Welcome to

1
Welcome to
Statistical Analysis in Business
(Math 1029)
2
Why study Statistics?

• To become a better consumer of other peoples
  data
• To facilitate communication
• To improve computer skills
• To overcome either too little or too much
  information
• To develop technical literacy
• To improve career mobility

3
Statistical Analysis in Business

• Course Coordinator
• Malgorzata Korolkiewicz Y3-72 8302 0237
• Lecturers
• Deborah Downes OC1-64 8302 5781
• Jim Gaetjens Y3-69 8302 0525
• Susan Kim Y3-78 TBA
• Zen Lu Y3-66 8302 0648
• School Administration
• Hayley Timms Y3-76 8302 0525

4
Lecture times for S.A.B.
5
Computer Practical Lab times

• Register for a PC Lab session using Enrol Online
  (Medici).
• These sessions are an introduction to Minitab,
  the Statistical computing package used in this
  course.
• PC Labs are at times different to tutorial times,
  and in week 1 only.

6
Lecture Overheads and Textbook

• Lecture overheads posted on the S.A.B. website
  before the week in which lectures are given,
  without lecture exercise solutions.
• Lecture exercise solutions will not be posted.
• Textbook Brown, B.M., Brien, C.J. Lu, Z-H.
  (2003), Statistical Analysis in Business,
  Prentice-Hall, Sydney.
• Note the textbook comes with the Student
  edition of Minitab.

7
Tutorials

• Held in weeks 2-13. Please bring your calculators
  and textbook to every tutorial.
• Activities include revision of tutorial questions
  (may be different from the review exercises in
  Blue booklet) and class discussion
• Tutorial questions posted on the web, solutions
  available at the end of the week following the
  tutorial
• Homework due in weeks 2-4, tutorial assessable
  tests held in weeks 6-12, trial test in week 5
• Solutions available on the web the following
  fortnight

8
Assessment

• Final exam is worth 68 of overall grade with a
  minimum mark of 40 required to pass this course.
• Individual assignment is worth 12 of overall
  grade (handed out in Week 6 and to be handed in
  on the Monday of Week 10)
• Homeworks and tutorial tests 20 of overall grade
• Exam 68
• Assignment 12
• H/work test 20
• Total 100
• Note See the Blue booklet (Study guide) for full
  details

9
Help Desk

• City West
• Room Y3-73 (Yungondi building, level 3, room 73)
• Times See SAB website for further details. URL
  http//www.unisanet.unisa.edu.au/10688

10
Graduate qualities

• Body of knowledge
• Lifelong learning
• Ethical action
• Problem solving
• Work alone or in teams
• Communicate effectively
• International perspective
• The University has identified these seven
  graduate qualities, to be promoted throughout all
  courses

11
The first step ...

• Read carefully pages 3 and 4 of the Blue booklet
  (study guide)
• Check the section What needs to be done now on
  page 4
• Commence a program of regular reading, study and
  preparation work

12
Regular work and study is the key to success

• It can never be emphasised strongly enough how
  true this is for Statistical Analysis in Business
• This course rewards those students who put
  constant effort into it over the semester

13
Week 1 objectives

• 1. Course overview
• 2. Data types
• Sample and population
• Sampling methods
• 5. Historical records
• 6. Scenario questions
• Types of survey questions
• 8. Types of variables

14
Please note

• The material in Week 1 is mainly descriptive, and
  is not difficult to understand
• But it is essential to understand the main ideas,
  because these ideas arise later in the course,
  when the validity is assessed of formal methods
  of Statistical Inference
• Particularly this applies to Sampling Methods

15
1.Course Overview major categories of
statistical procedures
()
() Including interval estimation
16
Major categories of statistical procedures,
continued

• What was the Australian market share for major
  credit and debit cards in the first quarter of
  2005?
• This question can be answered using descriptive
  summaries

17
Major categories of statistical procedures,
continued

• A major car manufacturer wants to know the
  average cost of engine warranty claims on a new
  hybrid engine.
• They have collected warranty cost information on
  engines produced so far.
• The procedure required here is interval estimation

18
Major categories of statistical procedures,
continued

• A shipment of 200 DVD players reveals 4 with
  defects. The suppliers historical defect rate is
  0.005. Has the defect rate really changed or is
  it simply a bad batch?
• This question can be answered using hypothesis
  testing

19
2. Types of Data

• Nature of the data source either observational
  study or controlled experiment
• Type of data either primary or secondary
• Data frame either a census or a sample

20
3. Sample vs. Population

• What are the common sampling methods?
• How likely is each method to produce a sample
  which is representative of the parent
  population?
• In examples, can we identify the sampling method
  used, and the
• Target population? observational units?
  sampling units?

