STAT 201

Chapter 8

Estimation: Additional Topics

 Chapter Goals
After completing this chapter, you should be able to:
 Form confidence intervals for the mean difference from
dependent samples
 Form confidence intervals for the difference between
two independent population means (standard deviations
known or unknown)
 Compute confidence interval limits for the difference
between two independent population proportions
 Estimation: Additional Topics

Chapter Topics

Population Population
Means, Means, Population Population
Dependent Independent Proportions Variance
Samples Samples Ch.7
Examples:
Same group Group 1 vs. Proportion 1 vs. Variance of a
before vs. after independent Proportion 2 normal distribution
treatment Group 2
 Dependent Samples
Tests Means of 2 Related Populations
Dependent  Paired or matched samples
samples
 Repeated measures (before/after)
 Use difference between paired values:

di = xi - yi

 Eliminates Variation Among Subjects

 Assumptions:
 Both Populations Are Normally Distributed
 Mean Difference
The ith paired difference is di , where
Dependent
samples
di = x i - yi
n

d
The point estimate for
the population mean i

paired difference is d : d i 1
n
n
The sample
standard  i
(d  d) 2

deviation is: Sd  i1

n 1
n is the number of matched pairs in the sample
 Confidence Interval for
Mean Difference
The confidence interval for difference
Dependent between population means, μd , is
samples

Sd Sd
d  t n1,α/2  μd  d  t n1,α/2
n n
Where
n = the sample size
(number of matched pairs in the paired sample)
 Confidence Interval for
Mean Difference
(continued)

Dependent
 The margin of error is
samples
sd
ME  t n1,α/2
n

 tn-1,/2 is the value from the Student’s t

distribution with (n – 1) degrees of freedom
for which
α
P(t n1  t n1,α/2 ) 
2
 Paired Samples Example

 Six people sign up for a weight loss program. You

collect the following data:

Weight: 
d = n
di
Person Before (x) After (y) Difference, di

1 136 125 11 = 7.0

2 205 195 10
3 157 150 7
4
5
138
175
140
165
-2
10 Sd 
 i
(d  d) 2

n 1
6 166 160 6
42  4.82
 Paired Samples Example
(continued)

 For a 95% confidence level, the appropriate t value is

tn-1,/2 = t5,.025 = 2.571
 The 95% confidence interval for the difference between
means, μd , is
Sd S
d  t n1,α/2  μd  d  t n1,α/2 d
n n
4.82 4.82
7  (2.571)  μd  7  (2.571)
6 6
 1.94  μd  12.06

Since this interval contains zero, we cannot be 95% confident, given this
limited data, that the weight loss program helps people lose weight
 Difference Between Two Means

Population means, Goal: Form a confidence interval

independent for the difference between two
samples population means, μx – μy
 Different data sources
 Unrelated
 Independent
 Sample selected from one population has no effect on the
sample selected from the other population
 The point estimate is the difference between the two
sample means:
x–y
 Difference Between Two Means
(continued)

Population means,
independent
samples

σx2 and σy2 known Confidence interval uses z/2

σx2 and σy2 unknown

σx2 and σy2

assumed equal Confidence interval uses a value
from the Student’s t distribution
σx2 and σy2
assumed unequal
 σx2 and σy2 Known

Population means, Assumptions:

independent
samples  Samples are randomly and
independently drawn
σx2 and σy2 known
*  both population distributions
σx2 and σy2 unknown are normal

 Population variances are

known
 σx2 and σy2 Known
(continued)

When σx and σy are known and

Population means,
both populations are normal, the
independent
samples variance of X – Y is
2
σx
2
σy
σ 2X Y  
σx2 and σy2 known
* nx ny

…and the random variable

σx2 and σy2 unknown
(x  y)  (μX  μY )
Z
σ 2x σ y
2


nX nY

has a standard normal distribution

 Confidence Interval,
σx2 and σy2 Known

Population means,
independent
samples

σx2 and σy2 known

* The confidence interval for
μx – μy is:
σx2 and σy2 unknown

σ 2X σ 2Y σ 2X σ 2Y
(x  y)  z α/2   μX  μY  (x  y)  z α/2 
nx ny nx ny
 σx2 and σy2 Unknown,
Assumed Equal

