Assignment 9: Quantitative Methods-II (BSAF – HCC)

Sampling
Sampling is a procedure where we select (usually randomly) a particular number of observations
from a larger number of observations called population. Hence sample is a subset of the
population.
Sampling can be with or without replacement. Sampling with replacement means that we
reconsider (put back) the selected observation so the already selected observation can be
selected again. In sampling without replacement, the selected observation cannot be selected
again as we do not reconsider or put back the selected observation before selecting the next
observation.
In sampling with replacement, the events of selection are independent of each other and have
the same probability to be selected every time.
In sampling without replacement, the event of selection do not remain independent of each
other. When an item is selected, the next selection considers the condition that the first item has
been selected so the probability of selection does not remain the same.
In Statistics, samples are used to infer about the population. We use probability theory for this
purpose. This means that we believe that the sample exhibits properties of the population from
where the sample has been drawn.
The relationship of mean and standard deviation of a distribution of sample means is shown
below: (N = number of items in the population, n = sample size)
Sampling with replacement Sampling without replacement
Number of samples 𝑁𝑛 𝑁 𝑁!
𝐶𝑛 = (𝑁−𝑛)!𝑛! (If order does not matter)
𝑁 𝑁!
𝑃𝑛 = (𝑁−𝑛)! (If order matters)

Mean 𝜇 = 𝜇𝑋̅ 𝜇 = 𝜇𝑋̅

Standard Deviation 𝜎
𝜎𝑋̅ = 𝜎 𝑁−𝑛
√𝑛 𝜎𝑋̅ = .√
√𝑛 𝑁−1

Example: Sampling with replacement

Consider the population of five numbers 1, 2, 3, 4, 5
Here number of items in the population = N = 5
If we want to draw samples of size 2 with replacement then n = sample size = 2, number of
samples 𝑁 𝑛 = 52 = 25
We can write all possible samples of size 2 with replacement as given below (numbers can
repeatedly occur)

Samples Sample means = 𝑋̅ 𝑋̅ 2 Mean and Standard Deviation of sample means

1,1 1 1 If we call 𝑁 𝑛 = 𝑚 then
∑ 𝑋̅ 75
1,2 1.5 2.25 Mean of means = 𝜇𝑋̅ = = 25 = 3
𝑚
1,3 2 4 Standard Deviation of means =
1,4 2.5 6.25
∑ 𝑋̅ 2 ∑ 𝑋̅ 2 250 75 2
𝜎𝑋̅ = √ − ( 𝑚 ) = √ 25 − (25) = 1
1,5 3 9 𝑚

2,1 1.5 2.25

Mean and Standard Deviation of the Population
2,2 2 4
𝑋 𝑋2
2,3 2.5 6.25
1 1
2,4 3 9
2 4
2,5 3.5 12.25
3 9
3,1 2 4
4 16
3,2 2.5 6.25
5 25
3,3 3 9
15 55
3,4 3.5 12.25
3,5 4 16 ∑𝑋 15
Mean = 𝜇 = = = 3 (= 𝜇𝑋̅ )
4,1 2.5 6.25 𝑁 5

4,2 3 Standard Deviation of the population =

9
4,3 3.5 12.25 ∑ 𝑋2 ∑𝑋 2 55 15 2
𝜎=√ − ( 𝑁 ) = √ 5 − ( 5 ) = √2
𝑁
4,4 4 16
We can see that
4,5 4.5 20.25
𝜎 √2
5,1 3 9 =
= 1 = 𝜎𝑋̅
√𝑛 √2
5,2 3.5 12.25 Hence, we prove that for means of samples with
5,3 4 16 replacement,
𝜎
5,4 4.5 20.25 𝜎𝑋̅ =
√𝑛
5,5 5 25
Sums 75 250
Sampling Distribution of Sample Means
An easy way to prove the above relationship between sample mean and standard deviation and
the population mean and standard deviation is to construct a sampling distribution of sample
means as follows: (write the sample means once and write the number of times they occur i.e.
the frequency next to them)
Sample Means
Frequency
1 1
1.5 2
2 3
2.5 4
3 5
3.5 4
4 3
4.5 2
5 1
25
Note: We can find the mean and variance of the sample means (as we did above) by using this
distribution and formula with frequency.

Example: Sampling without replacement

Here we will use the same example as above to draw samples without replacement in two ways.
First, samples without replacement where order does not matter (‘a,b’ is the same as ‘b,a’) and
then where the order matters (so we need to include both ‘a, b’ and ‘b,a’ separa in our samples)

Consider the population 1, 2, 3, 4, 5 (N=5)

Let Sample size be 2 (n=2)
Number of samples will be as follows:

If order does not matter

𝑁 𝑁! 5! 5.4..3.2.1
𝐶𝑛 = (𝑁−𝑛)!𝑛! = (5−2)!2! = (3.2.1)(2.1) = 10 (remember that a dot ‘.’ Means multiplication)

If order matters
𝑁 𝑁! 5! 5.4..3.2.1
𝑃𝑛 = (𝑁−𝑛)! = (5−2)! = = 20
3.2.1

Note that the number of possible samples in the second case is larger as we need to include
possibilities of the same set of numbers as the order changes (e.g. 1,2 and 2,1 will be treated as
different samples in the second case when order matters)
Mean and Standard Deviation of the Population
∑𝑋 15
𝑋 𝑋2 Mean of means = 𝜇 = = = 3 (= 𝜇𝑋̅ )
𝑁 5
1 1
2 4
Standard Deviation of the population =
3 9
∑ 𝑋2 ∑𝑋 2 55 15 2
4 16 𝜎=√ − ( 𝑁 ) = √ 5 − ( 5 ) = √2 = 1.4142
𝑁

