15/09/2022

Chapter 4: Sampling and

Estimation

Statistics, Data Analysis, and

Decision Modeling, Fifth Edition
James R. Evans

Copyright ©2013 Pearson Education, Inc. publishing as Prentice Hall

4-1

Need for Sampling

 Very large populations
 Destructive testing
 Continuous production process

The objective of sampling is to draw a valid inference about a

population.

Copyright ©2013 Pearson Education, Inc. publishing as Prentice Hall

4-2

1
 15/09/2022

Sample Design
 Sampling Plan – a description of the
approach that will be used to obtain
samples from a population
 Objectives
 Target population
 Population frame
 Method of sampling
 Operational procedures for data collection
 Statistical tools for analysis
Copyright ©2013 Pearson Education, Inc. publishing as Prentice Hall
4-3

Sampling Methods
 Subjective
 Judgment sampling
 Convenience sampling
 Probabilistic
 Simple random sampling – every subset of a
given size has an equal chance of being
selected

Copyright ©2013 Pearson Education, Inc. publishing as Prentice Hall

4-4

2
 15/09/2022

Other Sampling Methods

 Systematic sampling
 Stratified sampling
 Cluster sampling
 Sampling from a continuous process

Copyright ©2013 Pearson Education, Inc. publishing as Prentice Hall

4-5

Errors in Sampling
 Nonsampling error
 Poor sample design
 Sampling (statistical) error
 Depends on sample size
 Tradeoff between cost of sampling and
accuracy of estimates obtained by
sampling

Copyright ©2013 Pearson Education, Inc. publishing as Prentice Hall

4-6

3
 15/09/2022

Random Sampling from

Probability Distributions
Random number - uniformly distributed
between 0 and 1
 Excel function RAND( )

Copyright ©2013 Pearson Education, Inc. publishing as Prentice Hall

4-7

Sampling from Discrete

Probability Distributions
 Rolling two dice

Select random number, compare to interval, assign outcome

Copyright ©2013 Pearson Education, Inc. publishing as Prentice Hall
4-8

4
 15/09/2022

Sampling from Common

Probability Distributions
 Random variate - a value randomly generated
from a specified probability distribution
 We can transform random numbers into
random variates using the cumulative
distribution function or special formulas
 Uniform random variate:
U = a + (b – a)*RAND( )

Copyright ©2013 Pearson Education, Inc. publishing as Prentice Hall

4-9

Random Variates in Excel

 NORM.INV(RAND(),mean,standard_deviation )
 NORM.S.INV(RAND() )
 LOGNORM.INV(RAND(),mean,standard_deviation )
 BETA.INV(RAND(),alpha,beta,A,B )
 GAMMA.INV(RAND(),alpha,beta )

Copyright ©2013 Pearson Education, Inc. publishing as Prentice Hall

4-10

5
 15/09/2022

Statistical Sampling
Experiment in Finance
 In finance, one way of evaluating capital budgeting projects
is to compute a profitability index (PI ), which is defined as
the ratio of the present value of future cash flows (PV ) to
the initial investment (I ):
PI = PV/I
 Suppose that PV is estimated to be normally distributed

with a mean of $12 million and a standard deviation of

$2.5 million, and the initial investment is also estimated to be
normal with a mean of $3.0 million and standard deviation of
$0.8 million. What is the distribution of PI ?

Copyright ©2013 Pearson Education, Inc. publishing as Prentice Hall

4-11

Experiment

Copyright ©2013 Pearson Education, Inc. publishing as Prentice Hall

4-12

6
 15/09/2022

Histogram of Simulated PI Values

Conclusion: The ratio of two normal random

variables is NOT normal!
Copyright ©2013 Pearson Education, Inc. publishing as Prentice Hall
4-13

Sampling Distributions and

Sampling Error
 How good is an estimate of the mean
obtained from a sample?
 Sampling distribution of the mean is the
distribution of the means of all possible
samples of a fixed size from some
population.
 Understanding sampling distributions is
the key to statistical inference.
Copyright ©2013 Pearson Education, Inc. publishing as Prentice Hall
4-14

7
 15/09/2022

Experiment
 Assume that a random variable is uniformly
distributed between 0 and 10. The expected
value is (0 + 10)/2 = 5, the variance is (10 –
0) /12 = 8.33, and the standard deviation is
2

2.89.
 Generate a sample of size n = 25 for this
random variable using the Excel function
10*RAND( ) and compute the sample mean. Repeat
this experiment several more times, say
20, and obtain a set of 20 sample means.

