Estimation of Parameters

Goal: How can we use sample data to estimate values of population parameters?

Definition of Terrms:
Estimate/ statistic- it’s a measurement of samples. An estimate can be derived from samples. Statistically,
this value or range of values can be used to approximate a parameter. Examples are, sample mean (X),
sample standard deviation (s), and sample variance (s2)

Parameter- it’s a measurement of population. A parameter can be derived from the entire set of data
(population). Statistically, these values are usually unknown. Examples are population mean (𝝻) , population
standard deviation (𝞼) etc.

Estimation- is the process of determining parameter values.

Question: Why do we estimate? In statistics, we do estimation to give approximation wtithin a certain

confidence level.

• In order to feel confident about our estimators as approximations of the true parameter values, we
take as many random samples as possible through “RANDOMIZATION”.

A good estimator has the following properties:

1. When the mean of a sample statistic is equal to the true population parameter.
2. Across the many repeated samples, the estimates are not very far from the true parameter value.

Negative bias (under estimate) Unbiased estimate Positive bias (Over estimate)

Two types of estimate

Point estimate - A single statistic value that is the “best guess or estimate” for the parameter value.
Confidence Interval/ Interval estimate – A range of values that may contain the parameter of a population.

Most Common Point estimators

• Sample mean (X) estimates population mean 𝝻.
• Sample standard dev. (s) estimates population standard dev. (𝞼)
• Sample proportion (𝑝̂ ) estimates population proportion (p)

Confidence Interval/ Interval estimate

Assumption: CI’s is believed to contain the true parameter value

Most CI’s are in the form : POINT ESTIMATE ± E (Margin of Error)

** The best point estimate is “Sample Mean”.
Formula:
𝜎
𝐸 = 𝑧𝑐 ( 𝑛) infinite population where 𝞼 = pop.sd; n = sample size; 𝑧𝑐 = critical value/test statistic and E = margin or
√
error

𝜎 𝑁−𝑛
𝐸 = 𝑧𝑐 ( 𝑛) √𝑁−1 finite population where N = population size
√

𝜎 𝑠
When 𝞼 is unknown (often case), the sample sd (s) is used. 𝐸 = 𝑧𝑐 ( 𝑛) = 𝑧𝑐 ( 𝑛)
√ √
***Note: z (test statistic/ critical value)- is applicable for large sample size with known pop.sd.

Confidence Level – is the probability that the interval estimate the parameter. It describes the percentage.
Usually, 1 – 𝛼.

Example:
𝛼= 0.05 = 1 – 0.05 = 0.95 or 95% “ We say that, we are 95% confident that the true value/parameter is within the
interval.

Common alphas:
Alpha 1- 𝛼 Confidence level z (test statistic)
0.05 95% ±1.96
0.01 99% ±2.58
0.10 90% ±1.65

Example Problems:
1. A random sample of 100 students are asked how much they spend for a meal during weekdays. The average expense is
found to be Php 75.00. Past experience has revealed that the population standard deviation is Php 5.00. Find the point
estimate and the interval estimate of the population mean (𝝻) , using the 95% confidence level.

n = 100, 𝑋̅ = 75, 𝞼= 5, 𝛼 =0.05=95% confidence level, and zc= ±1.96

Solution:
point estimate = 𝑋̅ = 75
𝜎 5
margin of error E= 𝑧𝑐 ( 𝑛) = ±1.96 ( ) = ±0.98
√ √100

Confidence Interval :
point estimate ± margin of error
𝑋̅ ± E = 75 ± 0.98
= 75 – 0.98 to 75+ 0.98
CI = 74.02 – 75.98 (We say that, we are 95% confident the the population mean is between
this interval.

2. Given the information: n = 50; 𝞼 = 0.50, and 𝑋̅ = 18.

Find a. point estimate
b. confidence interval within 𝛼= 0.10

𝛼= 0.10= 90% confidence level, zc= ±1.65

𝜎 0.5
Solution: a. point estimate = 𝑋̅ = 18 margin of error E= 𝑧𝑐 ( 𝑛) = ±1.65 (
√50
) = ±0.12
√
b. CI = Point estimate ± margin of error
CI = 18 ± 0.12
CI = 18 – 0.12 to 18 +0.12
CI = 17.88 – 18.12 (We say that, we are 90% confident that the population mean is between
this interval.
3. In a survey, it was found that the average daily intake of coffee of 4000 adults, selected a random sample
of n= 80, was 745 grams per day with standard deviation of s= 40 grams per day. Construct a 99%
confidence interval for the mean.

Solution: n =80; s= 40; 𝛼 = 0.01 =99% confidence level = zc= ±2.58

Point estimate: X
̅ = 745
𝑠 40
Confidence Interval: ̅
X ± 𝑧𝑐 ( ) = 745 ± 2.58 ( )
√𝑛 √80
= 745 ± 11. 54
= 733.46 − 756.54 (we say that, we are 99% confident that the population mean is
within this interval.

Small samples (n < 30)

𝑠
Use t-statistic instead of z -statistic. 𝐸 = 𝑡𝑐 ( )
√𝑛

Question: When can we use t instead of z?

