Gender and preferred Gym Activity

H1: There is an association between Gender and preferred Gym Activity

χ²-crit = χ²(0.10)(2) = 4.605

Decision rule

Reject H0 in favour of H1 if χ²-stat ≥ 4.605

Spinning Swimming Circuit Total

Male 36 19 30 85

Female 29 16 10 55

Total 65 35 40 140

Spinning Swimming Circuit Total

Male 39.46 21.25 24.29 85

Female 25.54 13.75 15.71 55

Total 65 35 40 140

Chi-Squared components

Spinning Swimming Circuit

Male 0.304 0.238 1.345

Female 0.470 0.368 2.078

χ²-stat = 4.803

Conclusion

Since χ²-stat = 4.803 > χ²-crit = 4.605 (marginally), there is sufficient sample evidence

at the 10% significance level to reject H0 in favour of H1. Therefore conclude that

gender and preferred gym activity are associated.

The nature of the relationship is described in (a) above.

(c) New rejection region Use α = 0.05 with degrees of freedom = (2-1)(3-1) = 2

χ²-crit = χ²(0.05)(2) = 5.991

Decision rule Reject H0 in favour of H1 if χ²-stat ≥ 5.991

since χ²-stat = 4.803 < χ²-crit = 5.991.

New conclusion There is no statistical association between gender and gym activity

(i.e. they are statistically independent) at the 5% level of significant.

(d) Let πi = proportion of females who prefer each gym activity i (spinning, swimming, circuit)

H0: π1 = π2 = π3

H1: At least one πi is different (i = 1,2,3)

0.446 0.457 0.250

Statistical conclusion

The same statistical conclusion applies as in (b) (i.e. a statistical association exists)

or stated differently: at least one population proportion is different.

Management conclusion

By an inspection of the row percentage table in (a), it can be concluded that

spinning and swimming are the most preferred gym activities of females,

while doing the circuit is the least preferred gym activity of females.

Exercise 10.11 File: X10.11 - supermarket visits.xlsx

Daily 36 20.0 25

3 / 4 times 55 30.6 35

Twice 62 34.4 30

Once only 27 15.0 10

Total 180 100 100

Once-a-week visits are the least common shopping behaviour (only 15%).

Only one-in-five shoppers (20%) shop daily.

The most common shopping pattern is either 3/4 times a week (30.6%) or twice a week (34.4%).

These two shopping behaviours represent 65% of all the sampled shoppers.

H0: The frequency of store visits per week is as per the manager's belief.

H1: The frequency of store visits per week differs significantly from the manager's belief.

χ²-crit = χ²(0.05)(3) = 7.815

Decision rule

Reject H0 in favour of H1 if χ²-stat ≥ 7.815

χ²-stat Customers

Visits fo %fe fe χ²-stat

Daily 36 25 45 1.8

3 / 4 times 55 35 63 1.016

Twice 62 30 54 1.185

Once only 27 10 18 4.5

Total 180 100 180 8.501

Conclusion

Since χ²-stat = 8.501 > χ²-crit = 7.815, there is sufficient sample evidence

at the 5% significance level to reject H0 in favour of H1. Therefore conclude that

the shopping frequency of customers differs significantly from the manager's belief.

The nature of the relationship is described in (a) above.

(c) Interpretation

The manager's belief is that customers shop more frequently during a week.

frequently than the manager believes. For example, the survey found that

only 51% shop more than 3 times per week, while the manager assumed that this

percentage was 60%. Similarly, more customers prefer to shop only once or twice a

week (49%) compared to the manager's belief that this percentage was only 40%).

These differences are however not strongly significantly different.

Exercise 10.12 File: X10.12 - equity portfolio.xlsx

H0: There is no change in the equity portfolio mix between 2008 and 2012.

H1: There is a significant change in the equity portfolio mix between 2008 and 2012.

χ²-crit = χ²(0.05)(3) = 7.815

Decision rule

Reject H0 in favour of H1 if χ²-stat ≥ 7.815

χ²-stat Equities

Equity fo (2012) Ratio fe fe (2008) χ²-stat

Mining 900 2 900 0

Industrial 1400 3 1350 1.852

Retail 400 1 450 5.556

Financial 1800 4 1800 0

Total 4500 10 4500 7.407

Conclusion

Since χ²-stat = 7.407 < χ²-crit = 7.815, there is insufficient sample evidence

at the 5% significance level to reject H0 in favour of H1. Therefore conclude that

there is no significant change in the equity portfolio mix of the investor

between 2008 and 2012.

The equity portfolio profile is essentially the same in 2012 as it was in 2008.

Exercise 10.13 File: X10.13 - payment method.xlsx

H1: There is a significant change in the payment method for electronic goods.

χ²-crit = χ²(0.05)(2) = 5.991

Decision rule

Reject H0 in favour of H1 if χ²-stat ≥ 5.991

Payment fo % fe fe χ²-stat

Cash 41 23 46 0.543

Debit Card 49 35 70 6.3

Credit Card 110 42 84 8.048

Total 200 100 200 14.891

Conclusion

Since χ²-stat = 14.891 >> χ²-crit = 5.991, there is sufficient sample evidence

at the 5% significance level to reject H0 in favour of H1. Therefore conclude that

there is a significant shift in payment practices from the past.

There is a significant shift in payment practices for electronic goods.

There is more emphasis on credit card payment (55%) today than in the past (42%).

Exercise 10.14 File: XS10.14 - package sizes.xlsx

H1: Limpopo sales pattern does not follow the national sales pattern.

χ²-crit = χ²(0.05)(2) = 5.991

Decision rule

Reject H0 in favour of H1 if χ²-stat ≥ 5.991

Package fo Ratio fe fe χ²-stat

Large 190 3 162 4.840

Midsize 250 5 270 1.481

Small 100 2 108 0.593

Total 540 10 540 6.914

Conclusion

Since χ²-stat = 6.914 > χ²-crit = 5.991, there is sufficient sample evidence

at the 5% significance level to reject H0 in favour of H1. Therefore conclude that

the Limpopo sales pattern of cereal package sizes differs significantly from the

national sales pattern of package sizes sold.

The Limpopo sales patterns differs significantly from the national sales pattern

of package sizes sold. By an inspection of the Limpopo sales profile relative to the

national pattern, Limpopo tends to sell more large sized packages relative

to the national pattern.

Exercise 10.15 File: X10.15 - compensation plan.xlsx

Regions

Plan Cape Gauteng Free State KZN Total

Present 62 75.7 67.1 72.7 70.8

New 38 24.3 32.9 27.3 29.2

Total 100 100 100 100 100

When compared to the national population profile of all employees,

the present compensation plan enjoys more support in Gauteng, and the

least support in the Cape Province where the new plan is favoured more.

The evidence is however not overwhelming (i.e. the profile differences are not large)

(b) Let πi = proportion of sales staff in favour of the present payment plan in each province i .

H0: π1 = π2 = π3 = π4

H1: At least one πi is different (i = 1,2,3,4)

0.62 0.757 0.671 0.727

χ²-crit = χ²(0.10)(3) = 6.251

Decision rule

Reject H0 in favour of H1 if χ²-stat ≥ 6.251

Plan Cape Gauteng Free State KZN Total

Present 62 140 47 80 329

New 38 45 23 30 136

Total 100 185 70 110 465

Plan Cape Gauteng Free State KZN Total

Present 70.75 130.89 49.53 77.83 329

New 29.25 54.11 20.47 32.17 136

Total 100 185 70 110 465

Chi-Squared components

Plan Cape Gauteng Free State KZN

Present 1.0828 0.6337 0.1289 0.0606

New 2.6194 1.5330 0.3119 0.1466

χ²-stat = 6.5169

Conclusion

Since χ²-stat = 6.5169 > χ²-crit = 6.251 (marginally), there is sufficient sample evidence

at the 10% significance level to reject H0 in favour of H1. Therefore conclude that

the support for the present compensation plan is different in at least one of the provinces.

The nature of the relationship is described in (a) above.

(d) Formulate as a test for independence of association between payment plan and province.

H0: There is no association between payment plan preference and province.

H1: There is an association between payment plan preference and province.

(e) New rejection region Use α = 0.05 with degrees of freedom = (2-1)(4-1) = 3

χ²-crit = χ²(0.05)(3) = 7.815

Decision rule Reject H0 in favour of H1 if χ²-stat ≥ 7.815

since χ²-stat = 6.5169 < χ²-crit = 7.815.

New conclusion

There is no statistical association between payment plan preference and province.

(i.e. they are statistically independent) at the 5% level of significant.

The sample evidence is not strong enough (i.e. the sample proportion differences are

not great enough) to reject H0 in favour of H1 at the 5% level of significance.

Exercise 10.16 File: X10.16 - tyre defects.xlsx

technical mechanical material Total

Morning 22.1 61.8 16.2 100

Afternoon 30.2 46.5 23.3 100

Night 42.6 36.8 20.6 100

Total 31.5 48.2 20.3 100

When compared to the total production of defective tyres (total row %),

tyre defects due to mechanical problems tend to be more prevalent during morning shifts.

Technical defects however tend to be more prevalent during the night shift.

Thus there does appear to be an association between shift and nature of tyre defects.

(b) Formulate as a test for independence of association between nature of defect and shift

H1: There is an association between nature of tyre defect and shift.

χ²-crit = χ²(0.05)(4) = 9.488

Decision rule

Reject H0 in favour of H1 if χ²-stat ≥ 9.488

technical mechanical material Total

Morning 15 42 11 68

Afternoon 26 40 20 86

Night 29 25 14 68

Total 70 107 45 222

technical mechanical material Total

Morning 21.44 32.77 13.78 68

Afternoon 27.12 41.45 17.43 86

Night 21.44 32.77 13.78 68

Total 70 107 45 222

Chi-Squared components

technical mechanical material

Morning 1.9351 2.5967 0.5622

Afternoon 0.0460 0.0508 0.3782

Night 2.6646 1.8443 0.0034

χ²-stat = 10.0812

(c) Conclusion

Since χ²-stat = 10.0812 > χ²-crit = 9.488, there is sufficient sample evidence

at the 5% significance level to reject H0 in favour of H1. Therefore conclude that

the nature of tyre defects produced is related to the shift on which the defects occur.

The nature of the relationship is described in (a) above.

(d) Let πi = proportion of defective tyres caused by mechanical factors per shift i .

H0: π1 = π2 = π3

H1: At least one πi is different (i = 1,2,3)

The sample proportions being compared are: Morning Afternoon Night

0.618 0.465 0.368

Conclusion

Since H0 is rejected in favour of H1 at the 5% significance level, it can be concluded that there

is at least one shift that has a different proportion of defective tyres due to mechanical factors.

Based on the row percentages table in (a) above, it is clear that the morning shift

produces a proportionally larger percentage of defective tyres due to mechanical

factors than the afternoon or night shifts.

Exercise 10.17 File: X10.17 - flight delays.xlsx

(a) Histogram

Histogram of Flight Delay Times

Delays(min) Count

<5 0

5 - 7.5 8 35

7.5 - 10 28 28 29

30

10-12.5 29

25

12.5-15 13

No. of flights

15-17.5 2 20

Total 80 13

15

10 8

5 2

0

0

<5 5 - 7.5 7.5 - 10 10-12.5 12.5-15 15-17.5

Mean 10.324

Standard Error 0.261

Median 10.3

Mode 11

Standard Deviation 2.333

Sample Variance 5.444

Kurtosis -0.550

Skewness 0.084

Range 10.3

Minimum 5.3

Maximum 15.6

Sum 825.9

Count 80

H0: Flight time delays (in minutes) follows a normal distribution with μ = 10.324 min and σ = 2.333 min.

H1: Flight time delays (in minutes) do not follow a normal distribution with μ = 10.324 min and σ = 2.333 min.

χ²-crit = χ²(0.01)(4) = 13.277

Decision rule

Reject H0 in favour of H1 if χ²-stat ≥ 13.277

Delay Intervals fo Normal probability intervals (x and z) Probability fe χ²-stat

-∞ < x < 5 0 P(x < 5) P(z < -2.282) 0.01130 0.90 0.9040

5 < x < 7.5 8 P(5 < x < 7.5) P(-2.282 < z < -1.21) 0.10180 8.14 0.0025

7.5 < x < 10 28 P(7.5 < x < 10) P(-1.21 < z < -0.139) 0.33120 26.50 0.0854

10 < x < 12.5 29 P(10 < x < 12.5) P(-0.139 < z < 0.933) 0.37950 30.36 0.0609

12.5 < x < 15 13 P(12.5 < x < 15) P(0.933 < z < 2.004) 0.15400 12.32 0.0375

15 < x < 17.5 2 P(15 < x < 17.5) P(2.004 < z < 3.076) 0.02117 1.69 0.0554

17.5 < x < +∞ 0 P(x > 17.5) P(z > 3.076) 0.00103 0.08 0.0824

80 1 80

χ²-stat = 1.2282

Conclusion

Since χ²-stat = 1.2282 << χ²-crit = 13.277, there is insufficient sample evidence

at the 1% significance level to reject H 0 in favour of H1. Therefore conclude that

flight delay times (in minutes) follows a Normal distribution with μ = 10.324 min and σ = 2.333 min.

(c) (ii) NORMDIST() From Z-table

P(x < 5) =NORMDIST(5,10.324,2.333,1) 0.01124 0.01130

P(5 < x < 7,5) =NORMDIST(7.5,10.324,2.333,1) - NORMDIST(5,10.324,2.333,1) 0.10181 0.10180

P(7,5 < x < 10) =NORMDIST(10,10.324,2.333,1) - NORMDIST(7.5,10.324,2.333,1) 0.33172 0.33120

P(10 < x < 12,5) =NORMDIST(12.5,10.324,2.333,1) - NORMDIST(10,10.324,2.333,1) 0.37974 0.37950

P(12,5 < x < 15) =NORMDIST(15,10.324,2.333,1) - NORMDIST(12.5,10.324,2.333,1) 0.15297 0.15400

P(15 < x < 17,5) =NORMDIST(17.5,10.324,2.333,1) - NORMDIST(15,10.324,2.333,1) 0.02147 0.02117

P(x > 17,5) =1-NORMDIST(17.5,10.324,2.333,1) 0.00105 0.00103

1 1

Exercise 10.18 File: newspaper sections.xlsx

(a) Two-way Pivot table of Gender by Newspaper Section Most Preferred to Read.

Section

Gender Data Sport Social Business Grand Total

Female Count 14 28 18 60

Row % 23.3% 46.7% 30.0% 100%

Male Count 41 35 44 120

Row % 34.2% 29.2% 36.7% 100%

Total Count 55 63 62 180

Total Row % 30.6% 35.0% 34.4% 100%

Females tend to read the Social section most, with the Sports section read the least.

Males, alternatively, are marginally more interested in the Sport and Business sections.

These observational conclusions are, however, not overwhelmingly conclusive.

(b)

Stacked Bar Chart

Newspaper Section Read by Gender

120%

100%

30.0% 36.7%

80%

Business

60% Social

46.7% 29.2% Sport

40%

20% 34.2%

23.3%

0%

Female Male

(c) Formulate as a test for independence of association between gender and section read .

H0: There is no association between gender and the newspaper section most preferred.

H1 : There is an association between gender and the newspaper section most preferred.

χ²-crit = χ²(0.10)(2) = 4.605

Decision rule

Reject H0 in favour of H1 if χ²-stat ≥ 4.605

χ²-stat Observed frequencies (fo)

Sport Social Business Total

female 14 28 18 60

male 41 35 44 120

Total 55 63 62 180

Sport Social Business Total

female 18.3 21.0 20.7 60

male 36.7 42.0 41.3 120

Total 55 63 62 180

Chi-Squared components

Sport Social Business

female 1.0242 2.3333 0.3441

male 0.5121 1.1667 0.1720

χ²-stat = 5.5525

Conclusion

Since χ²-stat = 5.5525 > χ²-crit = 4.605 (marginal), there is sufficient sample evidence

at the 10% significance level to reject H0 in favour of H1. Therefore conclude that

gender and the newspaper section most preferred are statistically associated.

The nature of the relationship is described in (a) above.

(d) Let πi = proportion of females who most prefer each newspaper section i .

H0: π1 = π2 = π3

H1: At least one πi is different (i = 1,2,3)

The sample proportions being compared are:

Sport Social Business

0.255 0.444 0.290

Conclusion

Since H0 is rejected in favour of H1 at the 10% significance level, it can be concluded that

there is at least one newspaper section that females prefer differently to the other sections.

Based on the row percentages table in (a) above, it is clear that

females tend to read the Social section most, with the Sports section read the least.

Exercise 10.19 File: X10.19 - vehicle financing.xlsx

Loan size Total Loan size Total

Under 100 18 Under 100 6%

100 - <150 58 100 - <150 19%

150 - <200 110 150 - <200 37%

200 - <250 70 200 - <250 23%

Above 250 44 Above 250 15%

Grand Total 300 Grand Total 100%

The most popular car loan size was between R150 000 and R200 000 (37% of all

applications) followed by car loan sizes of between R200 000 and R250 000 (23%).

Only 6% of loan applications were for amounts below R100 000.

(b)

Bar Chart of Car Loan Applications

0.4

0.35

0.3

0.25

0.2

0.37

0.15

0.23

0.1 0.19

0.15

0.05

0.06

0

Under 100 100 - <150 150 - <200 200 - <250 Above 250

Size of Loan Applications (R 1000)

H0: There is no change in the size of car loan applications from four years ago.

H1: There is a significant shift in the size of car loan applications from four years ago.

χ²-crit = χ²(0.05)(4) = 9.488

Decision rule

Reject H0 in favour of H1 if χ²-stat ≥ 9.488

Loan Size fo % fe fe χ²-stat

< R100 18 10 30 4.800

R100 - R150 58 20 60 0.067

R150 - R200 110 40 120 0.833

R200 - R250 70 20 60 1.667

> R250 44 10 30 6.533

Total 300 100 300 13.900

Conclusion

Since χ²-stat = 13.9 > χ²-crit = 9.488, there is sufficient sample evidence

at the 5% significance level to reject H0 in favour of H1. Therefore conclude that

there has been a significant shift in the size of car loan applications from 4 years ago.

The shift has been towards larger car loan applications.

Exercise 10.20 File: X10.20 - milk products.xlsx

(a) Two-way Pivot table of milk type purchased and health-concious status of consumers

Question 2

Question 1 Data Yes No Grand Total

Fat-free Count 20 5 25

Percent 43.5% 16.7% 32.9%

Low fat Count 15 10 25

Percent 32.6% 33.3% 32.9%

Full cream Count 11 15 26

Percent 23.9% 50.0% 34.2%

Total Count 46 30 76

Total Percent 100% 100% 100%

The sample evidence points strongly towards health-concious consumers purchasing more

fat-free dairy products, while non-health-concious consumers tend to purchase more

full cream dairy products.

(b)

Stacked Bar Chart

Milk Type Purchased and Health-Concious Status

80%

70%

% of Consumers

60%

50% 16.7% 33.3% 50.0% No

40%

Yes

30%

20% 43.5%

32.6%

10% 23.9%

0%

Fat-free Low fat Full cream

Milk Categories

(c) Formulate as a test for independence of association between milk type purchased and

health-concious status of consumers.

H0: There is no association between milk type purchased and health-concious status.

H1: There is an association between milk type purchased and health-concious status.

χ²-crit = χ²(0.05)(2) = 5.991

Decision rule

Reject H0 in favour of H1 if χ²-stat ≥ 5.991

χ²-stat Observed frequencies (fo)

Yes No Total

fat-free 20 5 25

low fat 15 10 25

full cream 11 15 26

Total 46 30 76

Yes No Total

fat-free 15.13 9.87 25

low fat 15.13 9.87 25

full cream 15.74 10.26 26

Total 46 30 76

Chi-Squared components

Yes No

fat-free 1.566 2.402

low fat 0.001 0.002

full cream 1.426 2.186

χ²-stat = 7.583

Conclusion

Since χ²-stat = 7.583 > χ²-crit = 5.991, there is sufficient sample evidence

at the 5% significance level to reject H0 in favour of H1. Therefore conclude that

there is a significant statistical association between the health-concious status

of a consumer and their preference for certain type of dairy milk products.

The nature of the relationship is described in (a) above.

(d) Let πi = proportion of health-concious consumers who prefer each milk type i .

H0: π1 = π2 = π3

H1: At least one πi is different (i = 1,2,3)

The sample proportions being compared are: fat-free low-fat full cream

0.800 0.600 0.423

Conclusion

Since H0 is rejected in favour of H1 at the 5% significance level, it can be concluded that

there is at least one milk type that is purchased by a different proportion of health-concious

consumers. Based on the row percentages table in (a) above, it is clear that health-concious

consumers tend to purchase more fat-free dairy products than full cream products.

CHAPTER 11

ANALYSIS OF VARIANCE

Exercise 11.1 The purpose of one-factor Anova is to test for equality of means across

multiple (more than two) populations.

Exercise 11.2 Example: Compare the output performance of five identical machines.

Exercise 11.3 Variation between groups measures how similar or how different

(i.e. how close or how far apart) the sample means are from each other.