Copyright ©2013 Pearson Education, Inc. publishing as Prentice Hall

4-15

Sampling Experiment
Spreadsheet

Copyright ©2013 Pearson Education, Inc. publishing as Prentice Hall

4-16

8
 15/09/2022

Histogram of 20 Sample
Means

Copyright ©2013 Pearson Education, Inc. publishing as Prentice Hall

4-17

Results for Other Sample Sizes

Copyright ©2013 Pearson Education, Inc. publishing as Prentice Hall

4-18

9
 15/09/2022

Summary

Key observations:
1. All means appear close to the expected value of 5.
2. The standard deviation gets smaller as the sample
size increases.
Copyright ©2013 Pearson Education, Inc. publishing as Prentice Hall
4-19

Sampling Distribution of the

Mean
 Sampling distribution of the mean is the
distribution of the means of all possible
samples of a fixed size from some
population.
 Understanding sampling distributions is
the key to statistical inference.

Copyright ©2013 Pearson Education, Inc. publishing as Prentice Hall

4-20

10
 15/09/2022

Properties of the Sampling

Distribution of the Mean
 Expected value of the sample mean is
the population mean, m
 Variance of the sample mean is s 2 /n ,
where s 2is the variance of the
population
 Standard deviation of the sample mean,

called the standard error of the mean,

is s /n
Copyright ©2013 Pearson Education, Inc. publishing as Prentice Hall
4-21

Central Limit Theorem

 If the sample size is large enough (generally at
least 30, but depends on the actual
distribution), the sampling distribution of the
mean is approximately normal, regardlessof the
distribution of the population.
 If the population is normal, then the sampling
distribution of the mean is exactly normal for
any sample size.

Copyright ©2013 Pearson Education, Inc. publishing as Prentice Hall

4-22

11
 15/09/2022

Applying Sampling Distributions –

Example 1
 Suppose that the size of individual customer orders (in
dollars), X, from a major discount book publisher Web
site is normally distributed with a mean of $36 and
standard deviation of $8. What is the probability that the
next individual who places an order at the Web
site will purchase more than $40?
 P(X > 40) = 1 – NORMDIST(40,36,8,TRUE)
= 1 – 0.6915 = 0.3085
 The calculation uses the standard deviation of the
individual customer orders.

Copyright ©2013 Pearson Education, Inc. publishing as Prentice Hall

4-23

Applying Sampling Distributions –

Example 2
 Suppose that a sample of 16 customers is chosen.
What is the probability that the meanpurchase for
these 16 customers will exceed $40?
 The sampling distribution of the mean will have a
mean of $36, but a standard error of $8/16 = $2.
 P( X > 40) = 1 – NORMDIST(40,36,2,TRUE)
= 1 - .9772 = 0.0228
 The calculation uses the standard error of the
sampling distribution.

Copyright ©2013 Pearson Education, Inc. publishing as Prentice Hall

4-24

12
 15/09/2022

Sampling and Estimation

 Estimation – assessing the value of an unknown
population parameter using sample data.
 Point estimate – a single number used to estimate a
population parameter
 Confidence interval estimate – a range of values
between which a population parameter is believed to
be along with the probability that the interval
correctly estimates the true population parameter

Copyright ©2013 Pearson Education, Inc. publishing as Prentice Hall

4-25

Common Point Estimates

Copyright ©2013 Pearson Education, Inc. publishing as Prentice Hall

4-26

13
 15/09/2022

Example

Copyright ©2013 Pearson Education, Inc. publishing as Prentice Hall

4-27

Theoretical Issues
 Unbiased estimator – one for which the
expected value equals the population
parameter it is intended to estimate
 The sample variance is an unbiased
estimator for the population variance
n 2 n 2

 x  i  x  i

x i n  s 2mi 1
s2  1
N
1

Copyright ©2013 Pearson Education, Inc. publishing as Prentice Hall

4-28

14
 15/09/2022

Interval Estimates
 Range within which we believe the true
population parameter falls
 Example: Gallup poll – percentage of voters favoring
a candidate is 56% with a 3% margin of error.
 Interval estimate is [53%, 59%]
 A 100(1 – a)% probability interval is any
interval [A,B ] such that the probability of
falling between A and B is 1 – a

