Scribd
Upload
51 views58 pages

Sample Size Determination Methods

The document discusses methods for determining sample sizes for surveys and studies. It provides examples of calculating the sample size needed to estimate a population mean or proportion within a specified level of confidence and margin of error. The key factors in determining sample size are the confidence level, margin of error or precision level, and estimates of the population standard deviation or proportion. Higher confidence levels and smaller margins of error/precision levels require larger sample sizes. Sample size formulas provided are based on the normal distribution and z-scores.

Uploaded by

Gerald B. APOSTOL
Copyright
© All Rights Reserved
We take content rights seriously. If you suspect this is your content, claim it here.
Available Formats
Download as PPTX, PDF, TXT or read online on Scribd
51 views58 pages

Sample Size Determination Methods

The document discusses methods for determining sample sizes for surveys and studies. It provides examples of calculating the sample size needed to estimate a population mean or proportion within a specified level of confidence and margin of error. The key factors in determining sample size are the confidence level, margin of error or precision level, and estimates of the population standard deviation or proportion. Higher confidence levels and smaller margins of error/precision levels require larger sample sizes. Sample size formulas provided are based on the normal distribution and z-scores.

Uploaded by

Gerald B. APOSTOL
Copyright
© All Rights Reserved
We take content rights seriously. If you suspect this is your content, claim it here.
Available Formats
Download as PPTX, PDF, TXT or read online on Scribd

Engineering Data

Analysis
“How many participants
should be chosen for a
survey”?
Sample Size Determination

The sample size is typically


denoted by n and it is always a
positive integer.
Reminders:
Representativeness, not size,
is the more important
consideration.
Reminders:
Use no less than 30 subjects if
possible.

If you use complex statistics, you


may need a minimum of 100 or
more in your sample (varies with
method).
Methods in Determining the
Sample Size
Estimating the Mean or Average
The sample size required to estimate the
population mean µ to with a level of
confidence with specified margin
of error e, givenby
Z 2
n σ
Where: ≥ ( e )
Z is the z-score corresponding to level of
confidence e is the level of precision.
Example:
A soft drink machine is regulated so
that the amount of drink dispensed is
approximately normally distributed
with a standard deviation equal to 0.5
ounce. Determine the sample size
needed if we wish to be 95% confident
that our sample mean will be within
+.03 ounce from the true mean.
Z - Table
Solving for Z value
It was given that the confidence level is at
95%. ThusCL= 0.95
Solving for Z value
It was given that the confidence level is at
95%. ThusCL= 0.95
Solve for the are under the curve
Solving for Z value
It was given that the confidence level is at
95%. ThusCL= 0.95
Solve for the are under the curve
A = (1 + CL)/ 2 = ( 1 + 0.95 ) / 2 = 0.975
Solving for Z value
It was given that the confidence level is at
95%. ThusCL= 0.95
Solve for the are under the curve
A = (1 + CL)/ 2 = ( 1 + 0.95 ) / 2 = 0.975

-z 0 z
Solving for Z value
It was given that the confidence level is at
95%. ThusCL= 0.95
Solve for the are under the curve
A = (1 + CL)/ 2 = ( 1 + 0.95 ) / 2 = 0.975

-z 0 z
Solving for Z value
It was given that the confidence level is at
95%. ThusCL= 0.95
Solve for the are under the curve
A = (1 + CL)/ 2 = ( 1 + 0.95 ) / 2 = 0.975

-z 0 z
Solving for Z value
It was given that the confidence level is at
95%. ThusCL= 0.95
Solve for the are under the curve
A = (1 + CL)/ 2 = ( 1 + 0.95 ) / 2 = 0.975

-z 0 z
Solving for Z value
It was given that the confidence level is at
95%. ThusCL= 0.95
Solve for the are under the curve
A = (1 + CL)/ 2 = ( 1 + 0.95 ) / 2 = 0.975
0.95

-z 0 z
Solving for Z value
It was given that the confidence level is at
95%. ThusCL= 0.95
Solve for the are under the curve
A = (1 + CL)/ 2 = ( 1 + 0.95 ) / 2 = 0.975
0.95
+0.025

-z 0 z
Solving for Z value
It was given that the confidence level is at
95%. ThusCL= 0.95
Solve for the are under the curve
A = (1 + CL)/ 2 = ( 1 + 0.95 ) / 2 = 0.975
0.95 +0.025 = 0.975

-z 0 z
Solving for Z value
It was given that the confidence level is at
95%. ThusCL= 0.95
Solve for the are under the curve
A = (1 + CL)/ 2 = ( 1 + 0.95 ) / 2 = 0.975
0.95 +0.025 = 0.975