It is a measure of the level of influence of the treatment factor on the

response measure. Any differences can be attributed to (or explained by)

the influence of the treatment factor on the numeric response measure.

Numerator degrees of freedom = (4 - 1) = 3

SSE = 204.6 - 25.5 = 179.1

Denominator degrees of freedom = (4x10 - 4) = 36

MSE = 179.1 / 36 = 4.975

Exercise 11.6 H 0 : μ 1 = μ2 = μ3 = μ4

H1: At least one μi differs

Since F-stat = 1.7085 < F-crit = 2.866, do not reject H0.

Conclusion: All population means are equal.

Exercise 11.7 File: X11.7 - car fuel efficiency.xlsx

Peugot VW Ford

Σxi 32.4 29.3 34.4

ni 5 4 5

x(bar)i 6.48 7.325 6.88

Grand mean = 96.1/14 = 6.864

10

9

8 7.325

6.88

7 6.48

litres / 100 km

6

5

4

3

2

1

0

Peugot VW Ford

Car Types 6.48 7.325 6.88

Response measure = Fuel Consumption (l/100km)

H0: μ1 = μ2 = μ3

H1: At least one μi differs (i = 1, 2, 3)

Region of Rejection (Use α = 0.05 with df1 = 3-1 = 2 and df2 = 14-3 = 11)

F-crit = F(0.05)(2,11) = 3.98

Decision rule

Reject H0 in favour of H1 if F-stat ≥ 3.98

(6.8-7.325)²+(7.4-7.325)²+(7.9-7.325)²+(7.2-7.325)²+

(7.6-6.88)²+(6.8-6.88)²+(6-4.88)²+(7-6.88)²+(6.6-6.88)²

= 2.0635

SSB = (6.48-6.864)²+(7.325-6.864)²+(6.88-6.864)²

= 1.5886

SST = 2.0635 + 1.5886 = 3.6521

Anova Table

Source of

SS df MS F-stat

Variation

Between 1.5886 2 0.79432 4.234

Within 2.0635 11 0.18759

Total 3.6521 13

Conclusion

Since F-stat = 4.234 > F-crit = 3.98, there is sufficient sample evidence at the

5% significance level to reject H0 in favour of H1. Therefore conclude that there is at

least one motor vehicle type with a different average fuel consumption to the rest.

By inspection, it would appear that VW has an average fuel consumption that is significantly

different (higher, and hence least fuel efficient) from Peugot and Ford.

(c) New Region of Rejection (Use α = 0.01 with df1 = 3-1 = 2 and df2 = 14-3 = 11)

F-crit = F(0.01)(2,11) = 7.21

Decision rule

Reject H0 in favour of H1 if F-stat ≥ 7.21

since F-stat = 4.234 < F-crit = 7.21

New conclusion

There is no statistical evidence, at the 1% signficance level, to conclude that average

fuel consumption differs across motor vehicle types.

Note: The sample evidence must be more convincing (i.e. larger differences between

sample means) before one is prepared to reject the null hypothesis in favour of the

alternative hypothesis.

A level of significance of 1% indicates that the sample evidence is not strong

enough (meaningful differences) yet to reject the null hypothesis of equal means.

Exercise 11.8 File: X11.8 - package design.xlsx

Assumption 2 A normally distribution population for the response variable.

Response measure = Carton sales (units)

H0: μ1 = μ2 = μ3

H1: At least one μi differs (i = 1, 2, 3)

Region of Rejection (Use α = 0.05 with df1 = 3-1 = 2 and df2 = 21-3 = 18)

F-crit = F(0.05)(2,18) = 3.55

Decision rule

Reject H0 in favour of H1 if F-stat ≥ 3.55

Sample evidence

Design A Design B Design C

Sample means 35.75 32.86 34.17

Grand mean 721/21 = 34.33

SSW = (35-35.75)²+(37-35.75)²+(39-35.75)²+(36-35.75)²+(30-35.75)²+(39-35.75)²+(36-35.75)²+(34-35.75)²+

(35-32.86)²+(34-32.86)²+(30-32.86)²+(31-32.86)²+(34-32.86)²+(32-32.86)²+(34-32.86)²+

(38-34.17)²+(34-34.17)²+(32-34.17)²+(34-34.17)²+(34-34.17)²+(33-34.17)²

= 101.1905

SSB = (35.75-34.33)²+(32.86-34.33)²+(34.17-34.33)²

= 31.4762

SST = 101.1905+31.4762 = 132.6667

Anova Table

Source of

SS df MS F-stat

Variation

Between 31.4762 2 15.7381 2.7995

Within 101.1905 18 5.6217

Total 132.6667 20

Conclusion

Since F-stat = 2.7995 < F-crit = 3.55, there is insufficient sample evidence at the

5% significance level to reject H0 in favour of H1. Therefore conclude that there is

no difference in the mean volume of sales across the 3 package designs.

(b) Recommendation

There is no strong statistical evidence to conclude that sales volumes differ across the

three package designs. All are likely to generate the same average sales. Therefore the cereal

producer can choose any of the three package designs for their new muesli cereal.

Exercise 11.9 File: X11.9 - bank service.xlsx

Response measure = Rating score (1 to 10)

H0: μ 1 = μ2 = μ3

H1: At least one μi differs (i = 1, 2, 3)

Region of Rejection (Use α = 0.10 with df1 = 3-1 = 2 and df2 = 27-3 = 24)

F-crit = F(0.10)(2,24) = 2.538 =FINV(0.1,2,24)

Decision rule

Reject H0 in favour of H1 if F-stat ≥ 2.538

Sample evidence

Bank X Bank Y Bank Z

Sample means 6.875 5.778 6.2

Grand mean 169/27 = 6.259

SSW = (8-6.875)²+(6-6.875)²+...+(5-5.778)²+(6-5.778)²+…+(8-6.2)²+(7-6.2)²+…+(6-6.2)²

= 24.0306

SSB = (6.875-6.259)²+(5.778-6.259)²+(6.2-6.259)²

= 5.1546

SST = 24.0306 + 5.1546 = 29.1852

Anova Table

Source of

SS df MS F-stat

Variation

Between 5.1546 2 2.5773 2.574

Within 24.0306 24 1.0013

Total 29.1852 26

Conclusion

Since F-stat = 2.574 > F-crit = 2.538, there is sufficient sample evidence at the

10% significance level to reject H0 in favour of H1. Therefore conclude that there is

at least one bank that has a different mean service rating score to the other banks.

service rating score than the other two banks.

(b) New Region of Rejection (Use α = 0.05 with df1 = 3-1 = 2 and df2 = 27-3 = 24)

F-crit = F(0.05)(2,24) = 3.40

Decision rule

Reject H0 in favour of H1 if F-stat ≥ 3.40

since F-stat = 2.574 < F-crit = 3.40

New conclusion

There is no statistical evidence, at the 5% signficance level, to conclude that the

mean service ratings is different across the three banks.

The three banks are perceived similarly by customers in terms of their service levels.

Note: The reason for the change in conclusion between (a) and (b) is that the

statistical evidence is only weak (i.e. small differences in sample means) at

the 10% significance level, while it the 5% significance level, it is not seen as

strong enough (i.e. differences are not large enough) to reject the null hypothesis.

Exercise 11.10 File: X11.10 - shelf height.xlsx

Response measure = Sales volume (units sold)

Let μi = population mean sales of a drinking chocolate product displayed at shelf height i

H0: μ1 = μ2 = μ3 = μ4

H1: At least one μi differs (i = 1, 2, 3, 4)

Region of Rejection (Use α = 0.05 with df1 = 4-1 = 3 and df2 = 30-4 = 26)

F-crit = F(0.05)(3,26) = 2.990

Decision rule

Reject H0 in favour of H1 if F-stat ≥ 2,990

Sample evidence

Bottom Waist Shoulder Top

Sample means 76.143 81.375 82.444 74.833

Grand mean 2375/30 = 79.167

SSW = (78-76.143)²+….+(78-81.375)²+….+(83-82.444)²+….+(69-74.833)²+…+(75-74.833)²

= 727.788

SSB = (76.143-79.167)²+(81.375-79.167)²+(82.444-79.167)²+(74.833-79.167)²

= 312.379

SST = 727.788 + 312.379 = 1040.167

Anova Table

Source of

SS df MS F-stat

Variation

Between 312.379 3 104.126 3.720

Within 727.788 26 27.992

Total 1040.167 29

Conclusion

Since F-stat = 3.72 > F-crit = 2.99, there is sufficient sample evidence at the

5% significance level to reject H0 in favour of H1. Therefore conclude that there is

at least one shelf height that generates a different mean level of sales to the other shelves.

By inspection, it would appear that shoulder and waist high shelves generate higher average

sales of the drinking chocolate product than bottom or top shelves.

Exercise 11.11 File: X11.11 - machine evaluation.xlsx

Response measure = Processing time (in minutes)

Let μi = population mean processing time for shaping and labelling machine i

H0: μ1 = μ2 = μ3

H1: At least one μi differs (i = 1, 2, 3)

SUMMARY

Groups Count Sum Average Variance

Machine A 5 59 11.8 3.7

Machine B 5 56 11.2 5.7

Machine C 5 72 14.4 3.8

ANOVA

Source of Variation SS df MS F-stat P-value F crit

Between Groups 28.933 2 14.467 3.288 0.0727 2.8068

Within Groups 52.8 12 4.4

Total 81.733 14

Conclusion

Since F-stat = 3.288 > F-crit = 2.8068, there is sufficient sample evidence at the

10% significance level to reject H0 in favour of H1. Therefore conclude that there is at least

machine that has a different mean processing time.

By inspection, machine C has a significantly longer mean processing time than either

machines A and B. Machine C must not be considered for purchase.

(b) Two-Sample Test of Mean Processing Times between Machines A (1) and B (2)

H0: μ1 = μ2 (two tailed test)

H1: μ1 ≠ μ2

Machine A Machine B

Mean 11.8 11.2

Variance 3.7 5.7

Observations 5 5

Pooled Variance 4.7

Hypothesized Mean Difference 0

df 8

t Stat 0.4376

P(T<=t) one-tail 0.3366

t Critical one-tail 1.3968

P(T<=t) two-tail 0.6733 t-crit (0.10, 8) =

t Critical two-tail 1.8595 =TINV(0.1,8) 1.8595

Conclusion

Since t-stat = 0.4376 lies with the region of non-rejection of H 0 (i.e. within ± 1.8595),

the sample evidence is not strong enough to reject H 0 in favour of H1.

The population mean processing times between the two machines A and B are

likely to be identical. Thus the company can purchase either machine A or machine B.

(c) Recommendation

Based on the statistical evidence in (a) and (b), the company can purchase either

machine A or machine B - they are likely to be equally efficient in operation.

Exercise 11.12 File: X11.12 - earnings yield.xlsx

(1 = Financial; 2 = Retail; 3 = Industrial; 4 = Mining)

Response measure = Earnings yield (%)

H0: μ 1 = μ 2 = μ 3 = μ4

H1: At least one μi differs (i = 1, 2, 3, 4)

SUMMARY

Groups Count Sum Average Variance

Industrial 20 96.1 4.81 1.091

Retail 20 94.4 4.72 1.492

Financial 20 113.2 5.66 1.588

Mining 20 111 5.55 2.067

ANOVA

Source of Variation SS df MS F-stat P-value F crit

Between Groups 14.3894 3 4.7965 3.0757 0.0326 2.7249

Within Groups 118.5195 76 1.5595

Total 132.9089 79

Conclusion

Since F-stat = 3.076 > F-crit = 2.725, there is sufficient sample evidence at the

5% significance level to reject H0 in favour of H1. Therefore conclude that there is

at least one sector with a different mean earnings yield relative to the other sectors.

(c) Interpretation

By inspection, the Industrial and Retail sectors each appear to have significantly lower

mean earnings yields relative to the Financial and Mining sectors' mean earnings yields.

Exercise 11.13 File: X11.13 - advertising strategy.xlsx

Factor = Advertising strategy

(1 = Sophisticated; 2 = Athletic; 3 = Trendy)

Response measure = Sales (units of cans)

Let μi = population mean level of sales achieved under each advertising strategy i

H0: μ1 = μ2 = μ3

H1: At least one μi differs (i = 1, 2, 3)

SUMMARY

Groups Count Sum Average Variance

Sophisticated 20 8380 419 5075.579

Athletic 20 7241 362.05 6201.313

Trendy 20 8030 401.5 3781.737

ANOVA

Source of Variation SS df MS F-stat P-value F-crit

Between Groups 34039.03 2 17019.52 3.391 0.0406 3.159

Within Groups 286113.95 57 5019.54

Total 320152.98 59

Conclusion

Since F-stat = 3.391 > F-crit = 3.159, there is sufficient sample evidence at the 5% level of

significance to reject H0 in favour of H1. Therefore conclude that there is at least one

advertising strategy that results in a different mean level of deodorant sales relative

to the other strategies.

By inspection, the Athletic advertising strategy resulted in the lowest mean sales of all

three advertising strategies. On average, the Sophisticated and Trendy strategies

appear to be equally effective (the difference in sample means does not appear significant).

Recommendation:

Either the Trendy or the Sophisticated strategy can be adopted.

(d) Two-Sample Test of Means between the Sophisticated (1) and Trendy (2) Strategies

H0: μ1 = μ2 (two-tailed test)

H1: μ1 ≠ μ2

Sophisticated Trendy

Mean 419 401.5

Variance 5075.579 3781.737

Observations 20 20

Pooled Variance 4428.6579

Hypothesized Mean Difference 0

df 38

t Stat 0.8316

P(T<=t) one-tail 0.2054

t Critical one-tail 1.6860

P(T<=t) two-tail 0.4108 t-crit (0.05,38) =

t Critical two-tail 2.0244 =TINV(0.05,38) 2.0244

Conclusion

Since t-stat = 0.8316 lies within the region of non-rejection of H0 (i.e. ± 2.0244), there is

insufficient sample evidence to reject H0 in favour of H1.

Therefore the population mean sales from each of the two strategies is likely to be identical.

The two strategies are therefore equally effective and either can be adopted by the company.

(e) Recommendation

The Athletic strategy can be discarded. It appears the least effective.

The remaining two strategies (Sophisticated and Trendy) are equally effective and

therefore either can be adopted by the company to promote its new ladies deodorant.

Exercise 11.14 File: X11.14 - leverage ratio.xlsx

Technology Construction Banking Manufacturing

Standard Error 1.697 1.912 2.527 2.347

Median 72.5 78.5 68 79.5

Mode 60 80 68 81

Standard Deviation 9.296 10.471 13.841 12.853

Sample Variance 86.420 109.651 191.582 165.206

Kurtosis -1.309 -0.546 -0.453 0.495

Skewness 0.010 0.051 0.103 -0.857

Range 28 40 58 50

Minimum 60 58 41 44

Maximum 88 98 99 94

Sum 2215 2342 2092 2291

Count 30 30 30 30

Confidence Level(95.0%) 3.471 3.910 5.168 4.799

Interpretation

The mean leverage ratio is lowest for the banking sector and highest for the

construction and the manufacturing sectors.

These differences do appear to be significant.

(1 = Technology; 2 = Construction; 3 = Banking; 4 = Manufacturing)

Response measure = Leverage ratio

H0: μ1 = μ2 = μ3 = μ 4

H1: At least one μi differs (i = 1, 2, 3, 4)

SUMMARY

Groups Count Sum Average Variance

Technology 30 2215 73.83 86.42

Construction 30 2342 78.07 109.65

Banking 30 2092 69.73 191.58

Manufacturing 30 2291 76.37 165.21

ANOVA

Source of Variation SS df MS F-stat P-value F crit

Between Groups 1181.13 3 393.711 2.849 0.0406 2.683

Within Groups 16032.87 116 138.214

Conclusion

Since F-stat = 2.849 > F-crit = 2.683 (marginally), there is sufficient sample evidence at the

5% significance level to reject H0 in favour of H1. Therefore conclude that there is

at least one sector with a different mean leverage ratio relative to the other sectors.

By inspection, the banking sector has the lowest mean leverage ratio, while construction

and manufacturing appear to have similarly high mean leverage ratios.

Recommendation The investor is advised to consider either the banking sector (with the lowest

mean leverage ratio) or the technology sector (with a marginally higher mean

leverage ratio). (This difference may not be statistically significant).

(c) Two-Sample Test of Means - Leverage ratios between Technology (1) and Banking (2) sectors.

H0: μ1 = μ2 (two-tailed test)

H1: μ1 ≠ μ2

Technology Banking

Mean 73.833 69.733

Variance 86.420 191.582

Observations 30 30

Pooled Variance 139.001

Hypothesized Mean Difference 0

df 58

t Stat 1.347

P(T<=t) one-tail 0.092

t Critical one-tail 1.672

P(T<=t) two-tail 0.183 t-crit (0.05, 58) =

t Critical two-tail 2.002 =TINV(0.05,58) 2.002

Conclusion

Since t-stat = 1.347 lies within the region of non-rejection of H 0 (i.e. within ±2.002), the sample evidence is

not strong enough to reject H 0 in favour of H1. The population mean leverage ratios between the

Technology and the the Banking sector are therefore likely to be equal.

(d) Recommendation

Based on the statistical evidence in (c) and (d), an investor is advised to consider either

the Technology sector or the Banking sector. Their mean leverage ratios are likely to be equal.

Since both sectors offer an investor the same lower risk, either or both can be chosen for investment.

Exercise 11.15 File: X11.15 - training methods.xlsx

On-the-Job Lecture Role Play Audio-Visual

Mean 9 8.75 9.2 8.85

Standard Error 0.082 0.112 0.094 0.114

Median 9 8.65 9.2 8.9

Mode 9 8.5 9.6 8.9

Standard Deviation 0.327 0.420 0.325 0.426

Sample Variance 0.107 0.177 0.105 0.181

Kurtosis 0.109 0.339 -1.420 3.419

Skewness 0.472 0.759 0.229 -1.522

Range 1.2 1.5 0.900 1.7

Minimum 8.5 8.2 8.8 7.7

Maximum 9.7 9.7 9.7 9.4

Sum 144 122.5 110.4 123.9

Count 16 14 12 14

Confidence Level(95.0%) 0.174 0.243 0.206 0.246

Interpretation

The differences in mean performance scores appear marginal across the four

different training methods, with on-the-job training and role-play having the

highest average scores.

(1 = On-the-Job; 2 = Lecture; 3 = Role Play; 4 = Audio-Visual)

Response measure = Performance score (1 - 10)

H0: μ1 = μ2 = μ3 = μ4

H1: At least one μi differs (i = 1, 2, 3, 4)

SUMMARY

Groups Count Sum Average Variance

On-the-Job 16 144 9 0.107

Lecture 14 122.5 8.75 0.177

Role Play 12 110.4 9.2 0.105

Audio-Visual 14 123.9 8.85 0.181

ANOVA

Source of Variation SS df MS F-stat P-value F-crit

Between Groups 1.487 3 0.4957 3.479 0.0223 2.7826

Within Groups 7.41 52 0.1425

Total 8.897 55

Conclusion

Since F-stat = 3.479 > F-crit = 2.7826, there is sufficient sample evidence at the

5% significance level to reject H0 in favour of H1. Therefore conclude that there is at least one

training method with a different mean performance score relative to the other training methods.

By inspection, the lecture and audio-visual are the least effective (lower mean scores),

while on-the-job and the role-play methods are more effective (with higher mean scores).

Recommendation

The training manager is advised to consider either on-the-job training or use

role play methods (The difference does not appear to be statistically significant).

Performance scores between On-the-Job (1) and Role Play (2) methods.

H0: μ1 = μ2 (two-tailed test)

H1: μ1 ≠ μ2

Mean 9 9.2

Variance 0.10666667 0.105454545

Observations 16 12

Pooled Variance 0.10615385

Hypothesized Mean Difference 0

df 26

t Stat -1.6074361

P(T<=t) one-tail 0.06001825

t Critical one-tail 1.7056179

P(T<=t) two-tail 0.1200365 t-crit (0.05,26) =

t Critical two-tail 2.05552942 =TINV(0.05,26) 2.0555

Conclusion

Since t-stat = -1.6074 lies within the region of non-rejection of H 0, (i.e. within ±2.0555) the sample

evidence is not strong enough to reject H0 in favour of H1. The population mean performance scores

between the two training methods is likely to be the same.

(d) Recommendation

Based on the statistical evidence in (b) and (c), the HR manager is advised to select either

on-the-job training or role play methods of training.

Both are likely to produce similar high mean performance scores.

Exercise 11.16 In one-factor ANOVA, only one categorical factor is used to explain

possible differences between the observed sample means.

In two-factor ANOVA, two categorical factors are used to explain

possible differences between the observed sample means.

Factor 1 (x1) categorical (nominal / ordinal-scaled)

Factor 2 (x2) categorical (nominal / ordinal-scaled)

Exercise 11.18 Interaction term means that different combinations of the levels from the two factors

can have different influences (effects) on the numeric response variable.

Exercise 11.19 Interaction plot: It visually displays the nature and strength of the interaction

effect of combinations of factor levels on the dependent response variable.