Population means, Assumptions:

independent
 Samples are randomly and
samples
independently drawn

σx2 and σy2 known  Populations are normally

distributed
σx2 and σy2 unknown
 Population variances are
σx2 and σy2
assumed equal * unknown but assumed equal

σx2 and σy2

assumed unequal
 σx2 and σy2 Unknown,
Assumed Equal
(continued)

Population means, Forming interval

independent estimates:
samples
 The population variances
σx2 and σy2 known are assumed equal, so use
the two sample standard
deviations and pool them to
σx2 and σy2 unknown
estimate σ
σx2 and σy2
assumed equal *  use a t value with
(nx + ny – 2) degrees of
σx2 and σy2 freedom
assumed unequal
 σx2 and σy2 Unknown,
Assumed Equal
(continued)

Population means,
independent
samples
The pooled variance is
σx2 and σy2 known

σx2 and σy2 unknown (nx  1)s 2x  (ny  1)s2y

sp2 
σx2 and σy2 nx  ny  2
assumed equal *
σx2 and σy2
assumed unequal
 Confidence Interval,
σx2 and σy2 Unknown, Equal

σx2 and σy2 unknown

σx2 and σy2

assumed equal * The confidence interval for
μ1 – μ2 is:
σx2 and σy2
assumed unequal

sp2 sp2 sp2 sp2

(x  y)  t nx ny 2,α/2   μX  μY  (x  y)  t nx ny 2,α/2 
nx ny nx ny

(nx  1)s 2x  (ny  1)s2y

Where sp2 
nx  ny  2
 Pooled Variance Example 1
You are testing two computer processors for speed.
Form a confidence interval for the difference in CPU
speed. You collect the following speed data (in Mhz):

CPUx CPUy
Number Tested 17 14
Sample mean 3004 2538
Sample std dev 74 56

Assume both populations are

normal with equal variances,
and use 95% confidence
 Calculating the Pooled Variance

The pooled variance is:

S 
2
n x  1S x
2
 n y  1S y
2


17  174 2  14  156 2
 4427.03
(nx  1)  (ny  1) (17 - 1)  (14  1)
p

The t value for a 95% confidence interval is:

t nx ny 2 , α/2  t 29 , 0.025  2.045

 Calculating the Confidence Limits

 The 95% confidence interval is

sp2 sp2 sp2 sp2

(x  y)  t nx ny 2,α/2   μX  μY  (x  y)  t nx ny 2,α/2 
nx ny nx ny

4427.03 4427.03 4427.03 4427.03

(3004  2538)  (2.054)   μX  μY  (3004  2538)  (2.054) 
17 14 17 14

416.69  μX  μY  515.31

We are 95% confident that the mean difference in

CPU speed is between 416.69 and 515.31 Mhz.
 Pooled Variance Example 2
Two catalysts in a batch chemical process are being
compared for their effect on the output of the process
reaction. Form a confidence interval for the difference
between population means.

Catlyst 1 Catlyst2
Sample size 10 11
Sample mean 86 77
Sample std dev 3 6

Assume both populations are

normal with equal variances,
and use 95% confidence
 Calculating the Pooled Variance

The pooled variance is:

S 
2
n x  1S x
2
 n y  1S y
2


10  132  11 162
 23.21
(n x  1)  (n y  1) (10 - 1)  (11 1)
p

The t value for a 95% confidence interval is:

t n x  n y 2 , α/2  t19 , 0.025  2.093

 Calculating the Confidence Limits

 The 95% confidence interval is

sp2 sp2 sp2 sp2

(x  y)  t nx ny 2,α/2   μX  μY  (x  y)  t nx ny 2,α/2 
nx ny nx ny

23.21 23.21 23.21 23.21

(86 - 77)  (2.093)   μ X  μ Y  (86 - 77)  (2.093) 
10 11 10 11

4.59  μ X  μ Y  13.41

We are 95% confident that the mean difference in

catalyst performance is between 4.59 and 13.41.
 σx2 and σy2 Unknown,
Assumed Unequal

