BUSINESS STATISTICS

An Applied Approach
Chapter 1
Introduction
• “I hear and I forget.

• I see and I remember.

• I do and I    understand."                       
(Confucius)
 Why a Manager needs to Know
about Statistics
• Why? Because you want to know :
– Best use of imperfect information:
• e.g., 50,000 customers, 1,600 workers, 386,000
transactions,
– How to prepare ,present and describe
information .
– Good decisions in uncertain conditions:
 Why a Manager needs to Know
about Statistics
– How to draw conclusions about large populations
based on only information obtained from samples
• e.g., new product launch: Fail? OK? Make you rich?
•
– Competitive Edge
• e.g., for you in the job market!
– How to obtain reliable forecasts
 Why a Manager needs to Know
about Statistics
• Simply speaking the subject of statistics is

• The Art and Science of Collecting, Summarizing

,Interpreting and thus Understanding DATA:
– DATA = Recorded Information
• e.g., Sales, Productivity, Quality, Costs, Return,
Patients,Drugs etc
 Descriptive Statistics
• Descriptive statistics – summaries of data
presented in tabular, graphical, and
numerical forms that are easy for the reader
to understand.
• Most of newspapers, magazines, white
papers and reports use descriptive statistics.
 Example: Diagnostic Centre
• The manager of a Diagnostic Centre would
like to have a better understanding of the
chemicals used in the blood tests
performed in the centre. She examines 25
customer blood samples for tests. The
costs of chemical, rounded to the nearest
Rupee, are listed on the next slide
95 35 65 40 50

65 75 30 35 40

55 60 25 70 65

35 90 65 45 45

75 65 35 40 55
Chemical cost5- Frequency Freq. percentage

25-40

40-55

55-70

70-85

85-100
 Statistical Inference
• Population --he set of all elements of interest
in a particular study
• Sample -- a subset of the population
• Statistical inference --he process of using data
obtained from a sample to make estimates
and test hypotheses about the
characteristics of a population
• Census-- collecting data for a population
 Process of Statistical Inference
• 1. Population consists of all chemical costs for all diagnosis. Average
cost of chemical is unknown.
unknown
• A sample of 25 chemical costs is examined.
• The sample data provide a sample average chemical cost of Rs per
test .
• The sample average is used to estimate the population average.
• As an example ----a study was conducted l
comparing four groups of patients being treated for
their stomach ulcers under the following regime:
• Group 1 received placebo (an inert compound)
• Group 2 received active drug at dose d, twice per
day
• Group 3 received active drug at dose 2d, once per
day
• Group 4 received active drug at dose 2d, twice per
day
• So there were interesting comparisons to be
made particularly between a dose of 2d once
or twice a day and between a total daily dose
2d taken in one or two ‘shots’ per day. 500
patients were to be recruited into the study
across
• Table 1. Crude Cure Rates.
• Treatment Cure Rates
• Placebo 98/124 79%
• d twice/day 107/126 85%
• 2d once/day 106/130 82%
• 2d twice/day 115/130 88%
 How to approach
• Designing the study:
– First step

• Plan for data-gathering

• Random sample (control bias and error)
• Exploring the data:
• First step (once you have data)
• Look at, describe, summarize the data
• Are you on the right track?
 How to approach
• Estimating an unknown:
– Best “guess” based on data
– Wrong - buy by how much?
– Confidence interval - “we’re 95% sure that the unknown is
between …”
• 4. Hypothesis testing:
– Data decide between two possibilities
– Does “it” really work? [or is “it” just randomly better?]
– Is financial statement correct? [or is error material?]
– Whiter, brighter wash?
 Statistical View of the World

• Data are imperfect

– We do the best we can -- Statistics helps!
• Events are random
– Can’t be right 100% of the time
• Use statistical methods
– Along with common sense and good judgment
• Be skeptical!
– Statistics can be used to support contradictory
conclusions
– Look at who funded the study?
Statistics in Business: Examples

