Chapter 2
Statistical
Estimation
CONTENTS
I. Basic Concepts
II. Point estimators
of the mean and
proportion
[Link]
estimators of
the mean and
proportion
[Link]’s t-
distribution
Point and Interval
estimates of Population
parameters
• Point estimator is a sample statistic that is used to
estimate an unknown parameter.
• Point estimator is a single value of an estimator.
• The following are point estimators of population
parameter:
• is an estimator of the unknown µ
• is an estimator of the unknown P
• s is an estimator of the unknown ∂
• S is an estimator of the unknown
• S is an estimator of the unknown
Point and Interval
estimates
Example of Population
of Point estimators
Suppose we have the following random
parameters
sample of n = 6 elements from a population
whose parameter values are not known.
1 2 4 5 7 11;
a) xˉ = (∑xi)/n = 5, which is 5 is point
estimator of the unknown population
mean
b) pˉ of even numbers = 2/6 = 1/3 is a
point estimator of the unknown
population proportion
c) s = √∑(xi - xˉ)2 /n-1 = 3.63 is a point
estimator of the unknown population
standard deviation
d) sxˉ = s/√n = 1.483 a point estimator of
the unknown ∂xˉ
Point and Interval
estimates of Population
parameters
Interval Estimate: is an interval
estimate describes a range of values
within which a parameter might lie.
Example: if population mean is
estimated as within 1 value of the
sample mean value 5, the interval
estimate will be:
µ = xˉ± 1
= 5 ± 1 i.e. 5 – 1 & 5 + 1
Confidence Interval Estimates
of a Population mean µ
• The population estimated interval
is an inference based upon:
• The value of xˉ of a simple
random sample selected from
the population
• Known facts about the
sampling distribution of the xˉ
Assumptions
A. Assuming Normalof Confidence
population, ∂xˉ
Interval
is knownEstimates of a
•
Populationlargemean
sample µ size (n ≥ 30) –
approximates normality, or
• With small sample case
(n<30) – the population
need to be known normal
• Population standard
deviation (∂) is known
Example: The most recent ABC customer
satisfaction survey provided data on
the satisfactions course of a sample
of 100 customers the sample mean
satisfaction score was 82 &
Assumptions of Confidence
Interval Estimates of a
Population mean µ
Solution, Note:
• Confidence Coefficient (c): is the proportion of correct
statement
• Alpha (α): is the proportion of incorrect statements α = 1-c
The formula for population mean µ estimation is:
xˉ ± Z α/2*∂xˉ
Given:
n = 100; xˉ = 82; ∂ = 20; c = 0.95
α = 1-c = 0.05; at α = 0.05, Z α/2 = 1.96 from a
table.
µ = unknown population mean satisfaction score to be estimated.
∂xˉ = ∂/√n = 20/ √100 = 2
Table of Significance
level for Z α/2
Significance Level (α)
Assumptions of Confidence Interval
Z α/2
Estimates of a Population mean µ
0.10
Solution Contd…
1.64
xˉ ± Z α/2 *∂xˉ 0.05
82 ± 1.96*2 82 ± 3.92 1.96
0.025
78.08 to 85.92 2.24
0.01
2.57
0.005
Interpretation: At 95% confidence2.84
level, XYZ can conclude that, the
0.001
mean satisfaction score for the3.27
Assumptions of Confidence
Interval Estimates of a
Example 2: A normal population has a s.d. of
Population
12. A randommean
sampleµof size 25 has a
mean of 50. Construct a 99% confidence
interval estimate of the population
mean. Table of Significance level for
Solution Z α/2
Significance Level (α)
Given: Z α/2
0.10
n = 25 xˉ = 50; ∂ = 12; c = 0.99 1.64
0.05
α = 1-c = 0.01; at α = 0.01, Z α/2 = 2.57 from1.96a
table. 0.025
2.24
µ = unknown population mean satisfaction score0.01
2.57
to
be estimated. 0.005
2.84
∂xˉ = ∂/√n = 12/ √25 = 2.4 0.001
3.27
xˉ ± Z α/2*∂xˉ 50 ± 2.57*2.4
50 ± 6.2 43.8 to 56.2, we state that 99%
Unknown
Assumptions of Confidence
• large sample size
Interval Estimates of a(n ≥ 30) –
approximates normality, or
Population
•
mean µ
Population standard deviation
(∂) is unknown