Population means, Assumptions:

independent
 Samples are randomly and
samples
independently drawn

σx2 and σy2 known  Populations are normally

distributed
σx2 and σy2 unknown
 Population variances are
σx2 and σy2 unknown and assumed
assumed equal unequal
σx2 and σy2
assumed unequal *
 σx2 and σy2 Unknown,
Assumed Unequal
(continued)

Forming interval estimates:

Population means,
independent
 The population variances are
samples
assumed unequal, so a pooled
variance is not appropriate
σx2 and σy2 known
 use a t value with  degrees
σx2 and σy2 unknown of freedom, where
2
 s2x s2y 
σx2 and σy2 ( )  ( )
assumed equal  n x n y 
v 2
σx2 and σy2
2 2
 sx   s 2

  /(nx  1)   y  /(ny  1)
assumed unequal *  nx 
n 
 y
 Confidence Interval,
σx2 and σy2 Unknown, Unequal

σx2 and σy2 unknown

σx2 and σy2 The confidence interval for

assumed equal
μ1 – μ2 is:
σx2 and σy2
assumed unequal *
2 2
s2x s y s2x s y
(x  y)  t ,α/2   μX  μY  (x  y)  t ,α/2 
nx ny nx ny
2
 s2x s2y 
( )  ( )
v  n x n y 
Where 2
 s2 
2
 s2x 
  /(nx  1)   y  /(ny  1)
n 
 nx   y
 Example
From a random sample of seven students in a marketing
research class that uses group-learning techniques, the
mean examination score was found to be 78.25 and the
sample standard deviation was 2.87. For an
independent random sample of ten students in another
marketing research class that does not use group-
learning techniques, the sample mean and standard
deviation of exam scores were 74.94 and 9.15,
respectively.

Estimate with 95% confidence the difference

between the two population mean scores; do not
assume equal population variances.
 Example
ANSWER:
nx  7, x  78.25, sx  2.87, ny  10, y  74.94, s y  9.15

The degrees of freedom are given by

 s 2   s 2  
2
 s 2  2  s22 
2

   1    2     /(n1  1)    /(n2  1)  
1
11
 n1   n2    n1   n2  

  .05  t , / 2  t11,.025  2.201

The with 95% confidence interval for the difference
between the two population mean scores is
(x  y)  t  ( s / n ) ( s / n )  (78.25 – 74.94)  (2.201)((3.09)
 , / 2
2
x x
2
y y

3.31  6.80 or -3.49 < x   y < 10.11

Additional Exercises for the
difference between two
population means
 Example 1
A dependent random sample from two normally
distributed populations gives the following results:
n = 15, d = 20.5, and s = 2.4. d

Find the margin of error for a 95% confidence

interval for the difference in the means of the
two populations.

ANSWER:
n  15, df  n  1  14, tn1, / 2  t14,.025  2.1448

Margin of error: ME = tn1, / 2  sd / n  (2.1448)(2.4/ 15)  1.329

d  ME  20.5  1.329 or 19.171 < d < 21.829

 Example 2
Independent random sampling from two normally
distributed populations gives the following results:
n = 64, x = 441,  = 20, n = 36, y = 361, and  = 25.
x x y y

Find the margin of error for a 90% confidence

interval for the difference in the means of the two
populations.

ANSWER:
  0.10  z / 2  z0.05  1.645
Margin of error: ME  z . ( / n )  ( / n )  1.645
 /2
2
x x
2
y y (400 / 64)  (625 / 36)

= (1.645)(4.859) = 7.993
 Example 2

Find the 95% confidence interval for the difference in the

means of the two populations.