21
4. Common sampling methods

• Census, convenience sampling, judgement sampling,
  self-selected samples, simple random sampling,
  systematic sampling, stratified sampling,
  cluster sampling, quota sampling, historical
  records.
• Do you know what each of these terms means? Can
  you give an example of each type?
• the standard good method
• not examinable

22
Sampling methods (I)

• Census Complete enumeration
• Convenience Convenient
• Judgement By expert
• Self-selected eg. Phone-in, mail-back
• Simple random all units are equally likely to
  be chosen, all samples are equi-probable
• Systematic equi-spaced in list, random start

23
Sampling methods (II)

• Stratified population divided into strata
• Cluster population divided into clusters
• Quota see the textbook

24
Stratified sampling

• Divide population into groups
• Select a simple random sample from each group
• Groups are called strata

25
What is the difference between stratified and
cluster sampling?

• ALL strata sampled all groups
• Desirable that homogeneous (eg suburbs), sample
  within each stratum
• A sample of clusters taken some groups
• Desirable that each cluster heterogeneous (eg
  households)

26
How to choose a random sample?

• Use coins, dice, birthdates or other random
  devices
• Use a computer. With Minitab, select
• Calc gt RandomData gt Integer
• or
• Calc gt Random Data gt Sample from columns and
  choose the appropriate settings

27
Lecture exercise 1

• What types of sampling methods are these ?
• Official attendance figure at a football game
• Opinion Poll on voting preference in a uniform
  constituency
• Proportion of professional association members in
  favour of a constitutional change, as estimated
  by responses to a mail-out questionnaire
• Focus group (pre-election opinion polls,
  workplace consultation etc)
• Australian Bureau of Statistics estimate of
  national average weekly earnings

28
5. Historical records

• Not a true sample, but
• often the immediate past can be representative
  of whats most likely to happen in the near
  future, so
• the study of Time Series is important

29
6. Scenario questions

• What is the question to be investigated
  (including the group to which the conclusions
  apply)?
• What are the variables in the question to be
  investigated?
• What types of variables are they?
• Which major category of statistical procedures
  applies here? Why?
• Which particular statistical procedure(s) and/or
  diagram(s) should be used?

30
7. Classification of Survey Questions
31
Lecture exercise 2

• What question types are the following?
• (a) Who is your preferred telephone service
  provider?
• ? Telstra ? Optus ? AAPT ? Other
• (b) If you were to change your telephone service
  provider, what would be your main reason?

32
Lecture exercise 3

• What question types within closed questions
  are the following?
• Tick the box that most accurately describes the
  level of service you receive from this
  organisation
• ? courteous ? efficient ? lacking in
  care
• ? perfunctory ? non-existent ? friendly

33
Lecture exercise 3 (contd)

• Using the same boxes as in question 1, tick the
  boxes whose description applies to the level of
  service that you receive.
• "The service I receive from this organization is
  excellent". Categorize your response to this
  statement.
• ? strongly agree ? agree ? no opinion
• ? disagree ? strongly disagree

34
Variables

• Questions of interest can be investigated by
  measuring characteristics of individuals within
  the target population.
• These characteristics are called variables.
• The data for statistical analysis is always a
  collection of observed outcomes.
• For example, if the question to be investigated
  concerned the average family income in Adelaide,
  what variable(s) should be observed?

35
8. Types of variables
36
Binary variables are a special type of
categorical variable

• Variables which can take on only two possible
  values are called binary
• Binary variables are unique in that they can be
  regarded as nominal or ordinal

37
Lecture exercise 4

• What variable types are these?
• (nominal, ordinal, binary, numerical/discrete or
  numerical/continuous)
• The five subject choices of high school students
  in a term
• The golf scores of two players
• The annual income levels (to nearest 000) for
  families
• A classification of agricultural areas as
  tropical, temperate or cool
• The recording of satellite launches as successful
  or not
• Business profit or loss figures for the last nine
  quarters
• A recording of whether or not customers have been
  with a bank for longer than two years.