It is constructed from the sample means of the various combinations

of the different factor levels.

Exercise 11.20

Source of Variation df SS MS F-stat p -value F -crit

Factor A: PC O.S. 2 68 34 6.8 0.0025 5.08

Factor B: Laptop 3 42 14 2.8 0.0499 4.22

Interaction (AxB) 6 132 22 4.4 0.0013 3.20

Error (Residual) 48 240 5

Total 59 482

(g) Factor A is statistically signficant (at α = 0.01) since F-stat (6.8) > F-crit (5.08).

Alternatively, its p-value (0.0025) < α (0.01).

Factor B is not statistically signficant (at α = 0.01) since F-stat (2.8) < F-crit (4.22).

Alternatively, its p-value (0.0499) > α (0.01).

The interaction effect between factor A and factor B is statistically signficant (at α = 0.01)

since F-stat (4.4) > F-crit (3.20).

Alternatively, its p-value (0.0013) < α (0.01).

XS11.21

H0: μ(b) = μ(a) = μ(s) H0: μ(under 3) = μ(over 3)

H1: At least one μi differs (I = b, a, s) H1: At least one μi differs (i = under 3, over 3)

Factor: Interaction

H0: No interaction effect

H1: There is an interaction effect

Under 3 years

Count 10 10 10 30

Sum 400.1 376.6 325.9 1102.6

Average 40.01 37.66 32.59 36.7533333

Variance 61.31656 23.07156 14.56766667 40.628092

Over 3 years

Count 10 10 10 30

Sum 446.2 357.6 415.8 1219.6

Average 44.62 35.76 41.58 40.6533333

Variance 28.37956 92.68711 42.13066667 64.626023

Total

Count 20 20 20

Sum 846.3 734.2 741.7

Average 42.315 36.71 37.085

Variance 48.08029 55.78305 48.12555263

ANOVA

Source of Variation SS df MS F-stat p-value F-crit

Sample (Experience) 228.15 1 228.150 5.222 0.0263 4.020

Columns (Qualifications) 392.73 2 196.365 4.494 0.0157 3.168

Interaction (Exper x Qualif) 300.26 2 150.131 3.436 0.0394 3.168

Within 2359.38 54 43.692

Total 3280.52 59

(a) (ii) Factor: Qualifications Since F-stat (4.494) > F-crit (3.168), reject H0 and conclude that qualifications

is a statistically significant factor (α = 0.05)

Business graduates generate significantly higher average sales (42.32)

than graduates with Arts (36.71) or Social Science (37.09) degrees.

(a) (iii) Factor: Experience Since F-stat (5.22) > F-crit (4.02), reject H0 and conclude that experience

is a statistically significant factor (α = 0.05)

Graduates with over 3 years experience generate significantly higher average sales (40.65)

than graduates with less than 3 years experience (36.75).

(a) (iv) Factor: Interaction effect Since F-stat (3.436) > F-crit (3.168), reject H0 and conclude that there is a significant

interaction effect between experience and qualifications on graduates' sales performance (α = 0.05)

Business graduates with over 3 years experience perform the best (44.62) while social science

graduates with with less than 3 years experience perform the worst (32.59).

Under 3 years 40.01 37.66 32.59

Over 3 years 44.62 35.76 41.58

Page 19

XS11.21

Chart Title

46 Interaction Plot

44

42

40

38

36

34

32

30

Business Arts Social Science

Recruit predominantly business and social science graduates with over 3 years of work experience.

If Arts graduates are employed, they must be given intensive marketing training.

Page 20

Exercise 11.22 File: X11.22 - dropped calls.xlsx

(a) (i) Random variable: % of daily dropped calls per network switch and transmission type

Factor 1 = Switch type (SW1, SW2, SW3, SW4)

Factor 2 = Transmission type (Voice, Data bundles)

Management Question: Is the % of daily dropped calls the same across all network switching

devices and / or transmission types

H0: μ(1) = μ(2) = μ(3) = μ(4) H0: μ(voice) = μ(data)

H1: At least one μi differs (i = 1,2,3,4) H1: At least one μi differs (i = voice, data)

Factor: Interaction

H0: No interaction effect

H1: There is an interaction effect

Voice

Count 8 8 8 8 32

Sum 7.76 4.35 4.69 6.39 23.19

Average 0.97 0.54375 0.58625 0.79875 0.724688

Variance 0.098657 0.081884 0.086227 0.071155 0.106645

Data

Count 8 8 8 8 32

Sum 5.83 6.82 8.16 7.1 27.91

Average 0.72875 0.8525 1.02 0.8875 0.872188

Variance 0.047984 0.078707 0.057171 0.067279 0.067818

Total

Count 16 16 16 16

Sum 13.59 11.17 12.85 13.49

Average 0.849375 0.698125 0.803125 0.843125

Variance 0.083953 0.100363 0.11709 0.066703

ANOVA

Source of Variation SS df MS F-stat P-value F-crit

Sample (Transmission) 0.3481 1 0.3481 4.727498 0.033926 4.012973

Columns (Switch) 0.234819 3 0.078273 1.063014 0.372175 2.769431

Interaction (Trans x Switch) 1.050075 3 0.350025 4.753641 0.005055 2.769431

Within 4.12345 56 0.073633

Total 5.756444 63

(a) (ii) Factor: Switching device Since F-stat (1.063) < F-crit (2.769), do not reject H0 and conclude that switch devices are not statistically

significant (α = 0.05). Hence all switch devices are likely to have the same average dropped call rate.

(a) (iii) Factor: Transmission Since F-stat (4.73) > F-crit (4.01), reject H0 and conclude that transmission type

is a statistically significant factor (α = 0.05)

Data transmission lead to a higher average dropped call rate (0.872) than voice transmission (0.725).

(a) (iv) Factor: Interaction effect Since F-stat (4.75) > F-crit (2.769), reject H0 and conclude that there is a significant

interaction effect between switch devices and transmission type. (α = 0.05)

SW2 and SW3, transmitting voice (0.544 and 0.586), are likely to have the lowest average dropped call rate,

while SW1 transmitting voice (0.97) and SW3 transmitting data (1.02), are likely to have the

highest average dropped call rate.

(b) Summary of Sample Means Table

Voice 0.970 0.544 0.586 0.799

Data 0.729 0.853 1.020 0.888

Chart Title

1.100

1.000 1.020

0.970

0.900 0.888

0.853

0.800 0.799

0.700 0.729

0.600 0.586

0.544

0.500

0.400

SW1 SW2 SW3 SW4

Voice Data

For voice transmissions, use switching devices 2 and 3;

For data transmission, use only switching device 1.

Investigate the high % dropped calls rates of switching device 3 (for data transmission)

and switching device 4 for both transmission types.

Exercise 11.23 File: X11.23 - rubber wastage.xlsx

H0: μ(1) = μ(2) = μ(3) H0: μ(R) = μ(B)

H1: At least one μi differs (i = 1, 2, 3) H1: At least one μi differs (i = Radial, Bias)

Factor: Interaction

H0: No interaction effect

H1: There is an interaction effect

Radial

Count 5 5 5 15

Sum 19.71 33.52 19.44 72.67

Average 3.942 6.704 3.888 4.844667

Variance 3.96097 6.56833 3.44237 5.844455

Bias

Count 5 5 5 15

Sum 13.47 38.6 43.77 95.84

Average 2.694 7.72 8.754 6.389333

Variance 1.35283 9.83965 8.96443 13.26548

Total

Count 10 10 10

Sum 33.18 72.12 63.21

Average 3.318 7.212 6.321

Variance 2.794329 7.579173 12.09134

ANOVA

Source of Variation SS df MS F-stat p-value F-crit

Sample (Tyres) 17.89496 1 17.89496 3.146037 0.088802 4.259677

Columns (TAMS) 83.25042 2 41.62521 7.317951 0.003301 3.402826

Interaction (Tyre x TAM) 47.77433 2 23.88716 4.1995 0.0273 3.402826

Within 136.5143 24 5.688097

Total 285.434 29

Factor: Tyres Since F-stat (3.15) < F-crit (4.26), do not reject H0 and conclude that tyre types is not

a statistically significant factor (α = 0.05)

Regardless of TAM used, average rubber wastage is the same across both tyre types produced.

Factor: TAMs Since F-stat (7.318) > F-crit (3.403), reject H0 and conclude that TAM used

is a statistically significant factor (α = 0.05)

Regardless of tyre type produced, average rubber wastage is likely to be lowest on TAM1 (3.318).

compared to TAM2 (7.212) and TAM3 (6.321).

Factor: Interaction effect Since F-stat (4.199) > F-crit (3.403), reject H0 and conclude that there is a significant

interaction effect between tyre type produced and TAM used (α = 0.05)

Average rubber wastage for bias tyres is lowest on TAM1 (2.694) but highest on TAM3 (8.754).

By contrast, average rubber wastage of radial tyres is lowest on TAM3 (3.888) and highest on TAM2 (6.704).

Radial 3.942 6.704 3.888

Bias 2.694 7.720 8.754

Chart Title

10.000

9.000 8.754

8.000 7.720

7.000 6.704

6.000

5.000

4.000 3.942 3.888

3.000 2.694

2.000

1.000

0.000

TAM1 TAM2 TAM3

Radial Bias

(c) Recommendation to the production manager:

TAM1 machine is the most efficient in minimising wastage for both tyre types (purchase more of TAM1).

Allocate Radial tyres manufacturing to TAM3 (it has the minimum % wastage of all three machines on Radial tyres).

Investigate the high % wastage of TAM3 for Bias tyres and TAM2 for both tyre types.

Possibly replace TAM2 type machines with TAM1 type machines.

CHAPTER 12

Exercise 12.1 Regression analysis defines the structural relationship between two numeric

variables as a mathematical equation (usually a straight line equation).

Its purpose is to use the equation for estimation / prediction purposes.

two numeric variables used in the regression equation.

Degrees of freedom = 18 - 2 = 16

t-crit = t(0.05, 16) = 2.12 (Use =TINV(0.05,16) in Excel )

Decision Rule: Accept H0 if -2.12 ≤ t-stat ≤ +2.12

Conclusion: Do not reject H0. There is no statistically

significant relationship between x and y .

Exercise 12.8 File: X12.8 - training effectiveness.xlsx

80

output (units) 70

60

50

40

30

20

10 15 20 25 30 35 40 45

hours of training

Interpretation

There is a strong positive relationship between hours of training

and worker output. The more the training received, the higher the output.

20 40 400 800 1600

36 70 1296 2520 4900

20 44 400 880 1936

38 56 1444 2128 3136

40 60 1600 2400 3600

33 48 1089 1584 2304

32 62 1024 1984 3844

28 54 784 1512 2916

40 63 1600 2520 3969

n = 10 24 38 576 912 1444

Σ 311 535 10213 17240 29649

Coefficients b1 = (10*17240-(311*535))/(10*10213-(311²))

1.112

b0 = (535 - 1.112*311)/10

18.917

(c) Correlation coefficient (r) = (17240-(311*535)/10)/√((10213-(311²)/10)*(29649-(535²)/10))

= 0.8072

Coefficient of Determination (r²) = 0.8072² = 0.6516

Variation in training hours can explain 65.16% of the variability in worker output.

This is a high level of explained variation. Hence training input is very beneficial

to worker output and the training programmes should be continued.

A worker with 25 hours of training can be expected to produce 46.72 units of output, on average.

Exercise 12.9 File: X12.9 - capital utilisation.xlsx

Scatter Plot

Earnings Yield vs Inventory Turnover

18

16

14

earnings yield

12

10

8

6

4

2

0

0 1 2 3 4 5 6 7 8 9

inventory turnover

Interpretation

There is a strong positive relationship between inventory turnover and earnings yield.

As inventory turnover increases, earnings yields also increases.

3 10 9 30 100

5 12 25 60 144

4 8 16 32 64

7 13 49 91 169

6 15 36 90 225

4 10 16 40 100

8 16 64 128 256

6 13 36 78 169

n =9 5 10 25 50 100

Σ 48 107 276 599 1327

Coefficients b1 = (9*599-(48*107))/(9*276-(48²))

1.4167

b0 = (107-1.4167*48)/9

4.333

= 0.8552

There is a strong positive linear association between inventory turnover and earnings yield.

Thus the business analyst's view is supported by the strong sample evidence.

(d) Coefficient of Determination (r²) = 0.8552² = 0.7313 i.e. 73.13%

Inventory turnover (capital utilisation) can explain 73.12% of the variability in a company's

earnings yield. This is a high level of explained variation. Hence inventory turnover has been

shown to have a significant direct effect on a company's earnings yield.

Yes, the regression equation can by used with confidence to estimate earnings yield

based on a company's level of inventory turnover.

H1: ρ ≠ 0

t-crit = t(α=0.05,df = 7) = ± 2.364 2.364624

-2.364 ≤ t-stat ≤ +2.364

Conclusion

Since t-stat (4.3655) lies outside the area of acceptance for H 0,there is

sufficient sample evidence at the 5% significance level to reject H0 in favour of H1.

Conclude that there is a strong positive association between inventory turnover

and earnings yield.

A company with an expected inventory turnover of 6 next year can expect to achieve

an earnings yield of 12.833%.

Exercise 12.10 File: X12.10 - loan applications.xlsx

Independent variable (x) = interest rate (%)

Interest rate (%) is assumed to influence the number of loan applications received.

Scatter Plot

Loan applications Received vs Interest Rate (%)

35

no. loan applications received

30

25

20

15

10

4 5 6 7 8 9 10

Interpretation

A moderate to strong negative linear relationship between interest rate and

number of loan applications received is observed from the scatter plot.

7 18 49 324 126

6.5 22 42.25 484 143

5.5 30 30.25 900 165

6 24 36 576 144

8 16 64 256 128

8.5 18 72.25 324 153

6 28 36 784 168

6.5 27 42.25 729 175.5

7.5 20 56.25 400 150

8 17 64 289 136

n = 11 6 21 36 441 126

Σ 75.5 241 528.25 5507 1614.5

r= (11*1614.5 - 75.5*241)/√((11*528.25-75.5²)*(11*5507-241²))

= -0.8302

Interpretation

This correlation shows a strong negative (inverse) assocation between

interest rates (%) and the number of loan applications received.

(d) H0: ρ = 0 Area of Acceptance for H0

H1: ρ ≠ 0

t-crit = t(α=0.05,df = 9) = ± 2.262

-2.262 ≤ t-stat ≤ +2.262

t-stat = (-0.8302)*√((11-2)/(1-(-0.8302²))) =

'= -4.4677

Conclusion

Since t-stat (-4,4677) lies outside the area of acceptance for H0,there is

sufficient sample evidence at the 5% significance level to reject H0 in favour of H1.

Conclude that there is a strong negative relationship between interest rates (%)

and the number of loan applications received monthly.

b1 = (11*1614.5 - 75.5*241)/(11*528.25-75.5²)

-3.9457

b0 = (241 - (-3.9457)*(75.5))/11

48.991

For a 1% increase in interest rate, 3.95 fewer loan applications will be received.

If the rate of interest is 6%, the bank can expect to receive 25.32 (say, 25) applications.

Exercise 12.11 File: X12.11 - maintenance costs.xlsx

Independent variable (x) = Machine age (years)

Machine age is assumed to influence annual maintenance costs.

Scatter Plot

Annual Maintenance Cost (R) vs Age of Machines

70

annual maintenance cost (R)

60

50

40

30

20

10

0

0 2 4 6 8 10

Interpretation

A strong direct (positive) linear relationship between the ages of machines and

their annual maintenance costs (in Rands) is observed in the scatter plot.

4 45 16 2025 180

3 20 9 400 60

3 38 9 1444 114

8 65 64 4225 520

6 58 36 3364 348

7 50 49 2500 350

1 16 1 256 16

1 22 1 484 22

5 38 25 1444 190

2 26 4 676 52

4 30 16 900 120

n = 12 6 35 36 1225 210

Σ 50 443 266 18943 2182

r= (12*2182 - 50*443)/√((12*266-50²)*(12*18943-443²))

= 0.870028

Interpretation

There is a very strong positive association between the ages of

of machines and their annual maintenance costs (in Rands).

(d) H0: ρ = 0 Area of Acceptance for H0

H1: ρ ≠ 0

t crit = t(α=0.05,df = 10) = ± 2.228

-2.228 ≤ t-stat ≤ +2.228

Conclusion

Since t-stat (5.5806) lies outside the area of acceptance for H 0,there is

sufficient sample evidence at the 5% significance level to reject H 0 in favour of H1.

Conclude that there is a strong direct association between age of machines

and the level of annual maintance costs (in Rands).

5.8295

b0 = (443 - 5.8295*50)/12

12.627

If the age of a machine increased by 1 year, annual maintenance costs will rise by R5.8295.

(Alternatively, for every year older, the annual maintenance costs increase by R5.8295)

For a 5 year old machine, annual maintenance costs are expected to be R41.77.

Exercise 12.12 File: X12.12 - employee performance.xlsx

Scatter Plot

Performance Rating vs Aptitude Score

90

85

80

75

70

65

60

55

50

2 3 4 5 6 7 8 9 10

Interpretation

There appears to be a weak to moderate positive association between an employee's

aptitude score and their performance rating.

7 82 49 6724 574

6 74 36 5476 444

5 82 25 6724 410

4 68 16 4624 272

5 75 25 5625 375

8 92 64 8464 736

7 86 49 7396 602

8 69 64 4761 552

9 85 81 7225 765

6 76 36 5776 456

4 72 16 5184 288

n = 12 6 64 36 4096 384

Σ 75 925 497 72075 5858

r= (12*5858 - 75*925)/√((12*497-75²)*(12*72075-925²))

0.5194

Interpretation

There is a moderate positive association between the aptitude scores of

employees and their performance scores after one year.

(c) H0: ρ = 0 Area of Acceptance for H0

H1: ρ ≠ 0

t-crit = t(α=0.05,df = 10) = ± 2.228

-2.228 ≤ t-stat ≤ +2.228

Conclusion

Since t-stat (1.922) lies inside the area of acceptance for H0,there is

not sufficient sample evidence at the 5% significance level to reject H 0 in favour of H1.

Conclude, at a 5% significance level, that there is no statistically significant association

between an employee's aptitude score and their job performance rating score one year later.

b1 = (12*5858-75*925)/(12*497-75²) 2.7168

Interpretation of b1 coefficient

If an employee's aptitude score increases by one unit, their performance rating

score will increase by 2,7168 points.

(e) Estimation

x =8 ŷ= 60.1032 + 2.7168*(8) = 81.84

For an employee with an aptitude score of 8, they could expect a job performance rating

score of 81.84.

The association between aptitude score and performance rating is not statistically significant.

Therefore, the call centre manager should have low confidence in this

estimated performance rating score.

Exercise 12.13 File: X12.13 - opinion polls.xlsx

42 51 1764 2601 2142

34 31 1156 961 1054

59 56 3481 3136 3304

41 49 1681 2401 2009

53 68 2809 4624 3604

40 35 1600 1225 1400

65 54 4225 2916 3510

48 52 2304 2704 2496

59 54 3481 2916 3186

38 43 1444 1849 1634

n = 11 62 60 3844 3600 3720

Σ 541 553 27789 28933 28059

r= (11*28059 - 541*553)/√((11*27789-541²)*(11*2893-553²))

0.7448

Interpretation

There is a moderate to strong positive association between opinion poll predictions

and the actual election results.

H1: ρ ≠ 0

t-crit = t(α=0.05,df = 9) = ± 2.262

-2.262 ≤ t-stat ≤ +2.262

Conclusion

Since t-stat (3.3484) lies outside the area of acceptance for H0,there is

sufficient sample evidence at the 5% significance level to reject H0 in favour of H1.

Conclude, at a 5% significance level, that there is a statistically significant association

between opinion poll predictions and the actual election results.

Interpretation of b1 coefficient

For a one percentage point increase in an opinion poll prediction, the actual

election percentage is likely to increase by 0.729 percentage points.

r² = 0.7448² = 0.5547

Interpretation

Opinion poll predictions can explain 55.47% of variation in actual election results percentages

(e) Prediction

x = 58 ŷ= 14.4175 + 0.729*(58) = 56.70%

If an opinion poll predicts support at 58%, the actual election result is likely to be 56.7%.

If an opinion poll predicts support at 82%, the actual election result is likely to be 74.2%.

Because x = 82 is beyond the domain of x (34 ≤ x ≤ 65), this expected actual election result

is unreliable and possibly invalid due to extrapolation being used.

Exercise 12.14 File: X12.14 - capital investment.xlsx

Independent variable (x) = capital investment

Scatter Plot

Return on Investment (%) vs Capital Investment (%)

10

return on investment (%)

0

10 20 30 40 50 60 70 80 90

levels of capital investm ent (%)

Interpretation

There appears to be a weak to moderate association between a company's level of capital

investment and its return on investment.