ANSWER:
( x  y )  ME  80  7.993 or 72.007 <  X  Y < 87.993

Based on the above confidence interval, is there

evidence that the population means are different?

ANSWER:
This confidence interval does not include zero,
indicating strong evidence that the population
means are different.
 Example 3
Suppose, for a random sample of 250 firms that revalued their
fixed assets, the mean ratio of debt to tangible assets was
0.528 and the sample standard deviation was 0.151. For an
independent random sample of 500 firms that did not revalue
their fixed assets was 0.502 and the sample standard deviation
was 0.162. Find a 99% confidence interval for the difference
between the two population means.

ANSWER:
nx  250, x  0.528, sx  0.151, ny  500, y  0.502, sy  0.162, z.005  2.575
The 99% confidence interval for the difference between the two
population mean scores is
( x  y )  z  ( s / n ) ( s / n )  (0.528 – 0.502)  (2.575)((0.01199)
/2
2
x x
2
y y

0.026  0.031 or -0.005 < x   y < 0.057

Confidence Interval for Two
Population Proportion
 Two Population Proportions
Goal: Form a confidence interval for
Population the difference between two
proportions population proportions, Px – Py

Assumptions:
Both sample sizes are large (generally at
least 40 observations in each sample)

The point estimate for

the difference is
pˆ x  pˆ y
 Two Population Proportions
(continued)

 The random variable

Population
proportions
(pˆ x  pˆ y )  (px  p y )
Z
pˆ x (1 pˆ x ) pˆ y (1 pˆ y )

nx ny

is approximately normally distributed

 Confidence Interval for
Two Population Proportions

Population The confidence limits for

proportions
Px – Py are:

pˆ x (1 pˆ x ) pˆ y (1 pˆ y )
(pˆ x  pˆ y )  Z / 2 
nx ny
 Example:
Two Population Proportions
Form a 90% confidence interval for the
difference between the proportion of
men and the proportion of women who
have college degrees.

 In a random sample, 26 of 50 men and

28 of 40 women had an earned college
degree
 Example:
Two Population Proportions
(continued)

Men: ˆp x  26  0.52
50

Women: ˆp y  28  0.70
40

pˆ x (1 pˆ x ) pˆ y (1 pˆ y ) 0.52(0.48) 0.70(0.30)

    0.1012
nx ny 50 40

For 90% confidence, Z/2 = 1.645

 Example:
Two Population Proportions
(continued)

The confidence limits are:

pˆ x (1 pˆ x ) pˆ y (1 pˆ y )
(pˆ x  pˆ y )  Z α/2 
nx ny

 (.52  .70)  1.645 (0.1012)

so the confidence interval is

-0.3465 < Px – Py < -0.0135

Since this interval does not contain zero we are 90% confident that the
two proportions are not equal
 Example

In a random sample of 125 large retailers, 90

used regression as a method of forecasting.
In an independent random sample of 160
small retailers, 80 used regression as a
method of forecasting.

Find a 95% confidence interval for the

difference between the two population
proportions.
 Example

ANSWER:
nx  125, pˆ x  90 /125  0.72, ny  160, pˆ y  80 /160  0.50, z / 2  z.025  1.96

The margin of error ME is

pˆ x (1  pˆ x ) pˆ y (1  pˆ y ) (0.72)(0.28) (0.50)(0.50)
ME = z / 2
nx

ny
 1.96
125

160
 0.1104
The 95% confidence interval for the difference
between the two population proportions is
( pˆ x  pˆ y )  ME  (0.72  0.50)  0.1104

0.22  0.1104 or 0.1096 < P  P < 0.3304

x y
 Chapter Summary
 Compared two dependent samples (paired samples)
 Formed confidence intervals for the paired difference
 Compared two independent samples
 Formed confidence intervals for the difference between two
means, population variance known, using z
 Formed confidence intervals for the differences between two
means, population variance unknown, using t
 Formed confidence intervals for the differences between two
population proportions
 Determined required sample size to meet confidence
and margin of error requirements