• Advertising
– Effective? Which commercial? Which markets?
• Quality control
– Defect rate? Cost? Are improvements working?
• Finance
– Risk - How high? How to control? At what cost?
Statistics in Business: Examples
• Accounting
– Audit to check financial statements. Is error material?
• Marketing
• Consumer behaviour study , Test marketing
• Other
• Economic forecasting, background
information, measuring and controlling
productivity (human and machine), …
 Data Structures: Classifying the
Various Types of Data Sets
• Data Set:
– Measurements of items
• e.g., Yearly sales volume for your 23 salespeople
• e.g., Cost and number produced, daily, for the past month
• Elementary Units:
– The items being measured
• e.g., Salespeople, Days, Companies, Catalogs, …
• A Variable:
– The type of measurement being done
• e.g., Sales volume, Cost, Productivity, Number of defects, …
 How Many Variables?

• Univariate data set: One variable measured for

each elementary unit
– e.g., Sales for the top 30 computer companies.
– Can do: Typical summary, diversity, special features
• Bivariate data set: Two variables
– e.g., Sales and # Employees for top 30 computer firms
– Can also do: relationship, prediction
• Multivariate data set: Three or more variables
– e.g., Sales, # Employees, Inventories, Profits, …
– Can also do: predict one from all other variables
 Numbers or Categories?

• Quantitative Variable: Meaningful numbers

– e.g., Sales, # Employees
– Can add, rank, count
• Qualitative Variable: Categories
– Ordinal Variable: Categories with meaningful ordering
• e.g., Bond rating (AA, A, B, …), Gold (24,22, …)
• Can rank, count
– Nominal Variable: categories without meaningful ordering
• e.g., State, Type of business, Field of study

• Can count
 Attributes
• Certain description about any object
• e.g. Defective ,Non-defective
• B.P.High, B.P. Low
 Time-Series or Cross-Sectional?

• Time-Series Data: Data values recorded in

meaningful sequence
– Elementary units might be days or quarters or years
– e.g., Daily Dow-Jones stock market average close for
the past 90 days
– e.g., Your firm’s quarterly sales over the past 5 years
• Cross-Sectional Data: No meaningful sequence
– e.g., Sales of 30 companies
– e.g., Productivity of each sales division
– Easier than time series!
 Data Presentation
• Histogram
• Relative frequency Histogram
• Skewness
• Kurtosis
• Pie chart
• Bar Diagram
• Ogive
• Line Chart
 Histogram and Bar Chart

• Histogram is a bar chart of the frequencies of the

data
– Histogram: bar height represents number of cases within
the range
– Ordinary bar chart: bar height represents data value for
just one case
• Histogram shows overall distribution
– Histogram: the “big picture” of patterns in the data
– Ordinary bar chart: often too much detail (each individual
case)
Histogram Example
Frequency Histogram
Histogram Example
Relative Frequency Histogram
• Normal
– Symmetric
– Bell-Shaped
• Skewed
– Not symmetric
– Can cause trouble
Skewness
Skewed to left
Skewness
Symmetric
Skewness
Skewed to right
 Kurtosis
Platykurtic - flat distribution
 Kurtosis
Mesokurtic - not too flat and not too peaked
Pie Chart
 Bar Chart
Fig. 1-11  Airline Operating Expenses and Revenues

12
Average Revenues

10
Average Expenses

American
Continental
DeltaNorthwest
Southwest
United USAir
A i r li n e
 Frequency Polygon and Ogive

Relative Frequency Polygon Ogive

Cumulative Relative Frequency

0.3 1.0
Relative Frequency

0.2

0.5

0.1

0.0 0.0

0 10 20 30 40 50 0 10 20 30 40 50
Sales Sales
 Time Plot
M o nthly S te e l P ro d uc tio n
(P ro b le m 1 -4 6 )
8.5
Millions of Tons

7.5

6.5

5.5

Month J F M A M J J A S ON D J F M A M J J A S ON D J F M A M J J A S O