5 25
15 55

Samples (Without replacement)

Below are all possible samples without replacement in both cases.
Sampling without replacement (when order does not matter)
Sample 𝑋̅ 2 Number of samples
Samples
means = 𝑋̅ 𝑁
𝑁! 5! 5.4. .3.2.1
𝐶𝑛 = = = = 10
1, 2 1.5 2.25 (𝑁 − 𝑛)! 𝑛! (5 − 2)! 2! (3.2.1)(2.1)
1, 3 2 4
Mean and Standard Deviation of sample means
1, 4 2.5 6.25
1, 5 3 9 If we call the number of samples = 𝑚 then
2, 3 2.5 6.25 ∑ 𝑋̅ 30
Mean of means = 𝜇𝑋̅ = = 10 = 3
𝑚
2, 4 3 9
2, 5 3.5 12.25 Standard Deviation of means =
3, 4 3.5 12.25 ∑ 𝑋̅ 2 ∑ 𝑋̅ 2 97.5 30 2
3, 5 4 16 𝜎𝑋̅ = √ − ( 𝑚 ) = √ 10 − (10) = √0.75 = 0.86603
𝑚
4, 5 4.5 20.25
30 97.5
Note: See that we do not repeat any number and we do not write any sample in reverse order (e.g. if 1,2 is already
selected, we should not write 2,1 as the order does not matter)

Hence = 𝜇 = 𝜇𝑋̅ = 3 (Population mean and mean of the sample means are equal)
Now 𝜎𝑋̅ = 0.86603 and the population standard deviation 𝜎 = 1.4142
𝜎 𝑁−𝑛
We can check if 𝜎𝑋̅ = . √ 𝑁−1 holds true.
√𝑛

𝜎 𝑁 − 𝑛 1.4142 5−2
.√ = .√ = √0.75 = 0.86603
√𝑛 𝑁−1 √2 5−1

𝜎 𝑁−𝑛
Hence 𝜎𝑋̅ = . √ 𝑁−1
√𝑛
Sampling without replacement (when order matters)

Sample Number of samples

Samples means 𝑋̅ 2 𝑁
𝑁! 5! 5.4. .3.2.1
𝑃𝑛 = = = = 20
= 𝑋̅ (𝑁 − 𝑛)! (5 − 2)! 3.2.1
1, 2 1.5 2.25
1, 3 2 4 Mean and Standard Deviation of sample means
1, 4 2.5 6.25 If we call the number of samples = 𝑚 then
1, 5 3 9
∑ 𝑋̅ 60
2, 1 1.5 2.25 Mean of means = 𝜇𝑋̅ = = 20 = 3
𝑚
2, 3 2.5 6.25
2, 4 3 9 Standard Deviation of means =
2, 5 3.5 12.25 ∑ 𝑋̅ 2 ∑ 𝑋̅ 2 195 60 2
3, 1 2 4 𝜎𝑋̅ = √ − ( 𝑚 ) = √ 20 − (20) = √0.75 = 0.86603
𝑚
3, 2 2.5 6.25
3, 4 3.5 12.25
Hence = 𝜇 = 𝜇𝑋̅ = 3 (Population mean and mean of the
sample means are equal)
3, 5 4 16
4, 1 2.5 6.25 Now 𝜎𝑋̅ = 0.86603 and the population standard deviation
4, 2 3 9 𝜎 = 1.4142
4, 3 3.5 12.25
𝜎 𝑁−𝑛
4, 5 4.5 20.25 We can check if 𝜎𝑋̅ = . √ 𝑁−1 holds true.
√𝑛
5, 1 3 9
5, 2 3.5 12.25 𝜎 𝑁 − 𝑛 1.4142 5−2
5, 3 4 16 .√ = .√ = √0.75 = 0.86603
√𝑛 𝑁−1 √2 5−1
5, 4 4.5 20.25
60 195
Note: As order matters here, if we select 1,2, we do select 2,1 later. However, a number cannot be repeated in a
sample)

This shows that the relationship of means and standard deviations of the population and the
distribution of sample means hold in both cases (when order does not matter as well as when
order matters)
Questions to be solved
For the list of populations given below:
1. Perform the following:
a. Draw all possible samples of size 2 with replacement
b. For the sample mean prove that
i. 𝜇 = 𝜇𝑋̅ and
𝜎
ii. 𝜎𝑋̅ =
√𝑛
c. Construct a sampling distribution of sample means and prove the properties
mentioned above for this new grouped data
2. Perform the following:
 a. Draw all possible samples of size 2 without replacement (when order does not
matter)
b. For the sample mean prove that
i. 𝜇 = 𝜇𝑋̅ and
𝜎 𝑁−𝑛
ii. 𝜎𝑋̅ = . √ 𝑁−1
√𝑛
c. Construct a sampling distribution of sample means and prove the properties
mentioned above for this new grouped data
3. Perform the following:
a. Draw all possible samples of size 2 without replacement (when order matter)
b. For the sample mean prove that
i. 𝜇 = 𝜇𝑋̅ and
𝜎 𝑁−𝑛
ii. 𝜎𝑋̅ = . √ 𝑁−1
√𝑛
c. Construct a sampling distribution of sample means and prove the properties
mentioned above for this new grouped data

List of populations
I. 2, 4, 6, 8, 10
II. 1, 2, 3, 4, 5, 6
III. 3, 6, 9, 12,15
IV. 1, 2, 3, 4, 5, 6, 7
V. 1, 5, 9, 10, 12