SUMMARY OUTPUT

Regression Statistics

Multiple R 0.41447

R Square 0.17179

Adjusted R Square 0.15253

Standard Error 2.12495

Observations 45

ANOVA

df SS MS F-stat p-value

Regression 1 40.27335 40.27335 8.91907 0.00465

Residual 43 194.16309 4.51542

Total 44 234.43644

Intercept 1.27412 1.19293 1.06806 0.29145 -1.13166 3.67990

Capital 0.06783 0.02271 2.98648 0.00465 0.02203 0.11363

(d) Coefficient of Determination r²

r² = 0.41447² = 0.171788

Interpretation

The variation in the level of capital investment explains only 17.18% of the variation

in return on investment .

H1 : ρ ≠ 0

t-crit = t(α=0.05,df = 43) = ± 2.016

-2.016 ≤ t-stat ≤ +2.016

Conclusion

Since t-stat (2.9865) lies outside the area of acceptance for H0,there is

sufficient sample evidence at the 5% significance level to reject H0 in favour of H1.

Conclude, at a 5% significance level, that there is a statistically significant association

between a company's level of capital investment and its return on investment.

For a one percentage point change in capital investment , company return on investment

can be expected to change by 0.06783 percentage points.

(h) Estimation

x = 55 ŷ = 1.27412 + 0.06783 (55) = 5.005%

The expected return on investment for a company with a 55% level of capital

investment is 5.005%.

Exercise 12.15 File: X12.15 - property valuations.xlsx

Independent variable (x ) = Council Valuations

300

250

market values (R)

200

150

100

50 Market values

0

0 50 100 150 200

council valuations (R)

Interpretation

There appears to be a strong positive correlation between the council's valuation of

a residential property in Bloemfontein and its market value.

SUMMARY OUTPUT

Regression Statistics

Multiple R 0.78104

R Square 0.61002

Adjusted R Square 0.59975

Standard Error 26.40553

Observations 40

ANOVA

df SS MS F-stat p-value

Regression 1 41444.82 41444.82 59.4402 2.7522E-09

Residual 38 26495.58 697.2521

Total 39 67940.4

Intercept 71.3619 14.2370 5.0124 1.28E-05 42.541 100.183

Council valuations 1.0151 0.1317 7.7097 2.75E-09 0.749 1.282

Regression line (equation) ŷ = 71.3619 + 1.0151 x for 48 ≤ x ≤ 154

r² = 0.78104² = 0.61002

Interpretation

Variation in council valuations explain 61.002% of the variation in property market values .

(e) H0: ρ = 0 Area of Acceptance for H0

H1: ρ ≠ 0

t-crit = t(α=0.05,df = 38) = ± 2.024

-2.024 ≤ t-stat ≤ +2.024

Conclusion

Since t-stat (7.70975) lies well outside the area of acceptance for H0,there is

sufficient sample evidence at the 5% significance level to reject H 0 in favour of H1.

Conclude, at a 5% significance level, that there is a statistically significant association

between the Council's valuation and the resultant property market valuation.

For a R1 (in R1000) change (up / down) in council valuation , market valuation of a

property can be expected to change (up / down) by R1.0151 (in R1000) .

(g) Estimation

x = 100 ŷ = 71.3619 + 1.0151 (100) = R172.874 (in R1000s)

The expected market value of a property which the council values at R100 (in R1000s)

is likely to be R172.874 (in R1000s).

CHAPTER 13

Exercise 13.1 Simple linear regression has only one independent variable (x1) whereas multiple linear

regression has two or more independent variables (x1, x2, x3, … , xk) that are assumed

to influence the outcome of the dependent variable, y .

df SS MS F-stat p-value F-crit

Regression 5 84 16.8 4.541 0.002261 2.45

Residual 40 148 3.7

Total 45 232

36.2% of total variation in y can be explained by the 5 independent variables.

or H0: ρ = 0 vs H1: ρ ≠ 0

(c) F-stat = 4.5412 and F-crit = F(0.05,5,40) = 2.45 (See Anova Table)

(d) Reject H0. Conclude that the overall model is statistically significant.

(i.e. at least one x i is statistically significant in estimating y)

Exercise 13.3

Intercept 1.82 1.12 1.63 0.1215 -0.53 3.92

A 0.68 0.28 2.44 0.0253 0.09 2.78

B -2.35 0.984 -2.39 0.0140 -4.42 -0.25

C 0.017 0.012 1.42 0.1737 -0.01 2.12

D 1.96 1.16 1.69 0.1083 -0.48 4.06

(a) For each x i variable (A, B, C and D), test: H0: βi = 0 against H1: βi ≠ 0 for i = A, B, C and D.

For A : Since t-stat (2.44) > t-crit (+2.093); or p -value (0.0253) < α (0.05), or

{0.09 ≤ βA ≤ 2.78} does not cover zero, conclude variable A is statistically significant.

For B : Since t-stat (-2.39) < -t-crit (-2.093); or p -value (0.014) < α (0.05), or

{-4.42 ≤ βB ≤ -0.25} does not cover zero, conclude variable B is statistically significant.

For C : Since t-stat (1.42) < -t-crit (+2.093); or p -value (0.1737) > α (0.05), or

{-0.01 ≤ βC ≤ 2.12} covers zero, conclude variable C is not statistically significant.

For D : Since t-stat (1.69) < -t-crit (+2.093); or p -value (0.1083) > α (0.05), or

{-0.48 ≤ βD ≤ 4.06} covers zero, conclude variable D is not statistically significant.

Exercise 13.4

(b) t-crit = t(0.05,30) = ±2.042. Hence, do not reject H0 if -2.042 ≤ t-stat ≤ +2.042.

(c ) Since t-stat (2.44) > t-crit (+2.042), hence reject H0 in favour of H1 at α = 0.05.

Exercise 13.5

(a) Holding all other variables constant, a unit increase in x 2 will result in a 1.6 reduction in y^ .

Yes, since the 95% confidence interval for β2 does not cover zero.

Exercise 13.6 Binary coding scheme (Choose 'Lean' as the base category)'

F1 and F2 are the dummy variable names chosen.

Fuel type F1 F2

Leaded 1 0

Unleaded 0 1

Lean 0 0

Exercise 13.7 Binary coding scheme (Choose 'spring' as the base category)'

S1, S2 and S3 are the dummy variable names chosen.

Season S1 S2 S3

summer 1 0 0

autumn 0 1 0

winter 0 0 1

spring 0 0 0

Exercise 13.8 File: X13.8 - employee absenteeism.xlsx

SUMMARY OUTPUT

Regression Statistics

Multiple R 0.7384

R Square 0.5453

Adjusted R Square 0.5012

Standard Error 3.7204

Observations 35

ANOVA

df SS MS F-stat p-value

Regression 3 514.467 171.489 12.390 0.00001707

Residual 31 429.076 13.841

Total 34 943.543

Intercept 36.411 5.929 6.141 8.22345E-07 24.318 48.504

Tenure 0.220 0.065 3.376 0.001997 0.087 0.352

Satisfaction -0.184 0.109 -1.692 0.100765 -0.406 0.038

Commitment -0.332 0.080 -4.130 0.000254 -0.497 -0.168

F-crit = F(0.05,3,31) = 2.92 F-stat = 12.39 and p -value = 0.00001707

Reject H0. Conclude the overall model is statistically significant.

(i.e. at least one x i is statistically significant in estimating y)

(c), (d), (e) For each x i variable, test: H0: βi = 0 against H1: βi ≠ 0 with t-crit = t(0.05,31) = ±2.04

Tenure : Since t-stat (3.376) > t-crit (+2.04); or p -value (0.001997) < α (0.05), or

{0.087 ≤ βT ≤ 0.352} does not cover zero, conclude Tenure is statistically significant.

Satisfaction : Since t-stat (-1.692) lies within t-crit (±2.04); or p -value (0.100765) > α (0.05), or

{-0.406 ≤ βS ≤ 0.038} covers zero, conclude Satisfaction is not statistically significant.

Commitment : Since t-stat (-4.13) < t-crit (-2.04); or p -value (0.000254) < α (0.05), or

{-0.497 ≤ βC ≤ -0.168} does not cover zero, conclude Commitment is statistically significant.

(f) (i) No, organisational commitment is the most important explanatory factor because it has a

larger t-stat value (-4.13) and a smaller p- value (0.000254) than job tenure .

(f) (ii) No, job satisfaction is not a statistically signficant explanatory factor of employee

absenteeism (see (c), (d) and (e) above).

Yes, organisational commitment does play a statistically signficant role in explaining

employee absenteeism (see (c), (d) and (e) above).

(g) y(hat) = 36.411 + 0.22 (48) - 0.184 (50) - 0.332 (60) = 17.8008

Using t-crit = t(0.025,31) = ±2.04; standard error = 3.7204; n = 35

giving margin of error = 2.04 (3.7204)/√35 = 1.2826

Lower 95% confidence limit = 17.8008 - 1.2826 16.5182

Upper 95% confidence limit = 17.8008 + 1.2827 19.0835

{16.52 ≤ y(hat)(estimated) ≤ 19.08}

Management interpretation: We can be 95% confident that the true average number

of days absent per employee per annum is likely to lie between 16.5 days and 19.1 days.

Exercise 13.9 File: X13.9 - plastics wastage.xlsx

SUMMARY OUTPUT

Regression Statistics

Multiple R 0.8061

R Square 0.6498

Adjusted R Square 0.6109

Standard Error 0.5160

Observations 31

ANOVA

df SS MS F-stat p-value

Regression 3 13.3384 4.4461 16.6992 2.466E-06

Residual 27 7.1887 0.2662

Total 30 20.5271

Intercept 1.8179 1.2189 1.4914 0.1474 -0.6830 4.3188

Dexterity -0.1112 0.0286 -3.8816 0.0006 -0.1700 -0.0524

Speed 0.0173 0.0047 3.6770 0.0010 0.0077 0.0270

Viscosity 1.9189 1.2581 1.5252 0.1388 -0.6625 4.5004

F-crit = F(0.05,3,27) = 2.99

F-stat = 16.6992 and p-value = 0.000002466

Reject H0. Conclude the overall model is statistically significant.

(i.e. at least one x i is statistically significant in estimating y)

(c), (d), (e) For each x i variable, test: H0: βi = 0 against H1: βi ≠ 0 with t-crit = t(0.05,27) = ±2.052

Dexterity : Since t-stat (-3.886) < t-crit (-2.052); or p -value (0.0006) < α (0.05), or

{-0.17 ≤ βD ≤ -0.0524} does not cover zero, conclude Dexterity is statistically

significant.

Speed : Since t-stat (3.677) > t-crit (+2.052); or p -value (0.001) < α (0.05), or

{0.0077 ≤ βS ≤ 0.027} does not cover zero, conclude Speed is statistically significant.

Viscosity : Since t-stat (1.525) < t-crit (+2.052); or p -value (0.1388) > α (0.05), or

{-0.6625 ≤ βV ≤ 4.5} covers zero, conclude Viscosity is not statistically significant.

(f) The most important factor is operator dexterity (p -value (0.0006)), then machine speed with

p -value = 0.001. Plastic viscosity is not a significant influencing factor (p -value = 0.1388).

(g) y(hat) = 1.8179 - 0.1112 (25) + 0.0173 (200) + 1.9189 (0.25) = 2.9826

Using t-crit = t(0.025,27) = ±2.052; standard error = 0.516; n = 31

giving margin of error = 2.052 (0.516)/√31 = 0.1902

Lower 95% confidence limit = 2.9826 - 0.1902 = 2.7924

Upper 95% confidence limit = 2.9826 + 0.1902 = 3.1728

{2.79% ≤ y(hat)(estimated) ≤ 3.17%}

Management interpretation: We can be 95% confident that the true average % of plastic

wastage per shift is likely to lie between 2.79% and 3.17%.

Exercise 13.10 File: X13.10 - employee performance.xlsx

(a) Binary coding scheme (Choose 'method C' as the base category)

MA and MB are the dummy variable names chosen.

Method Code MA MB

A 1 0

B 0 1

C 0 0

(b) Sample input data for first 6 consultants (showing the binary coded data)

1 24 9 1 0

2 30 4 0 1

3 26 10 1 0

4 37 12 0 1

5 29 10 0 0

6 28 6 0 0

Excel's Data Analysis - Regression (using the binary coded data as in (a))

SUMMARY OUTPUT

Regression Statistics

Multiple R 0.7894

R Square 0.6231

Adjusted R Square 0.5524

Standard Error 2.3425

Observations 20

ANOVA

df SS MS F-stat p-value

Regression 3 145.151 48.384 8.817 0.001108

Residual 16 87.799 5.487

Total 19 232.95

Intercept 26.387 1.940 13.601 3.29E-10 22.274 30.499

Experience 0.389 0.167 2.335 0.0329 0.036 0.742

MA -3.659 1.376 -2.659 0.0172 -6.576 -0.742

MB 1.472 1.336 1.102 0.2867 -1.360 4.304

y^ = 26.387 + 0.389 Experience - 3.659 MA + 1.472 MB (based on data recoded as in (a))

F-crit = F(0.05,3,16) = 3.24

F-stat = 8.817 and p -value = 0.001108

Reject H0. Conclude the overall model is statistically significant.

(i.e. at least one x i is statistically significant in estimating y)

(e) For Experience , test: H0: βE = 0 against H1: βE ≠ 0 with t-crit = t(0.05,16) = ±2.12

Since t-stat (2.335) > t-crit (2.12), conclude work experience is statistically significant at α = 0.05.

(f) For each of MA and MB, test: H0: βi = 0 vs H1: βi ≠ 0 (i = MA, MB), wiith t-crit = t(0.05,16) = ±2.12

MA : Since t-stat (-2.659) < lower t-crit (-2.12), conclude MA is statistically significant (i.e. adopting

marketing method A results in significantly lower consultant productivity levels, on average,

compared to using marketing method C (i.e. the base category).

MB : Since –t-crit (-2.12) < t-stat (1.102) < +t-crit (-2.12), conclude MB is not statistically significant

(i.e. consultant productivity levels, on average, are the same for marketing method B and

and marketing method C (i.e. no difference to the base category average productivity level).

Overall conclusion: the independent variable ‘marketing method’ is statistically significant, but

only for marketing method A (when compared to method C (i.e. the base category).

Marketing methods B and C can be combined as there is no statistically significant difference

between them with regards to their average productivity levels across consultant.

Experience : Since its p -value (0.0329) < α = 0.05, or {0.036 ≤ βE ≤ 0.742} does not cover zero,

conclude that work experience is statistically significant.

MA : Since its p -value (0.0172) < α = 0.05, or {-6.576 ≤ βMA ≤ -0.742} does not cover zero,

conclude that marketing method A is statistically significantly different from marketing method C

(i.e. the base category) in terms of consultants' average productivity levels.

MB : Since its p -value (0.2867) > α = 0.05, or {-1.36 ≤ βMA ≤ 4.304} covers zero,

conclude marketing method B is not statistically significantly different from marketing method C

(i.e. the base category) in terms of consultants' average productivity levels.

(i) Employ consultants with longer work experience and avoid using marketing method A as it

produces lower productivity levels than either marketing methods B or C.

Using t-crit = t(0.025,16) = ±2.1199; standard error = 2.3425; n = 20

giving margin of error = 2.1199 (2.3425)/√20 = 1.1104

Lower 95% confidence limit = 30.9717 - 1.1104 = 29.86

Upper 95% confidence limit = 30.9717 + 1.1104 = 32.08

{29.86 deals ≤ y(hat)(estimated) ≤ 32.08 deals}

Management interpretation: The bank management can be 95% confident that the actual

average number of deals closed per month per consultant is likely to lie between 30 and 32 (rounded).

Exercise 13.11 File: X13.11 - corporate performance.xlsx

(a) Binary coding scheme (Choose region 'KZN' as the base category)

R1 and R2 are the dummy variable names chosen.

Region Code R1 R2

Gauteng 1 1 0

Cape 2 0 1

KZN 3 0 0

S is the dummy variable name chosen.

Sector Code S

Agriculture 1 1

Construction 2 0

(b) Sample input data for first 6 companies (showing the binary coded data)

19.7 7178 18.7 28.5 1 0 0

17.2 1437 18.5 24.3 1 0 1

17.1 3948 16.5 65.6 1 0 1

16.6 1672 16.2 26.4 1 0 1

16.6 2317 16.0 20.1 1 0 1

16.5 4123 15.6 46.4 0 1 0

Excel's Data Analysis - Regression (using the binary coded data as in (a))

Regression Statistics

Multiple R 0.9125

R Square 0.8327

Adjusted R Square 0.7769

Standard Error 0.9524

Observations 25

ANOVA

df SS MS F-stat p-value

Regression 6 81.2595 13.5433 14.9318 0.000004076

Residual 18 16.3261 0.9070

Total 24 97.5856

REGRESSION OUTPUT

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

Intercept 11.0146 0.8746 12.5936 0.0000 9.1771 12.8520

Sales 0.0002 0.0001 2.1668 0.0439 0.00001 0.00033

Margin% 0.1791 0.0672 2.6656 0.0158 0.03794 0.32030

Debt ratio(%) 0.0091 0.0154 0.5930 0.5606 -0.02321 0.04146

R1 3.1453 0.8389 3.7494 0.0015 1.38288 4.90770

R2 0.9213 0.5862 1.5716 0.1335 -0.31031 2.15295

S -0.9230 0.4205 -2.1950 0.0415 -1.80651 -0.03956

y^ = 11.0146 + 0.0002 Sales + 0.1791 Margin % + 0.0091 Debt ratio % + 3.1453 R1 + 0.9213 R2 - 0.923 S

(based on data recoded as in (a))

2

(c ) R = 81.2595/97.5856 = 83.27%

(d) H0: βS = βM% = βDR = βR1 = βR2 = βS = 0 versus H1: At least one βi ≠ 0

F-crit = F(0.05,6,18) = 2.66

F-stat = 14.932 and p -value = 0.000004076

Reject H0. Conclude that the overall model is statistically significant.

(i.e. at least one x i is statistically significant in estimating y)

(e), (h), (i) For all variables, test: H0: βi = 0 vs H1: βi ≠ 0 with t-crit = t(0.05,18) = ±2.101

Sales : Since t-stat (2.1668) > t-crit (2.101), or p -value (0.0439) < α = 0.05 or

{0.00001 ≤ βS ≤ 0.00033} does not cover zero, conclude sales is statistically significant.

Margin% : Since t-stat (2.6656) > t-crit (2.101), or p -value (0.0158) < α = 0.05 or

{0.0379 ≤ βM% ≤ 0.3203} does not cover zero, conclude margin% is statistically significant.

Debt ratio% : Since –t-crit (-2.101) < t-stat (0.593) < t-crit (2.101), or p -value (0.5606) > α = 0.05 or

{-0.0232 ≤ βDR% ≤ 0.0415} covers zero, conclude debt ratio% is not statistically significant.

(f) Region: For each dummy variable R1 and R2 , test: H0: βi = 0 against H1: βi ≠ 0 with t-crit = t(0.05,18) = ±2.101

R1 (Gauteng) : Since t-stat (3.7494) > upper t-crit (2.101); or its p -value (0.0015) < α = 0.05, or

{1.38288 ≤ βR1 ≤ 4.9077} does not cover zero, conclude that companies that operate in the Gauteng region

have a statistically significantly higher return on capital (%), on average, than companies that operate

in the KZN region (i.e. the base region).

R2 (Cape) : Since t-stat (1.5716) < upper t-crit (2.101); or its p -value (0.1335) > α = 0.05, or

{-0.31031 ≤ βR2 ≤ 2.15295} covers zero, conclude that companies that operate in the Cape region

do not have a statistically significant difference in their average return on capital (%)than companies

that operate in the KZN region (i.e. the base region).

Overall conclusion: the independent variable ‘region’ is statistically significant, but only with

respect to the Gauteng region (when compared to th KZN region (i.e. the base region).

The Cape and KZN regions can be merged into a single region as there is no statistically significant

difference in the average return on capital (%) of companies operating within these two regions.

(g) Sector: For the dummy variable, S, test: H0: βS = 0 against H1: βS ≠ 0 with t-crit = t(0.05,18) = ±2.101

S (Agriculture) : Since t-stat (-2.195) < lower t-crit (-2.101); or its p -value (0.0415) < α = 0.05, or

{-1.80651 ≤ βS ≤ -0.03956} does not cover zero, conclude that companies that operate in the

agricultural sector have a statistically significantly lower return on capital (%), on average, than

companies that operate in the construction sector (i.e. the base sector).

Overall conclusion: the independent variable ‘sector’ is statistically significant.

(j) Significant performance measures of ROC% are: Sales, Margin%, but not Debt ratio%.

For region, Gauteng has a significantly positive impact on average ROC% compared to Cape and KZN.

The agricultural sector has a significantly negative impact on average ROC% compared to the construction sector.

(j) y^ = 11.0146 + 0.0002 (8862) + 0.1791 (10) + 0.0091 (22) + 3.1453 (0) + 0.9213 (1) - 0.923 (0) = 15.6995

Using t-crit = t(0.025,18) = ±2.101; standard error = 0.9524; n = 25

giving margin of error = 2.101 (0.9524)/√25 = 0.400198

Lower 95% confidence limit = 15.6995 - 0.400198 = 15.299

Upper 95% confidence limit = 15.6995 + 0.400198 = 16.100

{15.299% ≤ y(hat)(estimated) ≤ 16.10%}

Management interpretation: The investment analyst can be 95% confident that the actual

average return on capital (%) of companies with the given profile lies between 15.3% and 16.1%.