Copyright ©2013 Pearson Education, Inc. publishing as Prentice Hall

4-29

Confidence Intervals
 Confidence interval (CI) – an interval
estimated that specifies the likelihood that
the interval contains the true population
parameter
 Level of confidence (1 – a) – the likelihood that
the CI contains the true population parameter,
usually expressed as a percentage (90%, 95%,
99% are most common).

Copyright ©2013 Pearson Education, Inc. publishing as Prentice Hall

4-30

15
 15/09/2022

Confidence Intervals and

PHStat Tools

Copyright ©2013 Pearson Education, Inc. publishing as Prentice Hall

4-31

Confidence Interval for the

Mean – s Known

za/2 may be found from Table A.1 or using the

Excel function NORM.S.INV(a/2)

Copyright ©2013 Pearson Education, Inc. publishing as Prentice Hall

4-32

16
 15/09/2022

Sampling From Finite

Populations
 When n > 0.05N , use a correction factor
in computing the standard error:
s Nn
s 
x
n N
1

Copyright ©2013 Pearson Education, Inc. publishing as Prentice Hall

4-33

Example: Cereal Samples

(PHStat Tool)

Copyright ©2013 Pearson Education, Inc. publishing as Prentice Hall

4-34

17
 15/09/2022

Key Observations

 As the confidence level (1 - a)

increases, the width of the
confidence interval also increases.
 As the sample size increases, the
width of the confidence interval
decreases.
Copyright ©2013 Pearson Education, Inc. publishing as Prentice Hall
4-35

Confidence Interval for the

Mean, s Unknown

t a/2,n-1 is the value from a t-distribution with

n-1 degrees of freedom, from Table A.2 or
the Excel function T.INV(1-a/2, n -1)

Copyright ©2013 Pearson Education, Inc. publishing as Prentice Hall

4-36

18
 15/09/2022

Relationship Between Normal

Distribution and t-distribution

The t-distribution yields larger confidence

intervals for smaller sample sizes.
Copyright ©2013 Pearson Education, Inc. publishing as Prentice Hall
4-37

Example: CreditApproval
Decisions

Copyright ©2013 Pearson Education, Inc. publishing as Prentice Hall

4-38

19
 15/09/2022

Sample Proportion
 Sample proportion: pˆ = x/n
 x = number in sample having desired
characteristic
 n = sample size
 The sampling distribution of p has
mean p and variance p(1–p) /n ˆ
 When np and n(1–p) are at least 5, the
sampling distribution of p approach a
normal distribution
Copyright ©2013 Pearson Education, Inc. publishing as Prentice Hall
4-39

Confidence Interval for the

Proportion

PHStat tool is available under Confidence

Intervals option

Copyright ©2013 Pearson Education, Inc. publishing as Prentice Hall

4-40

20
 15/09/2022

Example: InsuranceSurvey

Copyright ©2013 Pearson Education, Inc. publishing as Prentice Hall

4-41

Sampling Distribution of s
 The sample standard deviation, s , is a point
estimate for the population standard deviation,
s
 The sampling distribution of s has a chi-
square (c2) distribution with n-1 df
 See Table A.3
 CHISQ.DIST(x,deg_freedom ) returns probability to

the right of x
 CHISQ.INV(probability,deg_freedom ) returns the

value of x for a specified right-tail probability

Copyright ©2013 Pearson Education, Inc. publishing as Prentice Hall
4-42

21
 15/09/2022

Confidence Interval for the

Variance

Note the difference in the

denominators!