• In this case we simply use
sample s.d (s) to estimate ∂,
thus ∂xˉ = ∂/√n would be
estimated by sxˉ = s/√n interval
estimation for µ would be: xˉ ± Z
α/2*sxˉ
Example: Nyala Insurance Co. conducted a
sampling study selecting a sample
random sample of 36 life insurance
policy holders. The sample mean age is
39.5 years and the sample s.d. is 7.77
years. At 90% confidence level,
Assumptions of Confidence
Solution
Interval Estimates of a
Table of Significance
Given:
Population mean µ level for Z α/2
n = 36 xˉ = 39.5;s = 7.77; c Significance
= 0.90 Level
(α) Z α/2
α = 1-c = 0.10; at α = 0.10, Z α/2 = 1.64
from a table. 0.10
1.64
µ = unknown population mean satisfaction score to
be estimated. 0.05
1.96
sxˉ = s/√n = 7.77/ √36 = 1.295 0.025
xˉ ± Z α/2*sxˉ 39.5 ± 1.64*1.2952.24
0.01
2.57
0.005
39.5 ± 2.12 37.37 to 41.62 2.84 37.37 ≤
µ ≤ 41.62, we are 90% confident that, the
0.001
mean age for the population of Nyala
3.27 insurance
company’s life insurance policy holders is
• small sample size (n < 30) –
Assumptions of Confidence
approximates normality, or
Interval
• Estimates
Population ofdeviation
standard a (∂) is
Population
unknownmean µ
• The population is assumed to be
normally distributed
• In this case we simply use sample s.d
(s) to estimate ∂, thus ∂xˉ = ∂/√n
would be estimated by sxˉ = s/√n
interval estimation for µ would be:
xˉ ± t α/2*V*sxˉ , where V is the
degree of freedom. V = n-1
• Instead of Z we will use student t
table as, t α/2
Example: Management of Meta Brewery
Factory has agreed to train 15
employees in a computer aided program
so as to reduce the time needed to train
employees. The sample mean training
time is 53.87 days and a sample s.d. of
Assumptions of Confidence
Interval Estimates of a
Population mean µ
Solution
Given:
n = 15 xˉ = 53.87; s = 6.82; c = 0.95
α = 1-c = 0.05; V = n-1 = 14 at α = 0.05, tα/2 = 2.14 from a table.
µ = unknown population mean satisfaction score to be estimated.
sxˉ = s/√n = 6.82/√15 = 1.76
xˉ ± tα/2*sxˉ 53.87 ± 2.14*1.76
39.5 ± 3.766 50.10 to 57.60 50.10 ≤ µ ≤ 57.60, The 95%
confidence interval estimate of the population mean training
time is between 50.10 days and 57.60 days.
Interval Estimation of a
Population Proportion (P)
• The population proportion of success P is unknown
• pˉ will be used as the estimator of P
• Interval estimate of a population proportion is given by:
pˉ ± Zα/2*√(pˉqˉ/n), can be expressed as,
pˉ - Zα/2*√(pˉqˉ/n) ≤ P ≤ pˉ + Zα/2*√(pˉqˉ/n)
Where pˉ is sample proportion of success
qˉ is sample proportion of failure
α = 1-c
Interval Estimation of a
Population Proportion (P)
Example: From a sample of 45 employees of ABC co. who computed
computer aided program 36 have passed the exam given at the end of
the program. Construct the interval estimate of the population proportion
of employees who have passed the exam at 0.95 confidence coefficient.
Solution
Given: n = 45 employees, 36 passed, 9 failed
pˉ 36/45 = 0.8, qˉ 1- pˉ = 0.2; c = 0.95 α = 0.05 Zα/2= 1.96
pˉ ± Zα/2*√(pˉqˉ/n)
0.8 ± 1.96* √(0.8*0.2/45) 0.8 ± 0.12
0.68 ≤ P ≤ 0.92, At 95% confidence, the interval
estimate of the population proportion of employees who have passed the
exam is from 0.68 to 0.92
Interval Estimation of a
Population Proportion (P)
Example2: Last June a random sample of 400 members of the labor force in
a given region showed that 32 were unemployed. Construct the 95%
confidence interval for the proportion unemployed in the region.
Solution
Given: n = 400 employees, 32 unemployed (success) , 368 employed
(failure)
pˉ 32/400 = 0.08, qˉ 1- pˉ = 0.92; c = 0.95 α = 0.05 Zα/2= 1.96
pˉ ± Zα/2*√(pˉqˉ/n)
0.08 ± 1.96* √(0.08*0.92/400) 0.8 ± 0.027
0.053 ≤ P ≤ 0.107, With 95% confidence, we can state
that the population proportion unemployed in the region is between 5.3%
and 10.7%.
Statistical Table usage Summary