CHAPTER 14

INDEX NUMBERS

Exercise 14.1 An index number is a single summary value that measures the overall

change in the level of activity of a single item or a basket of related

items from one time period to the another.

Example 14.2 A price index measures changes in price levels over time, holding

quantities constant.

A quantity index measures changes in consumption levels over time, holding

prices constant.

Example 14.3 Items are 'weighted' in a basket to reflect the importance (or value) of each

item in the basket relative to the other items in the basket.

(ii) The selection of the basket of items (i.e. the mix of items)

(iii) The choice of item weights

(iv) The choice of a suitable base year

(v) The formulation of a substitution rule

A price relative is a change in the level of activity of an item in a given period

relative to a base period.

Example 14.7 Real (constant) values are found by dividing monetary values by an 'inflation' index.

This removes the influence of price increases.

Real (constant) values refer to the actual purchasing power of money / or the

real (actual) change in the level of activity.

Exercise 14.8 File: X14.8 - motorcycle sales.xlsx

Motorcycle Unit price Quantity Unit price Quantity

model (R1000) (units sold) (R1000) (units sold)

A 25 10 30 7

B 15 55 19 58

C 12 32 14 40

Motorcycle Price

(a) p1/p0*100

model Relative

A 30/25*100 = 120.0

B 19/15*100 = 126.7

C 14/12*100 = 116.7

Interpretation

Model A has risen in price by 20% from 2009 to 2010; model B by 26.7%

and model C by 16.7%.

Motorcycle

Base Value (p0*q0) Current Value (p1*q0)

model

A 25*10 = 250 30*10 = 300

B 15*55 = 825 19*55 = 1045

C 12*32 = 384 14*32 = 448

Totals 1459 1793

Weighted

Motorcycle

Base Value (p0*q0) Price Relative (p1/p0) Price

model

Relatives

A 25*10 = 250 30/25*100 = 120.0 120*250 +

B 15*55 = 825 19/15*100 = 126.7 126.7*825 +

C 12*32 = 384 14/12*100 = 116.7 116.7*384 =

Totals 1459 179300

(c) Interpretation

Motorcycle models A, B and C have risen in price by 22.9% on average

from 2009 to 2010.

Exercise 14.9 File: X14.8 - motorcycle sales.xlsx

Motorcycle Unit price Quantity Unit price Quantity

model (R1000) (units sold) (R1000) (units sold)

A 25 10 30 7

B 15 55 19 58

C 12 32 14 40

Motorcycle Quantity

(a) q1/q0*100

model Relative

A 7/10*100 = 70.0

B 58/55*100 = 105.5

C 40/32*100 = 125.0

Interpretation

Model A unit sales dropped 30% from 2009 to 2010; model B unit sales rose

by 5.55%; while model C unit sales rose the most by 25% over the year.

Motorcycle

Base Value (p0*q0) Current Value (p0*q1)

model

A 25*10 = 250 25*7 = 175

B 15*55 = 825 15*58 = 870

C 12*32 = 384 12*40 = 480

Totals 1459 1525

Weighted

Motorcycle

Base Value (p0*q0) Quantity Relative (q1/q0) Quantity

model

Relatives

A 25*10 = 250 7/10*100 = 70.0 70*250 +

B 15*55 = 825 58/55*100 = 105.5 105.5*825 +

C 12*32 = 384 40/32*100 = 125.0 125*384 =

Totals 1459 152537.5

(c) Interpretation

Unit sales of motorcycle models A, B and C have risen by 4.5% on average

from 2009 to 2010.

Exercise 14.10 File: X14.8 - motorcycle sales.xlsx

Motorcycle Unit price Quantity Unit price Quantity

model (R1000) (units sold) (R1000) (units sold)

A 25 10 30 7

B 15 55 19 58

C 12 32 14 40

Motorcycle

Base Value (p1*q0) Current Value (p1*q1)

model

A 30*10 = 300 30* 7 = 210

B 19*55 = 1045 19*58 = 1102

C 14*32 = 448 14*40 = 560

Totals 1793 1872

Weighted

Motorcycle

Base Value (p1*q0) Quantity Relative (q1/q0) Quantity

model

Relatives

A 30*10 = 300 7/10*100 = 70.0 70*300 +

B 19*55 = 1045 58/55*100 = 105.5 105.5*1045 +

C 14*32 = 448 40/32*100 = 125.0 125*448 =

Totals 1793 187200

(b) Interpretation

Unit sales of motorcycle models A, B and C have risen by 4.41% on average

from 2009 to 2010.

Exercise14.11 File: X14.11 - Telkom services.xlsx

Telkom Services Unit price Quantity Unit price Quantity Unit price Quantity

(cents/call) (100's calls) (cents/call) (100's calls) (cents/call) (100's calls)

TalkPlus 65 14 70 18 55 17

SmartAccess 35 27 40 29 45 24

ISDN 50 16 45 22 40 32

TalkPlus 70/65*100 = 107.7 55/65*100 = 84.6

SmartAccess 40/35*100 = 114.3 45/35*100 = 128.6

ISDN 45/50*100 = 90.0 40/50*100 = 80.0

Interpretation

TalkPlus services increased by 7.7% in price from 2009 to 2010, but then dropped by

15.4% in price from 2009 to 2011.

SmartAccess on the other hand showed an increase in price from 2009 to 2010 of 14.3%

and by 28.6% from 2009 to 2011.

ISDN showed a decrease in price by 10% in the first year (from 2009 to 2010) and by 20%

over the 2 year period from 2009 to 2011.

Telkom Services

(p0*q0) (p1*q0) (p1*q0)

TalkPlus 910 980 770

SmartAccess 945 1080 1215

ISDN 800 720 640

2655 2780 2625

The cost of Telkom services increased, on average, by 4.71% from 2009 to 2010,

while there was a net reduction in costs, on average of 1.13% from 2009 to 2011.

Base Value 2010 Value Base Value 2011 Value

Telkom Services

(p1*q0) (p1*q1) (p1*q0) (p1*q1)

TalkPlus 980 1260 770 935

SmartAccess 1080 1160 1215 1080

ISDN 720 990 640 1280

2780 3410 2625 3295

The printing company's usage of Telkom services has risen by 22.7% over one

year (from 2009 to 2010) and by 25.5%, on average over two years (from 2009 to 2011).

Exercise 14.12 File: X14.12 - computer personnel.xlsx

Job categories

2008 2011 2008 2011

Systems analyst 42 50 84 107

Programmer 29 36 96 82

Network manager 24 28 58 64

IT Job Category

2008 (p0*q0) 2011 (p1*q0)

Systems analyst 3528 4200

Programmer 2784 3456

Network manager 1392 1624

7704 9280

20.46% over the 3 years period from 2008 to 2011.

Price relatives

(c) IT Job Category

(p1/p0)

Systems analyst 119.05

Programmer 124.14

Network manager 116.67

in remuneration of 24.14% from 2008 to 2011.

Exercise 14.13 File: X14.12 - computer personnel.xlsx

Job categories

2008 2011 2008 2011

Systems analyst 42 50 84 107

Programmer 29 36 96 82

Network manager 24 28 58 64

Staff Relatives

(a) IT Job Category

(q1/q0)

Systems analyst 127.4

Programmer 85.4

Network manager 110.3

Interpretation

The staff complement of Systems Analysts grew by 27.4% and that of Network

managers grew by 10.3% over the period from 2008 to 2011. However, the

number of Programmers, on the other hand, reduced by 14.6% over this same period.

IT Job Category

2008 (p0*q0) 2011 (p0*q1)

Systems analyst 3528 4494

Programmer 2784 2378

Network manager 1392 1536

7704 8408

Weighted

Base Value 2008 Staff Relatives Ave of

IT Job Category

(p0*q0) (q1/q0) Quantity

Relatives

Systems analyst 3528 127.38 449400

Programmer 2784 85.42 237800

Network manager 1392 110.34 153600

7704 840800

(c) Interpretation

The overall IT staff complement across all job categories has increased by an

average of 9.14% from 2008 to 2011.

Exercise 14.14 File: X14.14 - printer cartridges.xlsx

Data

Printer

cartridges

Quantity Quantity

Unit price Unit price Quantity used Unit price

used used

HQ21 145 24 155 28 149 36

HQ25 172 37 165 39 160 44

HQ26 236 12 255 12 262 14

HQ32 314 10 306 8 299 11

(a) Price relative 2010

Cartridges (2008) (2010)

HQ26 236 262 262/236% = 111.02

HQ32 314 299 299/314% = 95.22

Interpretation

The price of the HQ26 printer cartridge has increased by 11.02% from 2008 to 2010.

The price of the HQ32 printer cartridge has decreased by 4.78% from 2008 to 2010.

Printer Base Value Current Value Base Value Current Value

cartridges (2008) (2009) (2008) (2010)

HQ21 4060 4340 5220 5364

HQ25 6708 6435 7568 7040

HQ26 2832 3060 3304 3668

HQ32 2512 2448 3454 3289

Totals 16112 16283 19546 19361

(b) (ii) Paasche Weighted Average of Relatives Price Index - for 2009

cartridges (2008) (2009) (2009)

HQ21 4060 106.9 434000

HQ25 6708 95.9 643500

HQ26 2832 108.1 306000

HQ32 2512 97.5 244800

Totals 16112 1628300

Printer Base Value Price relative Weighted Ave

cartridges (2008) (2010) (2010)

HQ21 5220 102.8 536400

HQ25 7568 93.0 704000

HQ26 3304 111.0 366800

HQ32 3454 95.2 328900

Totals 19546 1936100

(c) Interpretation

The average price of print cartridges increased marginally by 1.06% from 2008 to 2009.

However, the average price of print cartridges decreased by 0.95% from 2008 to 2010.

2008 2009 2010

Composite price indexes 100 101.06 99.05

Composite link relatives 101.06 98.01

Exercise 14.15 File: X14.15 - electrical goods.xlsx

Composite Price Index

88 96 100 109 114 112 115

(Electrical goods)

Composite Price Index

78.6 85.7 89.3 97.3 101.8 100 102.7

(Electrical goods)

Interpretation

In 2004, the average price of electrical goods was 21.4% below the 2009 (base) price levels,

while in 2006, it was only 10.7% below the 2009 (base) price levels.

In 2010, prices were 2.7% higher on average than in the base period of 2009.

Composite Price Index

100 109.1 104.2 109 104.6 98.2 102.7

(Electrical goods)

Interpretation

The annual average price changes in electrical goods starting in 2004 was 9.1% (for 2005);

4.2% (for 2006); 9% (for 2007); 4.6% (for 2008); -1.8% (decrease in 2009); and 2.7% (for 2010).

Exercise 14.16 File: X14.16 - insurance claims.xlsx

Federal Insurance

92.3 95.4 100 102.6 109.4 111.2

(base = 2008)

Baltic Insurance

93.7 101.1 98.2 100 104.5 107.6

(base = 2009)

Federal Insurance

84.4 87.2 91.4 93.8 100 101.6

(base = 2010)

Baltic Insurance

89.7 96.7 94 95.7 100 103

(base = 2010)

(b) Ferderal Insurance showed a 8.6/91.4% (9.41%) increase from 2008 to 2010.

Baltic Insurance showed a 6.0/94% (6.4%) increase from 2008 to 2010.

Hence Federal Insurance showed the bigger claims increase from 2008 to 2010.

Federal Insurance 100 103.4 104.8 102.6 106.6 101.6

Baltic Insurance 100 107.9 97.1 101.8 104.5 103

Federal showed a year on year increase of only 1.6%. Hence Baltic Insurance.

5

(e) Geometric mean Federal √(1.034*1.048*1.026*1.066*1.016) -1 = 3.785%

5

Baltic √(1.079*0.971*1.018*1.045*1.03) - 1 = 2.799%

between 2006 and 2011.

Exercise 14.17 File: X14.17 - micro-market basket.xlsx

market basket 2010 2011 2010 2011

Milk (litres) 7.29 7.89 117 98

Bread (loaves) 4.25 4.45 56 64

Sugar (kg) 2.19 2.45 28 20

Maize meal (kg) 5.25 5.59 58 64

Milk (litres) 7.29 7.89 108.23

Bread (loaves) 4.25 4.45 104.71

Sugar (kg) 2.19 2.45 111.87

Maize meal (kg) 5.25 5.59 106.48

Interpretation

The price of milk rose by 8.23%; bread by 4.71%; sugar by 11.87% and

maize meal by 6.48% per unit of measure from 2010 to 2011.

The largerst price change (increase) was sugar with a 11.87% increase.

Food items Price Relatives

(p0*q1) Average

Milk (litres) 714.42 108.23 77322

Bread (loaves) 272 104.71 28480

Sugar (kg) 43.8 111.87 4900

Maize meal (kg) 336 106.48 35776

Total 1366.22 146478

On average, the price of the micro-basket of items increased by 7.21% from 2010 to 2011.

Quantity

(c) Food items 2010 2011

Relatives

Milk (litres) 117 98 83.8

Bread (loaves) 56 64 114.3

Sugar (kg) 28 20 71.4

Maize meal (kg) 58 64 110.3

Interpretation

The consumption of milk decreased by 16.2%; bread consumption rose by 14.3%;

sugar consumption decreased significantly by 28.6% and maize meal consumption

rose by 10.3% from 2010 to 2011 respectively.

The largest consumption change (decrease) was sugar with a 28.6% decrease.

It is interesting to note that sugar showed the largest unit price increase while

simultaneously recorded the largest decrease in consumption from 2010 to 2011.

(d) Paasche Weighted Average of Quantity (Consumption) Relatives Index

Food items Quantity Relatives

(p1*q0) Average

Milk (litres) 923.13 83.8 77322

Bread (loaves) 249.20 114.3 28480

Sugar (kg) 68.60 71.4 4900

Maize meal (kg) 324.22 110.3 35776

Total 1565.15 146478

On average, consumption of the micro-basket items dropped by 6.4% from 2010 to 2011.

Exercise 14.18 File: X14.18 - utilities usage.xlsx

Utilities

2008 2009 2010 2008 2009 2010

Electricity 1.97 2.05 2.09 745 812 977

Sewage 0.62 0.68 0.72 68 56 64

Water 0.29 0.31 0.35 296 318 378

Telephone 1.24 1.18 1.06 1028 1226 1284

Price

Unit Price Unit Price

(a) Utilities relative

2008 2010

2010

Electricity 1.97 2.09 106.1

Sewage 0.62 0.72 116.1

Water 0.29 0.35 120.7

Telephone 1.24 1.06 85.5

Interpretation

From 2008 to 2010, electricity increased by 6.1%; sewage costs by 16.1%

water by 20.7%; while telephone costs decreased by 14.5% over this period.

Electricity showed the smallest change (increase) of 6.1% from 2008 to 2010.

Current Current

Base Value

Utilities Value Value

(2008)

(2009) (2010)

Electricity 1467.7 1527.3 1557.1

Sewage 42.2 46.2 49.0

Water 85.8 91.8 103.6

Telephone 1274.7 1213.0 1089.7

Totals 2870.4 2878.3 2799.3

Price Price

Base Value Weighted Weighted

Utilities relative relative

(2008) Ave (2009) Ave (2010)

(2009) (2010)

Electricity 1467.7 104.1 106.1 152782.4 155717.7

Sewage 42.2 109.7 116.1 4625.0 4894.8

Water 85.8 106.9 120.7 9176.3 10360.9

Telephone 1274.7 95.2 85.5 121353.3 108988.6

Totals 2870.4 287937.0 279961.9

(c) Interpretation

There was virtually no change in the average cost of household utilities between

2008 and 2009. A slight increase of only 0.3% was recorded.

From 2008 to 2010, however, the average cost of household utilities actually

decreased marginally by 2.5%.

Exercise 14.19 File: X14.18 - utilities usage.xlsx

Utilities

2008 2009 2010 2008 2009 2010

Electricity 1.97 2.05 2.09 745 812 977

Sewage 0.62 0.68 0.72 68 56 64

Water 0.29 0.31 0.35 296 318 378

Telephone 1.24 1.18 1.06 1028 1226 1284

Quantity

Usage Usage

(a) Utilities relative

2008 2010

2010

Electricity 745 977 131.1

Sewage 68 64 94.1

Water 296 378 127.7

Telephone 1028 1284 124.9

Interpretation

Consumption of electricity, water and telephone increased by 31.1%; 27.7% and

24.9% respectively from 2008 to 2010. Only sewage showed a decline in usage

by 5.9% from 2008 to 2010.

Current Current

Base Value

Utilities Value Value

(2008)

(2009) (2010)

Electricity 1467.7 1599.6 1924.7

Sewage 42.2 34.7 39.7

Water 85.8 92.2 109.6

Telephone 1274.7 1520.2 1592.2

Totals 2870.4 3246.8 3666.2

Quantity Quantity

Base Value Weighted Weighted

Utilities relative relative

(2008) Ave (2009) Ave (2010)

(2009) (2010)

Electricity 1467.7 109.0 131.1 159959 192468

Sewage 42.2 82.4 94.1 3474 3968

Water 85.8 107.4 127.7 9219 10962

Telephone 1274.7 119.3 124.9 152074 159213

Totals 2870.4 324726 366610

(c) Interpretation

On average, household consumption of utilities increased by 13.1% in 2009 from 2008

and also showed an overall consumption increase of 27.7% from 2008 to 2010.

Exercise 14.20 File: X14.20 - leather goods.xlsx

Composite

97 92 100 102 107 116 112

cost index

Composite

90.7 86 93.5 95.3 100 108.4 104.7

cost index

(b)

Composite cost index

120

110

100

Cost Index

90

80

60

2005 2006 2007 2008 2009 2010 2011

Interpretation

Prior to 2009, the average annual cost of leather goods inputs were below the 2009 level,

but showed a steady increase towards 2009 prices over this period.

Relative to 2009 unit costs, the cost of leather

goods inputs was higher by 8.4% and 4.7% respectively for 2010 and 2011.

(c) The average increase in unit costs of leather goods inputs was only 4.7% between

2009 and 2011.

Composite

100 94.8 108.7 102.0 104.9 108.4 96.6

cost index

The largest year-on-year change in overall unit costs of leather goods inputs was

between 2006 and 2007 with an increase of 8.7%.

6

(e) Geometric mean √(0.948*1.087*1.02*1.049*1.0841*0.966)

1.024244

(i.e. 2.424% increase on average per year)

=GEOMEAN(0.9485,1.087,1.02,1.049,1.0841,0.9655)

= 1.024244

Exercise 14.21 File: X14.21 - accountants' salaries.xlsx

2005 387 95 407.4

2006 406 100 406.0

2007 422 104 405.8

2008 448 111 403.6

2009 466 121 385.1

2010 496 126 393.7

2011 510 133 383.5

(b) Interpretation

The base real salary is R406 (in R1000s). Since 2006, real salaries have

declined relative to 2006 (base) and are continuing to fall further behind inflation (CPI).

Year Salary CPI Salary CPI Diff (Sal - CPI)

2005 387 95

2006 406 100 104.9 105.3 -0.4

2007 422 104 103.9 104.0 -0.1

2008 448 111 106.2 106.7 -0.6

2009 466 121 104.0 109.0 -5.0

2010 496 126 106.4 104.1 2.3

2011 510 133 102.8 105.6 -2.7

Interpretation

On a year-on-year basis, salary increases have lagged behind the inflation rate (CPI)

in all years except 2010 when salary adjustments exceeded CPI by 2.3%.

Exercise 14.22 File: X14.22 - school equipment .xlsx

Composite

94.8 97.6 100 105.2 108.5 113.9 116.7 121.1

cost index

Composite

83.2 85.7 87.8 92.4 95.3 100 102.5 106.3

cost index

(b) and (e) Plot and Interpretation of Re-based Cost Index Series

110

105

100

95

90

85

75

2003 2004 2005 2006 2007 2008 2009 2010

Interpretation

The average cost of school equipment has shown a steady increase of 3% to 5%

annually over the period 2003 to 2010.

Composite

100 102.95 102.46 105.20 103.14 104.98 102.46 103.77

price index

Interpretation

The link relatives confirm the year-on-year increases in the average cost of school

equipment of between 3% to 5% p.a.

(d) The 2003 budget must be adjusted annually by the link relative (year-on-year) indexes.

Therefore multiply each previous year's budget by the next year's link relative index.

Budgets 5000000 5147500 5274129 5548384 5722603 6007589 6155376 6387434

Exercise 14.23 File: X14.23 - coffee imports.xlsx

Data Coffee types

(2007) (2007) (2010) (2010)

Java 85 52 98 46

Colombia 64 75 74 90

Sumatra 115 18 133 20

Mocha 38 144 42 168

(a) Coffee types Price Relatives

(2007) (2010)

Java 85 98 115.3

Colombia 64 74 115.6

Sumatra 115 133 115.7

Mocha 38 42 110.5

Interpretation

The coffee types of Java, Colombia and Sumatra increased by over 15% from

2007 to 2010, while Mocha increased by only 10.5% over this same period.