Copyright ©2013 Pearson Education, Inc. publishing as Prentice Hall

4-43

Example: HomeMarketValue
Data

Copyright ©2013 Pearson Education, Inc. publishing as Prentice Hall

4-44

22
 15/09/2022

Confidence Interval for a

Population Total

Copyright ©2013 Pearson Education, Inc. publishing as Prentice Hall

4-45

Example
 Sample of 20
reimbursement claims
over 60 days old
yielded a mean of
$185 and standard
deviation of $22.

Copyright ©2013 Pearson Education, Inc. publishing as Prentice Hall

4-46

23
 15/09/2022

Confidence Intervals and

Decision Making
 Required weight for a soap product is
64 ounces. A sample of 30 boxes found a
mean of 63.82 and standard deviation of
1.05. A 95% CI is [63.43, 64.21]. What
conclusion can you reach?
 What if the standard deviation was 0.46
and the CI is [64.65, 63.99]?

Copyright ©2013 Pearson Education, Inc. publishing as Prentice Hall

4-47

Confidence Intervals and

Sample Size
 CI for the mean, s known
Sample size needed for half-width of at
most E is n  (za/2)2(s2)/E2
 CI for a proportion
 Sample size needed for half-width of at
most E is 2
( za / 2 ) p (1  p )
n E2
 Use the sample proportion as an estimate of
p, or 0.5 for the most conservative estimate
Copyright ©2013 Pearson Education, Inc. publishing as Prentice Hall
4-48

24
 15/09/2022

Example
 For the soap filling
process, the sampling
error associated with
the CI = [63.43,
64.21] is 0.39 (half-
width). What sample
size is required to
reduce this to 0.15?

Copyright ©2013 Pearson Education, Inc. publishing as Prentice Hall

4-49

Example
 How many voters
must be surveyed
to ensure a
sampling error of
at most ± 2%?

Copyright ©2013 Pearson Education, Inc. publishing as Prentice Hall

4-50

25
 15/09/2022

Prediction Intervals
 A prediction interval is one that
provides a range for predicting the
value of a new observation from the
same population.
 A 100(1 – a)% prediction interval for a
new observation is

Copyright ©2013 Pearson Education, Inc. publishing as Prentice Hall

4-51

Additional Types of
Confidence Intervals
 Difference in means
 Independent samples with unequal
variances
 Independent samples with equal
variances
 Paired samples
 Difference in proportions

Copyright ©2013 Pearson Education, Inc. publishing as Prentice Hall

4-52

26
 15/09/2022

Sampling Distribution of the

Mean – Theory

Copyright ©2013 Pearson Education, Inc. publishing as Prentice Hall

4-53

Interval Estimate Containing the

True Population Mean

Copyright ©2013 Pearson Education, Inc. publishing as Prentice Hall

4-54

27
 15/09/2022

Interval Estimate Not Containing

the True Population Mean

Copyright ©2013 Pearson Education, Inc. publishing as Prentice Hall

4-55

CI for Difference in Means

Copyright ©2013 Pearson Education, Inc. publishing as Prentice Hall

4-56

28
 15/09/2022

Independent Samples with

Unequal Variances

Copyright ©2013 Pearson Education, Inc. publishing as Prentice Hall

4-57

Independent Samples with

Equal Variances

Copyright ©2013 Pearson Education, Inc. publishing as Prentice Hall

4-58

29
 15/09/2022

Paired Samples

Copyright ©2013 Pearson Education, Inc. publishing as Prentice Hall

4-59

Differences Between
Proportions

Copyright ©2013 Pearson Education, Inc. publishing as Prentice Hall

4-60

30
 15/09/2022

Summary of CI Formulas

Copyright ©2013 Pearson Education, Inc. publishing as Prentice Hall

4-61

Summary of CI Formulas

Copyright ©2013 Pearson Education, Inc. publishing as Prentice Hall

4-62

31
 15/09/2022

All rights reserved. No part of this publication may be reproduced, stored in a retrieval system, or
transmitted, in any form or by any means, electronic, mechanical, photocopying, recording, or
otherwise, without the prior written permission of the publisher.
Printed in the United States of America.

Copyright ©2013 Pearson Education, Inc. publishing as Prentice Hall

4-63

32