• When samples are approximately normally distributed
• When sample size is small and the pop.sd is unknown.

To find for Confidence Interval : Point estimate ± Margin of error ̅ ± 𝑡𝑐 ( 𝑠 )

: X
√𝑛
To find for the tc, we are going to use the table below:
Examples:
1. A sample of 10 measurement of the length of a chip gave a mean 𝐗 ̅ = 𝟐. 𝟑𝟔 cm and a standard deviation s
= 0.06 cm. Find the confidence interval using a) 0.05 and b) 0.01 confidence levels.

̅ = 𝟐. 𝟑𝟔; s = 0.06
A. Solution: n= 10; point estimate: 𝐗

df (degree of freedom) = n – 1 𝛼 = 0.05= 95% = tc = 2.262

df = 10 – 1
df = 9

𝑠
Confidence Interval: = ̅
X ± 𝑡𝑐 ( )
√𝑛
0.06
= 2.36 ± 2.262 ( )
√10
= 2.36 ± 2.262(0.01897)
= 2.36 ± 2.262(0.01897)
= 2.36 ± 0.0429
= 2.3171 − 2.4029

̅ = 𝟐. 𝟑𝟔; s = 0.06
B. Solution: n= 10; point estimate: 𝐗

df (degree of freedom) = n – 1 𝛼 = 0.01= 99% = tc = 3.250

df = 10 – 1
df = 9

̅ ± 𝑡𝑐 ( 𝑠 )
Confidence Interval: = X
√𝑛
0.06
= 2.36 ± 3.250 ( )
√10
= 2.36 ± 3.250(0.01897)
= 2.36 ± 3.250(0.01897)
= 2.36 ± 0.06165
= 2.2983 − 2. 421
Exercises with Answers:

1. Find a 95% confidence interval for a population mean (𝝻) for these values.
̅ = 12.4, 𝞼 = 1.32
a. n = 40, X

Which test statistic is applicable?

Answer: pop. sd is known and n is larger than 30: z statistic (zc) = 95% = ± 1.96
𝜎
Confidence Interval: = ̅
X ± 𝑧𝑐 ( )
√𝑛

1.32
= 12.4 ± 1.96 ( )
√40
= 12.4 ± 1.96(0.2087)
= 12.4 ± 0.4091

𝐶𝐼 = 11.99 − 12.81

̅ = 15, s = 6.02
b. n = 18, X

Which test statistic is applicable?

Answer: pop. sd is unknown and n is small: t statistic (tc) = 95% = ± 2.110

df = n-1

df = 18 -1

df= 17

𝑠
̅ ± 𝑡𝑐 ( )
Confidence Interval: = X
√𝑛

6.02
= 15 ± 2.11 ( )
√18
= 15 ± 2.11(1.4189)
= 15 ± 2.9939

𝐶𝐼 = 12.01 − 17.99
What I Can Do?

A. Find the 95 % confidence interval for a population mean of the following values.
̅=442 , 𝞼 = 8.4
1. n = 125, X

2. n = 24, ̅
X=48 , s = 7.05

3. n = 15, ̅
X=18.8 , s = 4.08
̅=29.5 , 𝞼2 = 0.89
4. n = 100, X

5. n = 45, ̅
X=44.8 , 𝞼 = 1.11

B. Given the following values, find the E (margin of error). Choose the appropriate test statistic.

1. n =6 , s =2 , 𝛼= 0.10 (90%)

2. n= 44, 𝞼 = 4.32, 𝛼 =0.05 (95%)

3. n = 123, 𝞼= 8.88, 99% confidence level (0.01)

4. n = 20, s2= 5.25, 90% confidence level

5. n = 35, 𝞼2=1.25, 𝛼= 0.05

C.

1. A researcher wants to estimate the number of hours that a 5-year old child spend watching television. A
sample of 50 five-year old children was observed to have a mean viewing time of 3 hours. The population is
normally distributed with a population standard deviation of 0.5 hour, find:

a. the point estimate

b. the confidence interval at 𝛼= 0.05.

2. The mean scores of a random sample of 17 students who took a special examination in Electromagnetism
is 83.5. If the standard deviation of the scores is 4.1, find the:

a. point estimate and

b.confidence interval at 𝛼 = 0.01

PROBLEM SET:

1. Stephen has found out that the average weight of packed gummy candies selected from a normally
distributed population is 200 g. Given a random sample of 88 and population standard deviation of 3.4, use
95% confidence level and solve for the following:

a. point estimate;

b. margin of error; and

c. confidence interval/interval estimate.

2. Rochelle wants to know the mean of all entering trainees in a boot camp. The mean age of a random
sample of 25 trainees is 18 years and the standard deviation is 1.3 years. The sample comes from a normally
distributed population. Use 𝛼= 0.10 to find the following:

a. the point estimate;

b. the error/margin of error; and

c. the interval estimate/CI.

3. In an interview held among 18 randomly selected elementary teachers of a certain school, the mean
salary was found to be Php 21,000.00 with a standard deviation of Php 3, 000.00 Use 95% confidence to
estimate the population mean where the sample came from.

Find the following:

a. point estimate;

b. the margin of error (E); and

c. the interval estimate/CI.