Coffee type

(2007) (2010)

Java 4420 5096

Colombia 4800 5550

Sumatra 2070 2394

Mocha 5472 6048

Total 16762 19088

Interpretation

The cost of coffee imports has increased by 13.9% on average from 2007 to 2010.

(c) Cheaper.

The cost of coffee imports has increased by only 13.9% on average, since 2007.

(d) Laspeyres Composite Quantity Index (Weighted Aggregates approach)

Coffee type

(2007) (2010)

Java 4420 3910

Colombia 4800 5760

Sumatra 2070 2300

Mocha 5472 6384

Total 16762 18354

Interpretation

The quantity of coffee imported has increased by 9.5% on average from 2007 to 2010.

Exercise 14.24 File: X14.24 - medical claims.xlsx

Data Claim Type

(2008) (2008) (2010) (2010)

GPs 220 20 255 30

Specialists 720 30 822 25

Dentists 580 10 615 15

Medicines 400 50 438 70

(Use the Laspeyres Approach as the default method if none is specified)

Claim type

(2008) (2010)

GP's 4400 6600

Specialists 21600 18000

Dentists 5800 8700

Medicines 20000 28000

Total 51800 61300

Interpretation

The number of claims received has increased by 18.34% from 2008 to 2010.

Quantity

(b) Claim type Claims (2008) Claims (2010)

relatives

GP's 20 30 150.0

Specialists 30 25 83.33

Dentists 10 15 150.0

Medicines 50 70 140.0

Interpretation

GP's and Dentists showed the biggest increase in number of claims (both 50% increase).

(c) Laspeyres Composite Price (Claims Value) Index (Weighted Aggregates approach)

Claim type

(2008) (2010)

GP's 4400 5100

Specialists 21600 24660

Dentists 5800 6150

Medicines 20000 21900

Total 51800 57810

(d) Interpretation

The value of claims received between 2008 and 2010 has risen by 11.6%.

Exercise 14.25 File: X14.25 - tennis shoes.xlsx

Data Shoe Model

(2009) (2009) (2010) (2010)

Trainer 320 96 342 110

Balance 445 135 415 162

Dura 562 54 595 48

Shoe model

(2009) (2010) relatives

Trainer 96 110 114.6

Balance 135 162 120.0

Dura 54 48 88.9

Interpretation

Dura's sales volume is down by 11.1% from 2009.

Shoe model

(2009) (2010)

Trainer 30720 35200

Balance 60075 72090

Dura 30348 26976

Total 121143 134266

(c) Interpretation

The average increase in shoes sold from 2009 to 2010 was only 10.83%.

Overall, sales volumes do not meet the required growth of at least 12% p.a.

Exercise 14.26 File: X14.26 - energy fund.xlsx

Data

2001 2002 2003 2004 2005 2006 2007 2008 2009 2010

Fuel

cost 100.0 116.2 122.4 132.1 135.7 140.3 142.8 146.9 153.4 160.5

index

2001 2002 2003 2004 2005 2006 2007 2008 2009 2010

Fuel

cost 71.28 82.83 87.25 94.16 96.73 100.00 101.79 104.71 109.34 114.40

index

2001 2002 2003 2004 2005 2006 2007 2008 2009 2010

Fuel

cost 100.00 116.20 105.34 107.92 102.73 103.39 101.78 102.87 104.42 104.63

index

120.00

115.00

110.00

105.00

100.00

95.00

Fuel cost index

90.00

2001 2002 2003 2004 2005 2006 2007 2008 2009 2010

(c) Geometric Mean 9√(1.162*1.0534*1.0792*1.0273*1.0339*1.0178*1.0287*1.0442*1.0463)

1.05397

Annual Average Increase = 5.3974% p.a. on average.

(d) Interpretation

Annual fuel cost increases and decreases were high in the period 2001 to 2004.

Thereafter, year-on-year increases were moderate and stable at between 3% to 5% p.a.

Exercise 14.27 File: X14.27 - motorcycle distributor.xlsx

Data Model 2007 2008 2009 2010 2011

Blitz 18050 19235 21050 21950 22400

Cruiser 25650 26200 27350 28645 31280

Classic 39575 42580 43575 43950 46750

Units Sold

Model 2007 2008 2009 2010 2011

Blitz 205 185 168 215 225

Cruiser 462 386 402 519 538

Classic 88 70 111 146 132

Cruiser 100.0 102.1 106.6 111.7 121.9

The price of the Cruiser model rose marginally (between 2% and 7%) for 2008

and 2009 relative to 2007; but rose strongly by 11.7% in 2010 and 21.9% in 2011.

Model Value 2008

2007 2009 2010 2011

Blitz 3700250 3943175 4315250 4499750 4592000

Cruiser 11850300 12104400 12635700 13233990 14451360

Classic 3482600 3747040 3834600 3867600 4114000

Total 19033150 19794615 20785550 21601340 23157360

Price Index 100 104.0 109.2 113.5 121.7

130

120

110

100

90

80

70 Price Index

60

2007 2008 2009 2010 2011

(c) Interpretation

There has been a reasonably constant increase in the average price of motorcycles

over the past 5 years. In 2011, motorcycles cost 21.7% more on average than

they did in 2007.

Classic 100 107.6 102.3 100.9 106.4

Interpretation

The Classic model has shown a very modest year-on-year price increase of

between 2.3% and 7.6% between 2007 and 2011. In 2010, the annual

(year-on-year) increase was only 0.9% (less than 1%).

4

(e) Geometric Mean = √ (1.076*1.023*1.009*1.064) - 1= 4.263

Using Excel = GEOMEAN(link relatives) - 100 = 4.253

Model Value 2008

2007 2009 2010 2011

Blitz 3700250 3339250 3032400 3880750 4061250

Cruiser 11850300 9900900 10311300 13312350 13799700

Classic 3482600 2770250 4392825 5777950 5223900

Total 19033150 16010400 17736525 22971050 23084850

Quantity Index 100 84.1 93.2 120.7 121.3

130

120

110

100

90

80

Quantity Index

70

60

2007 2008 2009 2010 2011

(g) Interpretation

Sales of motorcycles dropped for two years after 2007 (i.e. by 15.9% in 2008 and

by 6.8% in 2009). Thereafter unit sales, on average per model, increased by 20.7%

in 2010 relative to 2007; but performed at the same level in 2011 compared to 2010.

Exercise 14.28 File: X14.28 - tyre production.xlsx

Cost/Tyre 2010-Jan Feb March April May June July Aug Sept Oct Nov Dec

Passenger 210.69 212.47 210.73 218.14 219.22 216.19 225.92 234 229.89 222.76 223.96 200.3

Light truck 376.45 361.7 361.76 363.94 363.62 364.06 376.9 375.4 375.55 375.04 375.59 376.3

Giant truck 1171.1 1109.6 1101.8 1119.7 1127.2 1120.32 1162.8 1181 1157.7 1166.75 1157.7 1148

Output (1000's) 2010-Jan Feb March April May June July Aug Sept Oct Nov Dec

Passenger 78 102 93 81 105 100 117 105 98 110 97 43

Light truck 11 14 13 12 16 16 19 17 14 15 13 6

Giant truck 10 14 13 11 15 15 16 16 16 15 14 5

Cost relative

Jan Feb March April May June July Aug Sept Oct Nov Dec

index series

Passenger tyre 100.0 100.8 100.0 103.5 104.0 102.6 107.2 111.1 109.1 105.7 106.3 95.1

Interpretation

Passenger tyre costs increased steadily throughout the year reaching a peak of 11.1% above January

2010 levels in August. Thereafter costs declined steadily and ended the year 4.9% below the

starting level in January 2010.

(b) PRODUCTION COST ANALYSIS Laspeyres Composite Cost Index (Weighted Aggregates)

Jan Feb March April May June July Aug Sept Oct Nov Dec

Passenger 16433.82 16573 16437 17015 17099 16862.8 17622 18250 17931 17375.3 17469 15623

Light truck 4140.95 3978.7 3979.4 4003.3 3999.8 4004.66 4145.9 4129 4131.1 4125.44 4131.5 4139

Giant truck 11711 11096 11018 11197 11272 11203.2 11628 11812 11577 11667.5 11577 11479

Total Cost 32285.77 31647 31435 32216 32371 32070.7 33396 34192 33639 33168.2 33177 31241

Jan Feb March April May June July Aug Sept Oct Nov Dec

Composite Cost

100.0 98.0 97.4 99.8 100.3 99.3 103.4 105.9 104.2 102.7 102.8 96.8

Index

108.0

106.0

104.0

102.0

100.0

98.0

96.0

94.0

Composite Cost Index

92.0

Jan Feb March April May June July Aug Sept Oct Nov Dec

Interpretation

Composite tyre manufacturing costs declined or remained constant for the first 6 months

until June 2010. Thereafter, costs, on average increased by 6% over July and August

before being brought under control. By December, composite costs were 3.2% below

the beginning of the year levels.

(c) Link Relatives - Costs - Light Truck Tyres

Jan Feb March April May June July Aug Sept Oct Nov Dec

Light truck 376.45 361.7 361.76 363.94 363.62 364.06 376.9 375.4 375.55 375.04 375.59 376.3

Link Relatives 100.0 96.1 100.0 100.6 99.9 100.1 103.5 99.6 100.0 99.9 100.1 100.2

Interpretation

Light truck radial tyres costs have remained almost constant and unchanged throughout

the year with most adjustments not exceeding 0.5%. Only in Feb (decrease of 3.9%) and

July (increase of 3.5%) did costs fluctuate to any degree.

Jan Feb March April May June July Aug Sept Oct Nov Dec

Passenger 16433.82 21490 19594 17066 22122 21069 24651 22122 20648 23175.9 20437 9060

Light truck 4140.95 5270.3 4893.9 4517.4 6023.2 6023.2 7152.6 6400 5270.3 5646.75 4893.9 2259

Giant truck 11711 16395 15224 12882 17567 17566.5 18738 18738 18738 17566.5 16395 5856

Total Value 32285.77 43156 39712 34465 45712 44658.7 50541 47260 44656 46389.2 41726 17174

Jan Feb March April May June July Aug Sept Oct Nov Dec

Volume Index 100.0 133.7 123.0 106.8 141.6 138.3 156.5 146.4 138.3 143.7 129.2 53.2

180.0

160.0

140.0

120.0

100.0

80.0

60.0

40.0

20.0 Volume Index

0.0

Jan Feb March April May June July Aug Sept Oct Nov Dec

(e) Interpretation

Production volumes of all three makes of tyres showed a steady increase above the January

level by up to nearly 60% over the first half of the year. However output showed a steady

decline in the second half of the year ending at only 53.2% of the beginning of the year levels.

CHAPTER 15

A FORECASTING TOOL

Time series data is recorded at fixed intervals over time.

Daily maximum temperature for Cape Town.

Seasonality tends to show the most regularity.

fluctuations in a time series to allow the longer-term

underlying trend and cyclical patterns to be revealed.

Exercise 15.6 Yes, averaging occurs over a longer time period (i.e. five periods)

producing a smoother curve.

Exercise 15.7 A seasonal index of 108 means that seasonal influences stimulate

the time series values by 8% above the trend / cyclical level.

the time series values by 12% below the trend / cyclical level.

Exercise 15.9 File: X15.9 - coal tonnage.xlsx

Uncentred 4 Centred 4- Centred 5-year

Year Coal Tonnage year moving year Moving Moving

total Average Average

1 118

2 124 470

3 108 484 119.25 120.4

4 120 475 119.875 119.8

5 132 489 120.5 119.4

6 115 517 125.75 127.4

7 122 545 132.75 135.4

8 148 618 145.375 146.6

9 160 697 164.375 163.8

10 188 723 177.5 174.2

11 201 754 184.625 182.8

12 174 744 187.25 186.4

13 191 689 179.125 178

14 178 676 170.625 170

15 146

16 161

Line Graphs of Coal mined, 4 and 5 year Moving Averages

250

coal mined (100 000 tonnes)

200

150

Coal Tonnage

100

Centred 4- year Moving

50 Average

Centred 5-year Moving

Average

0

1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16

years

(e) Interpretation

The annual tonnage of coal mined in the Limpopo province was constant for the first

7 years after which there was an expansion phase for the next 4 years. Thereafter coal

production has been declining steadily. This could be evidence of a cyclical effect

caused by economic cycles in the demand for coal worldwide.

Overall, a moderate upward cyclical trend.

Exercise 15.10 File: X15.10 - franchise dealers.xlsx

60

50

no. of new dealers

40

30

20

10

0

1 2 3 4 5 6 7 8 9 10

time periods

1 28 1 28

2 32 4 64

3 43 9 129

4 31 16 124

5 38 25 190

6 47 36 282

7 40 49 280

8 45 64 360

9 55 81 495

n = 10 10 42 100 420

Σ 55 401 385 2372

b1 = ((10*2372)-(55*401))/(10*385-55²) = 2.0182

b0 = (401-2.0182*55)/10 = 29

ŷ= 29 + 2.0182 x x = 1, 2, 3, …, 10

x substitute into ŷ ŷ

11 29 + 2.0182 (11) 51.20

12 29 + 2.0182 (12) 53.22

13 29 + 2.0182 (13) 55.24

periods 11, 12 and 13 respectively, based on trend estimates.

Exercise 15.11 File: X15.11 - policy claims.xlsx

90

80

70

60

no. of claims

50

40

30

20

10

0

1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16

quarters

1 84 1 84

2 53 4 106

3 60 9 180

4 75 16 300

5 81 25 405

6 57 36 342

7 51 49 357

8 73 64 584

9 69 81 621

10 37 100 370

11 40 121 440

12 77 144 924

13 73 169 949

14 46 196 644

15 39 225 585

n = 16 16 63 256 1008

Σ 136 978 1496 7899

b0 = =(978-(-1.21765)*136)/16 71.475

1 in Q1 2008

ŷ= 71.475 - 1.21765 x x= 2 in Q2 2008

3 in Q3 2008

Interpretation

There is a downward trend in household policy claims over the past 4 years.

Uncentred 2x4 Centred 4

Unadjusted Adjusted

Time 4 period period period Seasonal

(c) Claims Seasonal Seasonal

Periods Moving Moving Moving ratios

Indexes Indexes

Total Total Average

2008 Q1 84

2008 Q2 53 272

2008 Q3 60 269 541 67.625 88.725 79.688 78.705

2008 Q4 75 273 542 67.75 110.701 121.667 120.166

2009 Q1 81 264 537 67.125 120.670 122.940 121.423

2009 Q2 57 262 526 65.75 86.692 80.702 79.706

2009 Q3 51 250 512 64 79.688

2009 Q4 73 230 480 60 121.667

2010 Q1 69 219 449 56.125 122.940

2010 Q2 37 223 442 55.25 66.968

2010 Q3 40 227 450 56.25 71.111

2010 Q4 77 236 463 57.875 133.045

2011 Q1 73 235 471 58.875 123.992

2011 Q2 46 221 456 57 80.702

2011 Q3 39

2011 Q4 63 Totals 404.996 400.000

Q2 79.7

Q3 78.7

Q4 120.2

Interpretation

Household policy claims tend to increase significantly in Quarters 1 and 4 of each year,

by about 20% on average, while there is a significant decline in claims during

Quarters 2 and 3 by about 20% on average.

x = 17 in Quarter 1 2012

Trend estimate ŷ= 71.475 - 1.21765 (17) = 50.77495

x = 18 in Quarter 1 2012

Trend estimate ŷ= 71.475 - 1.21765 (18) = 49.5573

Q2 2012 ŷ (adj) = 49.55742*0.797 = 39.5

Interpretation

The insurance company can expect to receive 62 and 40 (rounded) household policy

claims in the first and second quarters of 2012 respectively.

Exercise 15.12 File: X15.12 - hotel occupancy.xlsx

100

90

occupancy rate (%)

80

70

60

50

40

1 2 3 4 5 6 7 8 9 10

1 74 1 74

2 82 4 164

3 70 9 210

4 90 16 360

5 88 25 440

6 74 36 444

7 64 49 448

8 69 64 552

9 58 81 522

n = 10 10 65 100 650

Σ 55 734 385 3864

1 in Sept

ŷ= 84.933 - 2.09697 x x= 2 in Oct

3 in Nov

Interpretation

There is a downward trend in hotel occupancy over the past 10 months since September.

July 11 61.87%

Aug 12 59.77%

Interpretation

The continued downward trend is reflected in the next 2 month's projections.

Exercise 15.13 File: X15.13 - electricity demand.xlsx

(a), (e), (f) Line Graph - Electricity Demand for a City (Cape Town)

y = 4.95x + 20.8

180

160

demand (1000 megawatts)

140

120

100

80

60

40

20

0

1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16

quarters (Q1 2008 - Q4 2011)

Interpretation

Electricity demand in Cape Town shows a clear seasonal pattern, peaking in quarter 3

and bottoming out in quarter 4 of each year.

1 21 1 21

2 42 4 84

3 60 9 180

4 12 16 48

5 35 25 175

6 54 36 324

7 91 49 637

8 14 64 112

9 39 81 351

10 82 100 820

11 136 121 1496

12 28 144 336

13 78 169 1014

14 114 196 1596

15 160 225 2400

n = 16 16 40 256 640

Σ 136 1006 1496 10234

1 in Q1 2008

ŷ= 20.8 + 4.95 x x= 2 in Q2 2008

3 in Q3 2008

(c)

Uncentred Centred 4

2x4 period Unadjusted Adjusted

Time 4 period period Seasonal

Demand Moving Seasonal Seasonal

Periods Moving Moving ratios

Total Indexes Indexes

Total Average

2008 Q1 21

2008 Q2 42 135

2008 Q3 60 149 284 35.5 169.01 178.65 175.61

2008 Q4 12 161 310 38.75 30.97 30.97 30.44

2009 Q1 35 192 353 44.125 79.32 79.32 77.97

2009 Q2 54 194 386 48.25 111.92 117.99 115.98

2009 Q3 91 198 392 49 185.71

2009 Q4 14 226 424 53 26.42

2010 Q1 39 271 497 62.125 62.78

2010 Q2 82 285 556 69.5 117.99

2010 Q3 136 324 609 76.125 178.65

2010 Q4 28 356 680 85 32.94

2011 Q1 78 380 736 92 84.78

2011 Q2 114 392 772 96.5 118.13

2011 Q3 160

2011 Q4 40 Totals 406.927 400.000

Q2 115.98

Q3 175.61

Q4 30.44

Interpretation

Electricity demand peaks in Q3 by 75% over the trend / cyclical level; and drops

to 70% below the trend /cyclical level during Q4.

Seasonally

Seasonal

Period x Trend ŷ adjusted

Index

Trend

Q3 2012 19 114.85 175.61 201.69

Q4 2012 20 119.8 30.44 36.47

Interpretation

Electricity demand in Cape Town is likely to peak at 201.69 megawatts in

Q3 of 2012 and bottom out at 36.47 megawatts in Q4 of 2012.

Exercise 15.14 File: X15.14 - hotel turnover.xlsx

Seaonal De-

(a) Actual

Index seasonalised

568 136 417.6

495 112 442.0

252 62 406.5

315 90 350.0

604 136 444.1

544 112 485.7

270 62 435.5

510 90 566.7

662 136 486.8

605 112 540.2

310 62 500.0

535 90 594.4

Hotel Industry Quarterly Turnover

700

600

turnover (R millions)

500

400

300

200

0

1 2 3 4 5 6 7 8 9 10 11 12

1 568 1 568

2 495 4 990

3 252 9 756

4 315 16 1260

5 604 25 3020

6 544 36 3264

7 270 49 1890

8 510 64 4080

9 662 81 5958

10 605 100 6050

11 310 121 3410

n = 12 12 535 144 6420

Σ 78 5670 650 37666

b1 = (12*37666 - 78*5670)/(12*650 - 78²) = 5.6713

1 in Summer 2008

ŷ= 435.64 + 5.6713 x x= 2 in Autumn 2008

3 in Winter 2008

Seasonally

Seasonal

Period x Trend ŷ adjusted

Index

Trend

Summer 2011 13 509.37 136 692.74

Autumn 2011 14 515.04 112 576.84

y = 5.6713x + 435.64

700

600

turnover (R millions)

500

400

300

100

1 2 3 4 5 6 7 8 9 10 11 12

quarters

Exercise 15.15 File: X15.15 - farming equipment.xlsx

y = 0.8544x + 52.05

75

70

65

no. units sold

60

55

50

45

Sales

40

35 Linear (Sales)

30

1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16

Uncentred Centred 4

2x4 period Unadjusted Adjusted

4 period period Seasonal

Periods Sales Moving Seasonal Seasonal

Moving Moving Ratios

Total Indexes Indexes

Total Average

2008 Q1 57

2008 Q2 51 214

2008 Q3 50 217 431 53.88 92.81 92.61 92.63

2008 Q4 56 222 439 54.88 102.05 102.95 102.97

2009 Q1 60 225 447 55.88 107.38 107.38 107.40

2009 Q2 56 230 455 56.88 98.46 96.97 97.00

2009 Q3 53 235 465 58.13 91.18

2009 Q4 61 239 474 59.25 102.95

2010 Q1 65 244 483 60.38 107.66

2010 Q2 60 251 495 61.88 96.97

2010 Q3 58 250 501 62.63 92.61

2010 Q4 68 252 502 62.75 108.37

2011 Q1 64 252 504 63.00 101.59

2011 Q2 62 254 506 63.25 98.02

2011 Q3 58

2011 Q4 70 Totals 399.92 400

Autumn 97.00

Winter 92.63

Spring 102.97

The influence of seasonal forces on farming equipment sales is modest.