-z 0 z
Z - Table
Z - Table

A = 0.975
Z - Table

A = 0.975
Z - Table

A = 0.975
Z - Table

A = 0.975

1.9
Z - Table

A = 0.975

1.9
Z - Table

A = 0.975

1.9
+ 0.06
Z - Table

A = 0.975

1.9
+ 0.06
1.96
Z - Table

A = 0.975

1.9
+ 0.06
Z=1.96
Solution:
A soft drink machine is regulated so that the amount of
drink dispensed is approximately normally distributed
with a standard deviation equal to 0.5 ounce. Determine
the sample size needed if we wish to be 95% confident
that our sample mean will be within 0.03 ounce from the
true mean.
Solution:
A soft drink machine is regulated so that the amount of
drink dispensed is approximately normally distributed
with a standard deviation equal to 0.5 ounce. Determine
the sample size needed if we wish to be 95% confident
that our sample mean will be within 0.03 ounce from the
true mean.

The z – score for confidence level 95% in the z – table is 1.96.


Solution:
A soft drink machine is regulated so that the amount of
drink dispensed is approximately normally distributed
with a standard deviation equal to 0.5 ounce. Determine
the sample size needed if we wish to be 95% confident
that our sample mean will be within 0.03 ounce from the
true mean.

2 in the z – table is 1.96.


The z – score for confidence level 95%
1.96(0.5)
n = 1067.11
( 0.03 )
≥
Solution:
A soft drink machine is regulated so that the amount of
drink dispensed is approximately normally distributed
with a standard deviation equal to 0.5 ounce. Determine
the sample size needed if we wish to be 95% confident
that our sample mean will be within 0.03 ounce from the
true mean.

2 in the z – table is 1.96.


The z – score for confidence level 95%
1.96(0.5)
n = 1067.11
( 0.03 )
≥
We need a 1068 sample for our study.
Example:
The administration at a college wishes
to estimate, the proportion of all its
entering freshmen who graduate
wi thi n f o ur y e ars , wi th 9
5 % confidence. Estimate the
minimum size sample required.
Assume that the population standard
deviation is σ =
1.3 and precision level is 0.05.
Methods in Determining the
Sample Size
Estimating Proportion (Infinite Population)
The sample size required to obtain a confidence
interval for p with specified margin of error e
given by
is 2
Z
n≥( p(1 −
Where: e)
Z is the z-score
p) corresponding to level of
confidence e is the level of precision.
P is population proportion
Example:
Suppose we are doing a study on the
inhabitants of a large town, and want to find
out how many households serve breakfast
in the mornings. We don’t have much
information on the subject to begin with, so
we’re going to assume that half of the
families serve breakfast: this gives us
maximum variability. So p = 0.5. We want
99% confidence and at least 1% precision.
Solving for Z value
It was given that the confidence level is at
99%. ThusCL= 0.99
Solving for Z value
It was given that the confidence level is at
9%. Thus CL = 0.99
Solve for the are under the curve
Solving for Z value
It was given that the confidence level is at
9%. Thus CL = 0.99

Solve for the are under the curve


A = (1 + CL)/ 2
Solving for Z value
It was given that the confidence level is at
9%. Thus CL = 0.99

Solve for the are under the curve


A = (1 + CL)/ 2 = ( 1+
0.99 ) /2
Solving for Z value
It was given that the confidence level is at
9%. Thus CL = 0.99
Solve for the are under the curve
A = (1 + CL)/ 2 = ( 1 + 0.99 ) / 2 = 0.995
Z - Table

A = 0.995
Z - Table

A = 0.995
Z - Table

A = 0.995

2.5
Z - Table

A = 0.995

2.5
+
0.07
Z - Table

A = 0.995

2.5
+
0.0
7
2.5
7
Z - Table

A =
0.995

2.5
+
0.0
7
Z
=
2.5
Solution:
Suppose we are doing a study on the inhabitants of a large
town, and want to find out how many households serve
breakfast in the mornings. We don’t have much information on
the subject to begin with, so we’re going to assume that half of
the families serve breakfast: this gives us maximum variability.
So p = 0.5. We want 99% confidence and at least 1% precision.
Solution:
Suppose we are doing a study on the inhabitants of a large
town, and want to find out how many households serve
breakfast in the mornings. We don’t have much information on
the subject to begin with, so we’re going to assume that half of
the families serve breakfast: this gives us maximum variability.
So p = 0.5. We want 99% confidence and at least 1% precision.
The z – score for confidence level 99% in the z – table is 2.57.
Solution:
Suppose we are doing a study on the inhabitants of a large
town, and want to find out how many households serve
breakfast in the mornings. We don’t have much information on
the subject to begin with, so we’re going to assume that half of
the families serve breakfast: this gives us maximum variability.
So p = 0.5. We want 99% confidence and at least 1% precision.
The z – score for confidence level 99% in the z – table is 2.57.
2
2.57
n 0.5(1 − 0.5) =
≥ ( 0.01 ) 16,512.25
Solution:
Suppose we are doing a study on the inhabitants of a large
town, and want to find out how many households serve
breakfast in the mornings. We don’t have much information on
the subject to begin with, so we’re going to assume that half of
the families serve breakfast: this gives us maximum variability.
So p = 0.5. We want 99% confidence and at least 1% precision.
The z – score for confidence level 99% in the z – table is 2.57.
2
2.57
n 0.5(1 − 0.5) =
≥ ( 0.01 ) 16,512.25
We need a 16,513 sample for our study.
Example:
A dermatologist wishes to estimate the
proportion of young adults who
apply sunscreen regularly before
going out in the sun in the
summer. Find the minimum sample
size required to estimate the
proportion with precision of 3%, p
= 0.35 and 90% confidence.
Methods in Determining the
Sample Size
Slovin’s Formula (Simplified form of Proportions for
finite population)

