Exercise 1. Illustrate the value of the following t-distributions.

a. two-tailed with 𝛼 = 10% and df=10. tcritical =±1.812

b. one-tailed with 𝛼 = 5% and df=18. tcritical = 1.734
c. two-tailed with 𝛼 = 5% and n=29. tcritical =±2.048

Chapter 4: ESTIMATION OF PARAMETERS

Lesson 4.1.

Estimating Population Mean ( μ ¿ When Variance (σ ¿ Is Known or Whenn ≥ 30.

Statistical inference is making a conclusion or generalization about the population based on the study of
samples.

Point estimate is a single value that estimates the population parameter, such as x as estimate for μ or
as s as estimate for σ .

Interval estimate, sometimes called confidence interval, is a range of interval (with lower and upper
limits) used to estimate the population parameter. It is usually in the form of a< θ<b ., which tells that
the estimate parameter (θ) is between two values (a and b) at a certain level of confidence.

When the population variance or standard deviation is known or when n ≥ 30 (by central limit theorem),
The formula below can be used as an interval estimate of population mean ( μ)at a certain degree of
confidence (α ¿ :

x−z α /2
( √σn )< μ< x + z ( √σn )
α/ 2

Where :

x=¿sample mean.
σ =¿Population standard deviation.
n = Sample size
z α / 2 = z-value leaves an area ofα / 2
 The values of are listed below with the usual confidence level used in estimating population mean.

Confidence level
or 90% 95% 97% 98% 99%
(1-α ¿ 100 %
zα/ 2 1.64 1.96 2.17 2.33 2.58

α
Confidence level: 90%, α =10 % /0.10, =0.05 z α / 2 = 1.64
2

α
Confidence level: 97%, α =3 % /0. 03, =0.015 z α / 2 = 2.17
2
α
Confidence level: 98%, α =2 % /0.0 2, =0.01 z α/ 2 = 2.33
2
α
Confidence level: 99%, α =1 % /0.01 , =0.005 z α / 2 = 2.58
2

From the formula, z α / 2 ( √σn ) is called margin of error.

Margin of error refers to the maximum acceptable difference (determined by α ) between the observed
sample statistic (mean or proportion) and the true population parameter (mean or proportion).
 ( )
2
zα/ 2 ∙ σ
Formula for sample size: n= E

Note that if the computed value of n has a decimal part, round the value to the next whole number.

Example 1

Compute the margin of error of the 95% confidence interval estimate of μ when σ =10, n=25.

Solution:

α
Confidence level: 95%, α =5 % /0.0 5, =0.025
2
z α / 2  0.5000-0.0250 = 0.4750 z α / 2=1.96

z α/ 2
( √σn )=( 1.96) √1025 =3.92
Example 2

Compute the 98% confidence interval estimate of μ given the following:

σ =6.5, n=40 ,∧x=42.

Solution:

α
Confidence level: 98%, α =2 % /0.0 2, =0.01
2
z α/ 2  0.5000-0.0100 = 0.4900 z α/ 2=2.33

( √σn )< μ< x + z ( √σn )

x−z α /2 α/ 2

42−2.33 ( ) < μ< 4 2+2.33 (

√ 40 )
6.5 6.5
√ 40
42−2.39< μ< 4 2+2.39
39.61< μ< 44.39
Example 3

A car company wants to estimate the miles-per-gallon rate of their new car model. It was known
through experience that similar models have standard deviation of 4.5 miles-per-gallon. A random
sample of 100 units of the new model showed a mean of 42 miles-per-gallon. Construct a 97%
confidence interval for the population mean (μ) miles-per-gallon rate of the new car model.

Given: σ =4.5 , n=100 ,∧x=42

α
Confidence level: 97%, α =3 % /0.0 3, =0.0150
2
z α/ 2  0.5000-0.0150 = 0.4850 z α/ 2=2. 17

( √σn )< μ< x + z ( √σn )

x−z α /2 α/ 2

42−2.17
( √10 0 )
4 .5
< μ<4 2+ 2.17
( √100 )
4.5

42−0.977< μ< 4 2+0.977

41.02< μ< 42.98

Example 4

In estimating the mean selling price of all college textbooks, which was reported to have a standard
deviation of 80 Php, how many college textbooks must be selected as samples if you want to be 95%
confident that the sample mean ( x ) is within 20 Php of the true population mean (μ)?

Given: σ =80 ,∧E=¿ 20

α
Confidence level: 95%, α =5 % /0.0 5, =0.02 50
2
z α/ 2  0.5000-0.0250 = 0.4850 z α/ 2=1.96

n= ( E ) (
z α / 2 ∙ σ 2 1.96 ∙80 2
=
20 ) 2
=( 7.84 ) =61.47 ≈ 62 textbooks
 Lesson 4.2.

Estimating Population Mean ( μ ¿ When Variance (σ ¿ Is Unknown and Whenn<30 .

We have established earlier that when n < 30 (small sample size, the Central Limit Theorem cannot be
applied, and thus, if population standard deviation (σ ) is unknown for small sample size, the sample
standard deviation s cannot take its place, therefore, the interval estimate using the z-table cannot be
used. For this case, t-distribution is used to make an interval estimate. The formula becomes.

x−t α / 2 ( √sn )< μ< x +t ( √sn )

α /2

Where :

x=¿sample mean.
s=¿sample standard deviation.
n = Sample size
t α / 2 = t-value with n-1 degree of freedom, that leaves an area ofα / 2.

From the formula, t α / 2 ( √sn ) is called margin of error.

Example 1

Compute the margin of error of the 90% confidence interval estimate of μ when s=5 ,n=16.

α
Confidence level: 90%, α =10 % /0.10 , =0.05 df = n-1 = 15
2
t α / 2  1.753

t α/ 2
( √sn )=( 1. 753) √516 =2.19
Example 2

Compute the 95% confidence interval estimate of μ given the following: s=9 , n=12 ,∧x=27.

Solution:

α
Confidence level: 95%, α =5 % /0.0 5, =0.025 @ df = n – 1 = 12 – 1 = 11
2
t α / 2  2.201

( √sn )< μ< x +t ( √sn )

x−t α / 2 α /2

27−2.201
( √912 )< μ< 27+2.201( √912 )
2 7−5.718< μ< 27+5.718
21.282< μ<32.718

Example 3

A printer manufacturer wants to estimate the pages-per-cartridge print rate of their new printer model.
A random sample of 10 units of the new printer model showed a mean of 1400 pages-per-cartridge with
a standard deviation of 20. Construct a 98% confidence interval to estimate the pages-per-cartridge rate
of their new printer model.

α
Confidence level: 98%, α =2 % /0.0 2, =0.01 @ df = n – 1 = 10 – 1 = 9
2
t α/ 2  2.821
 ( )
x−t α / 2
s
√n ( )
< μ< x +t α /2
s
√n

1400−2.821 ( ) < μ<1400+2.821 (

√ 10 )
20 20
√1 0
1400−17.84 2< μ <1400+17.84 2
1382.158< μ<1417.842