There is a small stimulatory effect during Spring and Summer (no more than 7%) and

a small depressing effect (also no more than 7%) during Autumn and Winter.

(c) Period (x) Sales (y) x² xy

1 57 1 57

2 51 4 102

3 50 9 150

4 56 16 224

5 60 25 300

6 56 36 336

7 53 49 371

8 61 64 488

9 65 81 585

10 60 100 600

11 58 121 638

12 68 144 816

13 64 169 832

14 62 196 868

15 58 225 870

n = 16 16 70 256 1120

Σ 136 949 1496 8357

b1 = (16*8357-136*949)/(16*1496-136²) = 0.8544

1 in Summer 2008

ŷ= 52.05 + 0.8544 x x= 2 in Autumn 2008

3 in Winter 2008

Seasonally

Seasonal

Period x Trend ŷ adjusted

Index

Trend

Summer 17 66.57 107.40 71.50

Autumn 18 67.43 97.00 65.41

Winter 19 68.28 92.63 63.25

Spring 20 69.14 102.97 71.19

Interpretation

The company can expect to sell between 63 and 72 farming implements

each quarter during 2012 with the higher sales expected in Summer and Spring.

Exercise 15.16 File: X15.16 - energy costs.xlsx

(a), (e) and (f) Line Graph of Office Complex Energy Costs (in R100 000)

5

4.5

energy cost (R100 000)

4

3.5

3

2.5

2

1.5

1

1 2 3 4 5 6 7 8 9 10 11 12

quarters (2009 - 2011)

(b)

Unadjusted Adjusted

Time period period period Seasonal

Costs Seasonal Seasonal

Periods Moving Moving Moving ratios

Indexes Indexes

Total Total Average

2009 Q1 2.4

2009 Q2 3.8 13.3

2009 Q3 4 13.5 26.80 3.35 119.40 118.27 117.6

2009 Q4 3.1 13.8 27.30 3.4125 90.84 90.49 90.0

2010 Q1 2.6 13.9 27.70 3.4625 75.09 73.41 73.0

2010 Q2 4.1 14.0 27.90 3.4875 117.56 120.21 119.5

2010 Q3 4.1 14.0 28.00 3.5 117.14

2010 Q4 3.2 14.4 28.40 3.55 90.14

2011 Q1 2.6 14.6 29.00 3.625 71.72

2011 Q2 4.5 14.7 29.30 3.6625 122.87

2011 Q3 4.3

2011 Q4 3.3 Totals 402.39 400

Autumn 119.5

Winter 117.6

Spring 90.0

Interpretation

Energy costs rise by nearly 20% (19,5% and 17,6%) over the colder months of

Autumn and Winter, and decline by between 10% (in Spring) and almost 30%

(27%) in Summer.

(c) Period (x) Cost (y) x² xy

1 2.4 1 2.4

2 3.8 4 7.6

3 4 9 12

4 3.1 16 12.4

5 2.6 25 13

6 4.1 36 24.6

7 4.1 49 28.7

8 3.2 64 25.6

9 2.6 81 23.4

10 4.5 100 45

11 4.3 121 47.3

n = 12 12 3.3 144 39.6

Σ 78 42 650 281.6

2

b1 = [(12*281.6)-(78*42)]/[12*650-78 ] = 0.0601

1 in Summer 2009

ŷ= 3.1091 + 0.0601 x x= 2 in Autumn 2009

3 in Winter 2009

(d) Seasonally-adjusted trend estimate of Office Complex Energy Costs for 2012

Seasonally

Seasonal

Period x Trend ŷ adjusted

Index

Trend

Summer 13 3.89 72.97 2.84

Autumn 14 3.95 119.50 4.72

Winter 15 4.01 117.57 4.71

Spring 16 4.07 89.95 3.66

Interpretation

The office complex manager must budget between R284 000 and R472 000 each

quarter during 2012 with higher costs expected in the Autumn and Winter periods.

Exercise 15.17 File: X15.17 - business registrations.xlsx

2500

no. of new registrations

2000

1500

1000

500

0

1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20

(b) and (c) 4-Period Moving Average and Quarterly Seasonal Indexes

(b) (c)

2x4 Centred 4

Uncentred 4 Unadjusted Adjusted

Time New period period Seasonal

period Moving Seasonal Seasonal

Periods Registrations Moving Moving ratios

Total Indexes Indexes

Total Average

2007 Q1 1005

2007 Q2 1222 4724

2007 Q3 1298 4892 9616 1202.00 107.99 106.526 106.41

2007 Q4 1199 5041 9933 1241.63 96.57 98.514 98.41

2008 Q1 1173 5199 10240 1280.00 91.64 91.820 91.72

2008 Q2 1371 5376 10575 1321.88 103.72 103.568 103.46

2008 Q3 1456 5517 10893 1361.63 106.93

2008 Q4 1376 5677 11194 1399.25 98.34

2009 Q1 1314 5826 11503 1437.88 91.38

2009 Q2 1531 5980 11806 1475.75 103.74

2009 Q3 1605 6125 12105 1513.13 106.07

2009 Q4 1530 6265 12390 1548.75 98.79

2010 Q1 1459 6422 12687 1585.88 92.00

2010 Q2 1671 6569 12991 1623.88 102.90

2010 Q3 1762 6714 13283 1660.38 106.12

2010 Q4 1677 6880 13594 1699.25 98.69

2011 Q1 1604 7034 13914 1739.25 92.22

2011 Q2 1837 7176 14210 1776.25 103.42

2011 Q3 1916

2011 Q4 1819 Totals 400.429 400

Q2 103.46

Q3 106.41

Q4 98.41

(d) Interpretation of Seasonal Influences on new business registrations

New business registrations show a modest seasonal pattern, ranging between

8,3% below the annual average in quarter 1 to 6.4% above the annual average

in quarter 3.

(b) 4-Period Moving Average Line Graph and Original Data Line Graph

2500

2000

no. of new businesses

1500

1000

New Registrations

500

Centred 4 period Moving

Average

0

1 in Q1 2007

ŷ= 1078.2 + 39.338 x x= 2 in Q2 2007

3 in Q3 2007

(f) Seasonally-adjusted Trend estimates

Adjusted

Trend

Periods y Seasonal ŷ (adj)

Estimate ŷ

Indexes

2007 Q1 1005 91.72 1117.5 1025.0

2007 Q2 1222 103.46 1156.9 1196.9

2007 Q3 1298 106.41 1196.2 1272.9

2007 Q4 1199 98.41 1235.6 1215.9

2008 Q1 1173 91.72 1274.9 1169.3

2008 Q2 1371 103.46 1314.2 1359.7

2008 Q3 1456 106.41 1353.6 1440.3

2008 Q4 1376 98.41 1392.9 1370.8

2009 Q1 1314 91.72 1432.2 1313.7

2009 Q2 1531 103.46 1471.6 1522.5

2009 Q3 1605 106.41 1510.9 1607.8

2009 Q4 1530 98.41 1550.3 1525.6

2010 Q1 1459 91.72 1589.6 1458.0

2010 Q2 1671 103.46 1628.9 1685.3

2010 Q3 1762 106.41 1668.3 1775.2

2010 Q4 1677 98.41 1707.6 1680.5

2011 Q1 1604 91.72 1746.9 1602.3

2011 Q2 1837 103.46 1786.3 1848.1

2011 Q3 1916 106.41 1825.6 1942.6

2011 Q4 1819 98.41 1865.0 1835.3

(g) Plot of Actual vs Seasonally adjusted Trend estimates (New Business Registrations)

2000

no. of new registrations

1800

1600

1400

1200

1000 y ŷ (adj)

800

2007 Q1

2007 Q2

2007 Q3

2007 Q4

2008 Q1

2008 Q2

2008 Q3

2008 Q4

2009 Q1

2009 Q2

2009 Q3

2009 Q4

2010 Q1

2010 Q2

2010 Q3

2010 Q4

2011 Q1

2011 Q2

2011 Q3

2011 Q4

Comment

The seasonally adjusted trend estimates track the actual number of new business

registrations very closely. It is a good fitting graph.

(h) Seasonally-adjusted trend estimate of New Business Registrations (2012)

Seasonally

Seasonal

Period x Trend ŷ adjusted

Index

Trend

Q1 2012 21 1904.3 91.72 1746.6

Q2 2012 22 1943.6 103.46 2010.9

Q3 2012 23 1983.0 106.41 2110.1

Q4 2012 24 2022.3 98.41 1990.2

Exercise 15.18 File: X15.18 - engineering sales.xlsx

Seasonally

Seasonal

Quarter Trend ŷ Adjusted Trend

Index

Estimate ŷ(adj)

1 95 12 11.4

2 115 12 13.8

3 110 12 13.2

4 80 12 9.6

48 48

Exercise 15.19 File: X15.19 - Table Mountain.xlsx

Wi 2009 18.1 62 29.19

Sp 2009 26.4 89 29.66

Su 2009 41.2 162 25.43

Au 2009 31.6 87 36.32

Wi 2010 22.4 62 36.13

Sp 2010 33.2 89 37.30

Su 2010 44.8 162 27.65

Au 2010 32.5 87 37.36

(Actual vs De-Seasonalised)

50

45

no. of visitors (1000's)

40

35

30

25

20

15

10 Visitors

5 De-Seasonlised

0

Wi 2009 Sp 2009 Su 2009 Au 2009 Wi 2010 Sp 2010 Su 2010 Au 2010

Comment

The trend (after de-seasonalising the quarterly data) is only marginally upwards.

Seasonally

Quarter Seasonal Index Trend ŷ Adjusted Trend

Estimate ŷ(adj)

Spring 89 37.5 33.375

Summer 162 37.5 60.75

Autumn 87 37.5 32.625

150 150

Exercise 15.20 File: X15.20 - gross domestic product.xlsx

2001 45

2002 47

2003 61

2004 64

2005 72

2006 74

2007 84

2008 81

2009 93

2010 90

2011 98

Domestic Product y = 5.2636x + 41.964

120

100

GDP (in R100 millions)

80

60

40

GDP (Millions)

20 Linear (GDP (Millions))

0

1 2 3 4 5 6 7 8 9 10 11

1 in 2001

(b) Trend line (see graph in (i)) ŷ = 41.964 + 5.2636 x x = 2 in 2002

3 in 2003

(c) Expected GDP for 2012 and 2013 (in R100 million)

2012 12 41.964 + 5.2636 (12) 105.1

2013 13 41.964 + 5.2636 (13) 110.4

Exercise 15.21 File: X15.21 - pelagic fish.xlsx

Unadj

3 Period Seasonal Seasonal

Year Months Pelagic fish Seasonal

Moving Ave ratio Index

Index

2007 Jan - Apr 44

May - Aug 36 38.00 94.74 96.85 97.79

Sept - Dec 34 38.33 88.70 88.70 89.56

2008 Jan - Apr 45 40.33 111.57 111.57 112.65

May - Aug 42 40.33 104.13

Sept - Dec 34 38.00 89.47

2009 Jan - Apr 38 34.67 109.62

May - Aug 32 32.33 98.97

Sept - Dec 27 33.00 81.82

2010 Jan - Apr 40 32.67 122.45

May - Aug 31 33.00 93.94

Sept - Dec 28 Totals 297.12 300

Seasonal

(c) Periods

Indexes

Jan - Apr 112.65

May - Aug 97.79

Sept - Dec 89.56

- significantly higher (12.65%) than the trend /cyclical volumes in the Jan - April period.

- only marginally lower (2.21%) than the trend /cyclical volumes in the May - Aug period.

- more than 10% lower than the trend /cyclical volumes in the Sept - Dec period.

(b)

(Actual vs 3-Period Moving Average)

50

45

40

35

no. of tonnes

30

25

20

15

10 Pelagic fish

5

3 Period Moving Ave

0

1 2 3 4 5 6 7 8 9 10 11 12

Comment

The trend in pelagic fish catches over the past 4 years is decidedly downwards.

(d) Period (x) Catch (y) x² xy

1 44 1 44

2 36 4 72

3 34 9 102

4 45 16 180

5 42 25 210

6 34 36 204

7 38 49 266

8 32 64 256

9 27 81 243

10 40 100 400

11 31 121 341

n = 12 12 28 144 336

Σ 78 431 650 2654

ŷ= 42.621 - 1.0315 x x= 2 in May - Aug 2007

3 in Sept - Dec 2007

Seasonally

Period Seasonal

x Trend ŷ adjusted

2011 Index

Trend

Jan - Apr 13 29.21 112.65 32.91

May - Aug 14 28.18 97.79 27.56

Sept - Dec 15 27.15 89.56 24.32

Exercise 15.22 File: X15.22 - share price.xlsx

January 1 90 = 89.428 - 2.7143(1) 86.7

February 2 82 = 89.428 - 2.7143(2) 84.0

March 3 78 = 89.428 - 2.7143(3) 81.3

April 4 80 = 89.428 - 2.7143(4) 78.6

May 5 74 = 89.428 - 2.7143(5) 75.9

June 6 76 = 89.428 - 2.7143(6) 73.1

July 7 70 = 89.428 - 2.7143(7) 70.4

Aug 8 = 89.428 - 2.7143(8) 67.7

Sept 9 = 89.428 - 2.7143(9) 65.0

Oct 10 = 89.428 - 2.7143(10) 62.3

Nov 11 = 89.428 - 2.7143(11) 59.6

1 in January

ŷ = 89.429 - 2.7143 x x = 2 in February

3 in March

y = 89.429 - 2.7143 x

100

90

price (cents)

80

70

60

Share price

50

Trend Price

40

1 2 3 4 5 6 7 8 9 10 11

Conclusion

The trend estimate of the share price is likely to fall below 60c (the selling level)

by November of the same year if the downward trend continues uninterrupted.

Exercise 15.23 File: X15.23 - Addo park.xlsx

Seaonal De-

(a) Season Actual

Index seasonalised

Su 2008 196 112 175.0

Au 2008 147 94 156.4

Wi 2008 124 88 140.9

Sp 2008 177 106 167.0

Su 2009 199 112 177.7

Au 2009 152 94 161.7

Wi 2009 132 88 150.0

Sp 2009 190 106 179.2

Su 2010 214 112 191.1

Au 2010 163 94 173.4

Wi 2010 145 88 164.8

Sp 2010 198 106 186.8

(b) and (d) Line Plots of Actual and De-Seasonalised Visitor Numbers (by Season)

(Actual vs De-Seasonalised)

250

no. of visitors (in 1000s)

200

150

100

Actual

50

De-seasonalised

0

Su Au Wi Sp Su Au Wi Sp Su Au Wi Sp

2008 2008 2008 2008 2009 2009 2009 2009 2010 2010 2010 2010

(c) Conclusion

There is a very slight upward trend in visitors to the Addo National Park. There is

almost no growth in visitors over the past 3 years.

Exercise 15.24 File: X15.24 - healthcare claims.xlsx

Uncentred Centred 4

2x4 period Unadjusted Adjusted

4 period period Seasonal

Periods Claims Moving Seasonal Seasonal

Moving Moving Ratios

Total Indexes Indexes

Total Average

2007 Q1 11.8

2007 Q2 13.2 60.5

2007 Q3 19.1 59.6 120.1 15.0125 127.23 139.67 136.28

2007 Q4 16.4 58.8 118.4 14.8 110.81 106.75 104.15

2008 Q1 10.9 62.1 120.9 15.1125 72.13 72.13 70.37

2008 Q2 12.4 63.5 125.6 15.7 78.98 91.41 89.19

2008 Q3 22.4 64.8 128.3 16.0375 139.67

2008 Q4 17.8 68.6 133.4 16.675 106.75 409.96 400

2009 Q1 12.2 70.3 138.9 17.3625 70.27

2009 Q2 16.2 67.1 137.4 17.175 94.32

2009 Q3 24.1 67.7 134.8 16.85 143.03

2009 Q4 14.6 66 133.7 16.7125 87.36

2010 Q1 12.8 62.7 128.7 16.0875 79.56

2010 Q2 14.5 64.2 126.9 15.8625 91.41

2010 Q3 20.8

2010 Q4 16.1

Q2 89.19

Q3 136.28

Q4 104.15

1 11.8 1 11.8

2 13.2 4 26.4

3 19.1 9 57.3

4 16.4 16 65.6

5 10.9 25 54.5

6 12.4 36 74.4

7 22.4 49 156.8

8 17.8 64 142.4

9 12.2 81 109.8

10 16.2 100 162

11 24.1 121 265.1

12 14.6 144 175.2

13 12.8 169 166.4

14 14.5 196 203

15 20.8 225 312

n = 16 16 16.1 256 257.6

Σ 136 255.3 1496 2240.3

b1 = (16*2240.3 - 136*255.3)/(16*1496 - 136²) = 0.2066

1 in Q1 2007

ŷ= 14.2 + 0.2066 x x= 2 in Q2 2007

3 in Q3 2007

Seasonally

Seasonal

Period x Trend ŷ adjusted

Index

Trend

Q1 2011 17 17.71 70.37 12.46

Q2 2011 18 17.92 89.19 15.98

Q3 2011 19 18.13 136.28 24.71

Q4 2011 20 18.33 104.15 19.09

Interpretation

Healthcare claims are expected to rise during 2011 from a low of

R12,46 (mill) in Q1 to R24,71 (mill) in Q3.

2007 Q1 11.8 70.37 1 14.41 10.14

2007 Q2 13.2 89.19 2 14.61 13.03

2007 Q3 19.1 136.28 3 14.82 20.20

2007 Q4 16.4 104.15 4 15.03 15.65

2008 Q1 10.9 70.37 5 15.23 10.72

2008 Q2 12.4 89.19 6 15.44 13.77

2008 Q3 22.4 136.28 7 15.65 21.32

2008 Q4 17.8 104.15 8 15.85 16.51

2009 Q1 12.2 70.37 9 16.06 11.30

2009 Q2 16.2 89.19 10 16.27 14.51

2009 Q3 24.1 136.28 11 16.47 22.45

2009 Q4 14.6 104.15 12 16.68 17.37

2010 Q1 12.8 70.37 13 16.89 11.88

2010 Q2 14.5 89.19 14 17.09 15.24

2010 Q3 20.8 136.28 15 17.30 23.58

2010 Q4 16.1 104.15 16 17.51 18.23

2011 Q1 70.37 17 17.71 12.46

2011 Q2 89.19 18 17.92 15.98

2011 Q3 136.28 19 18.13 24.70

2011 Q4 104.15 20 18.33 19.09

Healthcare Claims

(Actual vs Seasonally-adjusted Trend Estimates)

30

25

20

15

10

Exercise 15.25 File: X15.25 - financial advertising.xlsx

2005 1 9.6 1 9.6

2006 2 11.8 4 23.6

2007 3 12 9 36

2008 4 13.6 16 54.4

2009 5 14.1 25 70.5

2010 6 15 36 90

2011 7 17.8 49 124.6

Sum 28 93.9 140 408.7

n =7

b1 = (7*408.7-28*93.9)/(7*140-282) = 1.1821

b0 = (93.9-1.182143*28)/7 8.6857

1 in 2005

ŷ = 8.6857 + 1.1821 x x= 2 in 2006

3 in 2007

(c)

Financial Services Sector

Advertising Expenditure y = 1.1821x + 8.6857

20

Adspend (in R10 millions)

18

16

14

12

10

8

1 2 3 4 5 6 7

(d) Trend line equation (using Excel 's Add Trendline function)

1 in 2005

ŷ = 8.6857 + 1.1821 x x= 2 in 2006

3 in 2007

(e) Year Time (x) Actual (y) Trend (ŷ)

2005 1 9.6 9.87

2006 2 11.8 11.05

2007 3 12 12.23

2008 4 13.6 13.41

2009 5 14.1 14.60

2010 6 15 15.78

2011 7 17.8 16.96

2012 8 18.14

20

adspend (in R10 million)

18

16

14

12

Actual (y)

10 Trend (ŷ)

8

1 2 3 4 5 6 7 8

Exercise 15.26 File: X15.26 - policy surrenders.xlsx

240

220

200

no. of policies

180

160

140

120 Policies

100

2008 2008 2008 2008 2009 2009 2009 2009 2010 2010 2010 2010

Q1 Q2 Q3 Q4 Q1 Q2 Q3 Q4 Q1 Q2 Q3 Q4

Interpretation

There is a distinct moderate downward trend in the number of surrendered policies

over the period 2008 - 2010. This reduction could be due to the improved client

communication strategy adopted by the company in recent years.