Slovin’s formula is used to calculate the


sample size n given the population size and
error. It is computed as
N
n≥
1 + Ne2
Where:
N is the total population
e is the level of
Example:

A researcher plans to conduct a survey


about food preference of BioStat
students. If the population of students is
1000, find the sample size if the error is
5%.
Solution:
A researcher plans to conduct a survey about food
preference of BioStat students. If the population of
students is 1000, find the sample size if the margin
of error is 5%.

1000
n ≥ 1 + 1000(0.05)2 = 285.71

The researcher need to survey 286 BioStat students.


Example:
Using slovin's formula compute for n.
a. N = 2000 and e = 0.05.
b. N = 5000 and e = 0.05.
c. N = 10000 and e = 0.05.
What do you observe?
Computation of Sample size Using
Sample Size Online Calculator

[Link]
Computation of Sample size Using
Sample Size Online Calculator

[Link]

Common questions

Powered by AI

The Z-score used in sample size calculations is directly tied to the chosen confidence level. Higher confidence levels indicate larger Z-scores because a greater area under the curve must be accounted for, leading to larger sample sizes. For instance, a 95% confidence level corresponds to a Z-score of 1.96, while a 99% confidence level corresponds to a Z-score of 2.57 .

When determining sample size for estimating a proportion, maximum variability is assumed by choosing p = 0.5. This assumption maximizes the product p(1-p) and, consequently, the required sample size, ensuring precise estimation even with no prior proportion knowledge. This assumption illustrates generalization and represents the most uncertain scenario, thereby needing the largest sample size to achieve the desired level of confidence .

Increased precision in survey design requires narrower confidence intervals, leading to larger sample sizes. The precision level, denoted e, is inversely proportional to the sample size; lower e values mandate that sample results closely approximate the true population parameter, thus necessitating more samples to statistically reduce uncertainty and variability .

Standard deviation (σ) represents the population's variability and directly influences the calculated sample size for estimating means. Higher variability (i.e., larger standard deviation) results in increased uncertainty, thus requiring a larger sample size to meet the same confidence level and margin of error. This ensures that sample results are more likely to accurately reflect the true population parameter .

A sample size of no less than 30 is often recommended because, according to the Central Limit Theorem, the sampling distribution of the sample mean can be approximated to be normal if the sample size is 30 or more, regardless of the shape of the population distribution. This facilitates the application of normal distribuition-based inference methods, enhancing reliability and robustness of the results .

To determine the sample size for estimating a population mean with a specific level of confidence and margin of error, the formula n ≥ (Zσ/e)^2 is used, where Z is the z-score corresponding to the desired confidence level, σ is the population standard deviation, and e is the margin of error. For example, when estimating a soft drink machine's mean dispense amount with a standard deviation of 0.5 ounces and a desired 95% confidence level with a ±0.03 ounces margin of error, the sample size required is 1,068 .

Slovin’s formula calculates the sample size for finite populations using the formula n = N / (1 + Ne^2), where N is the total population size and e is the level of precision. This formula is used when the population's variability is unknown, providing a simple yet effective method to estimate an appropriate survey size based on both population size and desired precision .

In survey design, representativeness is more critical than sample size. Ensuring the sample adequately reflects the population's diversity is key for generalizable results. Additionally, the complexity of statistical analyses intended may necessitate larger samples, as complex methods require at least 100 subjects or more for valid and reliable results .

A sample size greater than 16,512 would be required when estimating the proportion of a large population with a high confidence level (e.g., 99%) and very high precision (e.g., 1%). This large sample size arises from the need to ensure the findings are extremely precise and reliable, accounting for maximum population variability, such as assuming p = 0.5 .

Choosing a higher confidence level in a college survey on graduation rates increases the sample size requirement significantly because it involves a higher Z-score, extending the confidence interval and demanding more data to maintain precision. For a 95% confidence level, the resulting larger sample size ensures that the observed proportion is more likely to closely mirror the actual graduation rates within the specified precision level, often compounded by the chosen level of variability .

You might also like