(b)

Uncentred Centred 4

2x4 period Unadjusted Adjusted

Time 4 period period Seasonal

Policies Moving Seasonal Seasonal

Periods Moving Moving ratios

Total Indexes Indexes

Total Average

2008 Q1 212

2008 Q2 186 795

2008 Q3 192 769 1564 195.5 98.21 97.84 97.99

2008 Q4 205 748 1517 189.625 108.11 107.47 107.62

2009 Q1 186 725 1473 184.125 101.02 100.62 100.77

2009 Q2 165 702 1427 178.375 92.50 93.49 93.62

2009 Q3 169 685 1387 173.375 97.48

2009 Q4 182 678 1363 170.375 106.82

2010 Q1 169 671 1349 168.625 100.22

2010 Q2 158 667 1338 167.25 94.47

2010 Q3 162

2010 Q4 178 Totals 399.41 400

Q2 93.62

Q3 97.99

Q4 107.62

Interpretation

Endowment policy surrenders are highest in Q4 by 7,62% over the quarterly average;

and lowest in Q2 by 6,38% below the quarterly average. There is very little - to none -

seasonal impact on policy surrenders in Q1 and only 2,01% below the quarterly

average in Q3.

(c) Period (x) Cost (y) x² xy

1 212 1 212

2 186 4 372

3 192 9 576

4 205 16 820

5 186 25 930

6 165 36 990

7 169 49 1183

8 182 64 1456

9 169 81 1521

10 158 100 1580

11 162 121 1782

n = 12 12 178 144 2136

Σ 78 2164 650 13558

1 in Q1 2008

ŷ= 203.42 - 3.5524 x x= 2 in Q2 2008

3 in Q3 2008

Seasonally

Seasonal

Period x Trend ŷ adjusted (Rounded)

Index

Trend

Q1 2011 13 157.24 100.77 158.45 158

Q2 2011 14 153.69 93.62 143.88 144

Q3 2011 15 150.13 97.99 147.12 147

Q4 2011 16 146.58 107.62 157.75 158

Exercise 15.27 File: X15.27 - company liquidations.xlsx

(a)

Period Liquidations 3-per M A 5-per M A

1 246

2 243 252.7

3 269 289.7 255.6

4 357 263.0 237.2

5 163 224.7 210.4

6 154 142.0 189.0

7 109 141.7 162.0

8 162 164.3 184.0

9 222 219.0 210.0

10 273 259.7 249.2

11 284 287.3 275.4

12 305 294.0 300.6

13 293 315.3 330.6

14 348 354.7 332.0

15 423 354.0 335.0

16 291 344.7 327.0

17 320 288.0 304.2

18 253 269.0 252.0

19 234 216.3 241.8

20 162 212.0 237.4

21 240 233.3 239.6

22 298 267.3 243.4

23 264 271.7 269.6

24 253 270.0 282.0

25 293 282.7 260.0

26 302 261.0

27 188

(b)

Company Liquidations

(Actual vs 3 and 5 period Moving Averages)

450

400

350

no. of liquidations

300

250

200

150

100

50

Liquidations 3-per M A 5-per M A

0

1 3 5 7 9 11 13 15 17 19 21 23 25 27

periods

(c) Interpretation

The level of business liquidations shows no actual upward / downward trend over the

past 27 periods. There is a distinct cyclical pattern wih the longest lasting from

period 7 to period 20.

Exercise 15.28 File: X15.28 - passenger tyres.xlsx

Centred 4

Uncentred 4 Unadjusted Adjusted

Time 2x4 period period Seasonal

Tyre Sales period Seasonal Seasonal

Periods Moving Total Moving ratios

Moving Total Indexes Indexes

Average

2005 Q1 64876

2005 Q2 58987 240829

2005 Q3 54621 244699 485528 60691.0 90.00 90.64 90.72

2005 Q4 62345 252285 496984 62123.0 100.36 100.81 100.91

2006 Q1 68746 258591 510876 63859.5 107.65 107.65 107.76

2006 Q2 66573 267480 526071 65758.9 101.24 100.51 100.61

2006 Q3 60927 277522 545002 68125.3 89.43

2006 Q4 71234 282186 559708 69963.5 101.82

2007 Q1 78788 289357 571543 71442.9 110.28

2007 Q2 71237 292567 581924 72740.5 97.93

2007 Q3 68098 291438 584005 73000.6 93.28

2007 Q4 74444 296653 588091 73511.4 101.27 399.62 400.00

2008 Q1 77659 302011 598664 74833.0 103.78

2008 Q2 76452 306475 608486 76060.8 100.51

2008 Q3 73456 313379 619854 77481.8 94.80

2008 Q4 78908 318170 631549 78943.6 99.95

2009 Q1 84563 319592 637762 79720.3 106.07

2009 Q2 81243 327440 647032 80879.0 100.45

2009 Q3 74878 334433 661873 82734.1 90.50

2009 Q4 86756 338248 672681 84085.1 103.18

2010 Q1 91556 340405 678653 84831.6 107.93

2010 Q2 85058 333794 674199 84274.9 100.93

2010 Q3 77035 345161 678955 84869.4 90.77

2010 Q4 80145 356559 701720 87715.0 91.37

2011 Q1 102923

2011 Q2 96456

Q2 100.61

Q3 90.72

Q4 100.91

Passenger Car Tyre Sales y = 1302x + 58114

120000

100000

80000

60000

1 in Q1 2005

(b) ŷ = 58114 + 1302 x x = 2 in Q2 2005

3 in Q3 2005

(b) Seasonally-adjusted Trend estimates of passenger car tyre sales (in units) for 2012 / 2013.

Seasonally

Time Seasonal Adjusted

Time Trend (ŷ)

Periods Indices Trend

Estimate

2012 Q1 29 95872 107.76 103312

2012 Q2 30 97174 100.61 97767

2012 Q3 31 98476 90.72 89337

2012 Q4 32 99778 100.91 100686

2013 Q1 33 101080 107.76 108924

2013 Q2 34 102382 100.61 103007

2013 Q3 35 103684 90.72 94062

2013 Q4 36 104986 100.91 105941

(c) Interpretation

The pattern of passenger car tyre sales is very stable.

The trend is linear and upward, and seasonal variations are consistent over time.

Hence Hillstone management could have high confidence in the estimates.

Exercise 15.29 File: X15.29 - outpatient attendances.xlsx

Uncentred Centred 4

2x4 period Unadjusted Adjusted

Time 4 period period Seasonal

Visits Moving Seasonal Seasonal

Periods Moving Moving ratios

Total Indexes Indexes

Total Average

2006 Q1 12767

2006 Q2 16389 64041

2006 Q3 19105 65472 129513 16189.1 118.01 117.31 117.46

2006 Q4 15780 68951 134423 16802.9 93.91 92.62 92.75

2007 Q1 14198 70745 139696 17462.0 81.31 78.99 79.10

2007 Q2 19868 73269 144014 18001.8 110.37 110.53 110.69

2007 Q3 20899 73712 146981 18372.6 113.75

2007 Q4 18304 74048 147760 18470.0 99.10

2008 Q1 14641 74227 148275 18534.4 78.99

2008 Q2 20204 72000 146227 18278.4 110.53

2008 Q3 21078 71434 143434 17929.3 117.56

2008 Q4 16077 72489 143923 17990.4 89.36 399.46 400.00

2009 Q1 14075 72378 144867 18108.4 77.73

2009 Q2 21259 72484 144862 18107.8 117.40

2009 Q3 20967 70821 143305 17913.1 117.05

2009 Q4 16183 71386 142207 17775.9 91.04

2010 Q1 12412 72569 143955 17994.4 68.98

2010 Q2 21824 74365 146934 18366.8 118.82

2010 Q3 22150 79370 153735 19216.9 115.26

2010 Q4 17979 78114 157484 19685.5 91.33

2011 Q1 17417 80274 158388 19798.5 87.97

2011 Q2 20568 85413 165687 20710.9 99.31

2011 Q3 24310 67996 153409 19176.1 126.77

2011 Q4 23118 47428 115424 14428.0 160.23

Q2 110.69

Q3 117.46

Q4 92.75

Outpatient Visits

y = 224.66x + 15591

30000

25000

no. of patients

20000

15000

10000

Visits

5000

Linear (Visits)

0

1 3 5 7 9 11 13 15 17 19 21 23

1 in Q1 2006

ŷ = 15591 + 224.66 x x = 2 in Q2 2006

3 in Q3 2006

for 2012 / 2013 (first half).

Seasonally

Time Seasonal Adjusted

Time Trend (ŷ)

Periods Indices Trend

Estimate

2012 Q1 25 21207.5 79.10 16775

2012 Q2 26 21432.2 110.69 23723

2012 Q3 27 21656.8 117.46 25438

2012 Q4 28 21881.5 92.75 20295

2013 Q1 29 22106.1 79.10 17486

2013 Q2 30 22330.8 110.69 24718

(c) Interpretation

The pattern of outpatient attendances at the Butterworth Clinic is very stable.

It shows a steady upward trend with highly consistent seasonal variations.

Demand increases in the winter months ((Q3) and is lowest in the summer months (Q1).

Exercise 15.30 File: X15.30 - construction absenteeism.xlsx

Construction Absenteeism

Unadjusted Adjusted

Time period period period Seasonal

Days_lost Seasonal Seasonal

Periods Moving Moving Moving ratios

Indexes Indexes

Total Total Average

2006 Q1 933

2006 Q2 865 3584

2006 Q3 922 3618 7202 900.3 102.42 102.74 102.88

2006 Q4 864 3689 7307 913.4 94.59 95.80 95.93

2007 Q1 967 3698 7387 923.4 104.72 104.07 104.21

2007 Q2 936 3736 7434 929.3 100.73 96.85 96.98

2007 Q3 931 3661 7397 924.6 100.69

2007 Q4 902 3570 7231 903.9 99.79

2008 Q1 892 3546 7116 889.5 100.28

2008 Q2 845 3445 6991 873.9 96.70

2008 Q3 907 3368 6813 851.6 106.50

2008 Q4 801 3238 6606 825.8 97.00 399.45 400.00

2009 Q1 815 3110 6348 793.5 102.71

2009 Q2 715 3020 6130 766.3 93.31

2009 Q3 779 3027 6047 755.9 103.06

2009 Q4 711 3168 6195 774.4 91.82

2010 Q1 822 3151 6319 789.9 104.07

2010 Q2 856 3162 6313 789.1 108.47

2010 Q3 762 3125 6287 785.9 96.96

2010 Q4 722 2984 6109 763.6 94.55

2011 Q1 785 2962 5946 743.3 105.62

2011 Q2 715 2944 5906 738.3 96.85

2011 Q3 740 2159 5103 637.9 116.01

2011 Q4 704 1444 3603 450.4 156.31

Q2 96.98

Q3 102.88

Q4 95.93

y = -9.917x + 952.75

1000

900

no. of days lsot

800

700

600

Days_lost

500

Linear (Days_lost)

400

1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24

1 in Q1 2006

ŷ = 952.75 - 9.917 x x = 2 in Q2 2006

3 in Q3 2006

(b) Seasonally-adjusted Trend estimates of Days Lost in Construction Industry for 2012.

Seasonally

Time Seasonal Adjusted

Time Trend (ŷ)

Periods Indices Trend

Estimate

2012 Q1 25 704.8 104.21 734

2012 Q2 26 694.9 96.98 674

2012 Q3 27 685.0 102.88 705

2012 Q4 28 675.1 95.93 648

(c) Interpretation

The pattern of days lost due to absenteeism in the Construction industry shows

a distinct downward trend but with inconsistent seasonal variations.

CHAPTER 16

FINANCIAL CALCULATIONS

Interest is computed on the original lump sum for each period.

Compound interest

For each period, interest is computed on the original lump sum

plus all accummulated interest of the preceeding periods.

simple interest investment.

capitalised sooner - and therefore earning more interest than

an annual compounded investment.

Exercise 16.4 Nominal interest rate is the quoted per annum interest rate

Effective interest rate is the actual interest rate achieved

when interest is compounded more than once per year.

(or received) at regular intervals over a period of time.

Exercise 16.6 Ordinary annuity - regular payments begin the first period of the annuity term

Deferred annuity - regular payments begin only at some future period into

the term of the annuity.

Exercise 16.7 Ordinary annuity certain - the series of regular payments take place at the

end of each payment period.

Ordinary annuity due - the series of regular payments take place at the

beginning of each payment period.

Exercise 16.8 NPV is the term used to convert all cash inflows (and outflows) over time

into present value terms by dividing by the annual rate of interest

It represents future cash flows in current terms.

Exercise 16.9

Exercise 16.10

Exercise 16.11

(or 7.003 years)

Exercise 16.12

Exercise 16.13

Exercise 16.14

Exercise 16.15

n = 8.0013 years

Exercise 16.16

Fv = 2000*(1+0.10/2)0.5 = R 2,049.39

Fv = 2049.39*(1+0.12/12)21 = R 2,525.65

Exercise 16.17

Exercise 16.18

Exercise 16.19

Pv = 25000/(1+0.09/4)11 = R 19,572.37

Exercise 16.20

n = log(30000/21353.4)/log(1+0.12) = 3 years

Exercise 16.21

Let Pv = R1 and Fv = R2

Exercise 16.22

Exercise 16.23

Simple Interest Fv2 = Pv*(1+0.07*9)

Pv = 334.16/((1+0.07)9 - (1+0.07*9)) = 1602.999

Exercise 16.24

R= 89989.12/(((1+0.09/12)(3*12) - 1)/(0.09/12))

R 2,186.71

R= 89989.12/(((1+0.09/12)(3*12) - 1)*(1+0.09/12)/(0.09/12))

R 2,170.43

Exercise 16.25

Exercise 16.26

Pv = 8750*((1-(1+0.1/12)^(-(5*12)))/(0.1/12)) = R 411,821.98

Exercise 16.27

Exercise 16.28

(2*12)

Fv1 (OAC) = 540*((1+0.12/12) -1)/(0.12/12) = R14 565.67

Then compute Fv on the capital sum after 2 years until maturity (for 7 years).

(7*12)

Fv2 (OAC) = 750*((1+0.12/12) -1)/(0.12/12) = R98 004.21

Exercise 16.29

Exercise 16.30

Deposit = 20000 R20 000,00

Total Purchase Price of Motor Vehicle = R108 406.52

= R108 406.52

Exercise 16.31

FV1 (1 year) =25933.19*(1+0.1/12)(1*12) = R28 648.73

(1*12)

Fv2 (1 year) = 1000*((1+0.10/12) -1)/(0.10/12) = R12 565.57

12565.56809

Maturity value after 3 years R28 648.73 + R12 565.57 = R41 214.30

Exercise 16.32

(a) Months 1 - 5

FV1 = 200*((1+0.12/12)5-1)/(0.12/12) = R1020.20

Withdrawal R 300.00

Balance R1020.20 - R300.00 = R 720.20

7

Compound Interest 720.2*(1+0.12/12) = R 772.15

Months 6 - 10

5

FV2 = 200*((1+0.12/12) -1)/(0.12/12) = R1020.20

Withdrawal R 300.00

Balance R1020.20 - R300.00 = R 720.20

Compound Interest 720.2*(1+0.12/12)2 = R 734.68

Months 11, 12

2

FV3 = 200*((1+0.12/12) -1)/(0.12/12) = R 402.00

R772.15 + R734.68 + R402.00 = R1908.83

7

=(-(FV(0.12/12,5,200,,0))-300)*(1+0.12/12) R 772.15

2

=(-(FV(0.12/12,5,200,,0))-300)*(1+0.12/12) R 734.68

=(-FV(0.01,2,200,,0)) R 402.00

R772.15 + R734.68 + R402.00 = R1908.83

Exercise 16.33

R= 26000/((1-(1+0.14/4)(-3*4))/(0.14/4)) = R2 690.58

R= 26000/(((1-(1+0.14/4)(-3*4))*(1+0.14/4))/(0.14/4)) = R2 599.60

(c) The student should select to repay at the beginning of every month.

Total repayment will be less than repaying at the end of every month.

Exercise 16.34

Present Value (Pv) of a Deferred Annuity

(-(36))

Factor 2 =(1-(1+0.16/12) )/(0.16/12) 28.44381

R= 30000/12.6779 = R2 366.32

Exercise 16.35

Exercise 16.36

n = LOG(1.51879)/LOG(1+0.12/12) = 42 months

The house owner will take 3 years and 6 months to save R103 757.98.

Exercise 16.37

INVESTMENT OPTIONS

Trucking Laundry

Initial investment (R) 60000 60000

Annual Cash Flow (R)

year 1 32000 0

year 2 38500 7500

year 3 26000 45000

year 4 13000 37500

year 5 9500 55500

1 28571 0

2 30692 5979

3 18506 32030

4 8262 23832

5 5391 31492

91422 93333

- Linking State Education Standards to Your School GardenUploaded byfriends of the Native Plant Society of New Mexico
- Chapter1 Introduction to StatisticsUploaded byMarie Angeline Manzano
- Chapter_08_revisesd0.pptUploaded byMeth
- Probability and statistic slidsUploaded bypankaj kumar
- StatisticsUploaded byJonny Beat
- Project Synopsis Template for MbaUploaded bySushanth
- Chap007 Sampling aaUploaded byThiện Thảo
- BAJAJUploaded bybuddysmbd
- Unit I - The Nature of Probability and StatisticsUploaded byKevin Bono
- Young Chinese Consumer´s Brand Perceptions towards Global BrandsUploaded byJelena Zuta Obradovic
- NTC ISO 9001 - 2015Uploaded bydjfredito
- A CQMS 202 Course Outline W2012-1Uploaded bycliao123
- Project Management QuizUploaded byNikky GS
- Fyp Present NazreUploaded byMohdNazreMdNajib
- Prevalence Survey4[1]Uploaded bysanjivdas
- A03 - IndexUploaded byTom
- Research PaperUploaded bySamuel Krypton
- Decision MakingUploaded byJacob Gibbs
- rsUploaded byapi-25887578
- Example of Research CHAPTER 3Uploaded byJoan Eser Mina
- Lesson 1 Quantitative Research and VariablesUploaded byBenedict Manaog
- A Marketing ReportUploaded byHarsh Patel
- Solomon BUploaded byaganapathy1
- How to Get Basic Statistics About Your Survey Response Data in LimeSurvey_2Uploaded byHappy Deal
- SIP Report Guidelines and Certificates-2Uploaded bysidhartha jain
- Metodologías de MuestreoUploaded byGabrielSantelices
- Writing Reports Based on Statistical DataUploaded bySahil Sethi
- reflection1Uploaded byapi-242194552
- 02 Scientific Invest Reprtg001Uploaded byOriginaltitle
- Excel GyanUploaded bygsethuanandh

- Kumon Simple AdditionUploaded bybooksdocs
- Diversity Management in South African CompaniesUploaded bymzansigirl
- an_overview_of_the_plastic_material_selection_process_for_medical_devices (1).pdfUploaded bycatiav5r18
- Cambridge Roadmapping Rob Phaal NewUploaded bymzansigirl
- How to Read a Financial Report1Uploaded byGeorge Kuruvilla
- GSMA Decision Support Toolkit SA _FINALUploaded bymzansigirl
- wfun15_birds_lowercase_letter_tracing_1.pdfUploaded bymzansigirl
- 9781137430359_sampleUploaded bymzansigirl
- GIBS Case Studies Booklet 2017Uploaded bymzansigirl
- Google's HR DilemmaUploaded bymzansigirl

- 4.1 Sampling and SurveysUploaded byStephen Pak
- Sampling Report PptUploaded byIonacer Viper
- Bhartiya International Ltd. Recruitment & Selection FinalUploaded byAvnindra Gangwar
- Thermo HWUploaded byIvalinov Kolev
- two wheeler (1).docxUploaded byDeepak Baliyan
- TATA NANO 1 (1)Uploaded byniteshomray25
- Chapter 3Uploaded byrajiram87
- 1. Introduction to Statistics and 2. Sampling TechniquesUploaded byLargil Masayon Taojo
- 06 BRM PowerPointUploaded bydilpreet92
- L1 StatisticsUploaded byMagcuyao Dan Andrew
- Quantitative Techniques_Case StudyUploaded byArathi Deepak Nair
- hypercity reseach project.docxUploaded byAnamika Singh
- AzizUploaded byazimeraw mosneh
- monica martell-1040 project parts 12Uploaded byapi-242261365
- Big BazaarUploaded byamriteshbhai
- CHAPTER (1)Uploaded byVinoth Kumar
- Summary of the Research SessionsUploaded byAbdiqadir Jibril
- Consumer Perception of Soft DrinksUploaded byViral Patel
- Research MethodologyUploaded bysamwel
- Apostila PAPPUploaded byAnderson Roberto
- Chapter 1 Descriptive Statistcs 1_2012Uploaded byjollychang_1
- probability and statisticsUploaded byarpitsingh063
- Caribbean StudiesUploaded byRenaisha Hood
- Asad ProjectUploaded byAsadullah Rahim
- ICICI LAMBORDUploaded byKing Nitin Agnihotri
- Sampling Methods PptUploaded byJasim Shaikh
- How to Design and Evaluate Research in Education(1)Uploaded byJuli Briana
- HIL organiUploaded byanju
- Research MethodologyUploaded byRavi Hooda
- No Survey Can Achieve Success Without a Wel1Uploaded byYuni Rahmi