Sampling and Sampling Distribution

The concept of sampling

Sampling unit: the ultimate unit to be sampled or elements of the
population to be sampled.
Examples: If somebody studies the Socioeconomic status of the
households. The sampling unit is households
If one studies the performance of freshman students in some college.
The sampling unit is the student
Sampling frame: is the list of all elements in a population.
Examples:
List of households.
List of students in the registrar’s office.
Abriham S.
Generalizability: the extent to which inferences drawn from a sample
are true for the population.
Homogenous population: a population whose every element is
similar regarding the variables required for the research.
Heterogeneous population: a population the elements of which are
not similar to each other.
Sampling bias: a situation where the selected sample does not truly
reflect the characteristics of the population.
Systematic errors: the errors that are caused by over or under-
representation of some characteristics of the population in the sample.

Abriham S.
 A main concern in sampling:
 Make sure that the sample represents the population, and
 The findings can be generalized.
Advantages of sampling:
Feasibility: Sampling may be the only feasible method of collecting
information.
Reduced cost: Sampling reduces demands on resources such as finance,
personnel, and material.
Greater accuracy: Sampling may lead to better accuracy in collecting data
Greater speed: Data can be collected and summarized more quickly resources

Abriham S.
 Error in Sampling
o No sample is the exact mirror image of the population

o Can not be avoided or eliminated

o The chance and random variation in variables that occurs when any
sample is selected from the population

o Sampling error is to be expected

o To avoid sampling error, a census of the entire population must be

taken

o To control for sampling error, researchers use various sampling

methods
Abriham S.
 Types of Errors in Sampling
Sampling error: the discrepancy b/n the population value & sample value.
o is any type of bias that is attributable to mistakes in either drawing a sample
or determining the sample size.
o the uncertainty associated with an estimate that is based on data gathered
from a sample of the population rather than the full population.
Non-sampling errors: are errors due to procedure bias:
o Observational error
o Respondent error
o Lack of preciseness of definition
o Errors in editing and tabulation of data
Abriham S.
There are two types of sampling methods.

I. Probability Sampling Methods

II. Non-Probability Sampling Methods

Abriham S.
 PROBABILITY SAMPLING

o Any method of sampling that utilizes some form of random selection.

o Involves random selection of a sample

o Every sampling unit has a known and non-zero probability of

selection into the sample

o Involves the selection of a sample from a population based on chance

o Probability sampling is:

◦ more complex,

◦ more time-consuming and

◦ usually more costly than non-probability sampling

Abriham S.
Most common probability sampling methods

1. Simple random sampling

2. Systematic random sampling

3. Stratified random sampling

4. Cluster sampling

Abriham S.
 Simple Random Sampling(SRS)
o It is the technique of selecting samples in which every individual or
item of the population has an equal chance of being selected into the
sample. Which item or individual gets selected is purely a matter of
chance. The personal bias of the investigator does not influence the
selection.

To ensure the randomness of the selection we can use the:

Lottery method

Table of random numbers

Abriham S.
Lottery Method: If the size of the population is relatively small, we may
number or write their names on separate pieces of paper of identical size and
shape.

These papers are identically folded and mixed up in a container.

The required items are picked up from the container one after the other. Each
piece of paper has the same chance of being picked up until the desired sample
size has been taken.

Random number table: can be used to select a representative sample from a

population of a large size.

o Is one in which every digit (0, 1, 2, …, 9) has been entered randomly. The
chance of getting any one of the ten digits (0, 1, 2, …, 9) would be the same
and exactly 1/10.
Abriham S.
o The occurrence of any number in one part of the table is independent
of the occurrence of any number in any other part of the table.

o In selecting a random sample from a finite population, we first assign

a code number to each member of the population.

Example 1: If there are 100 members in the population, then each

member will be assigned a code number 00, 01, 02, 03, 04, 05, …, 99.
we want to select a random sample of 10 members from this population.
Now let us a random number to select our 10 samples.

Example 2: Let us assume that there are 2000 students at Rift Valley
University arranged in alphabetical order. We want to a sample of 200
students out of these 2000 to study their average age.
Abriham S.
 Systematic Random Sampling
o Sometimes called interval sampling
o Selection of individuals from the sampling frame systematically rather than
randomly
o Individuals are taken at regular intervals down the list
o The starting point is chosen at random
Steps in systematic random sampling
 Number the units in the population from 1 to N
❖Decide on the n (sample size) that you want or need
❖k = N/n = the interval size
❖Randomly select an integer between 1 to k
❖Then, take every 𝑘 𝑡ℎ unit
Abriham S.
Example: N=1200, and n=60

1200
Sampling fraction (k)= =20
60

 List persons from 1 to 1200

 Randomly select a number between 1 and 20

 Let the first person selected = the 8𝑡ℎ on the list

 The 2𝑛𝑑 person =8+20 = the 28𝑡ℎ

 The 3𝑟𝑑 person = 28+20= the 48𝑡ℎ

 The 4𝑡ℎ person = 48+20= the 68𝑡ℎ … etc.

Abriham S.
Abriham S.
 Stratified Random Sampling
A method of probability Sampling in which the population is divided
into different subgroups and samples are selected from each subgroup.

Some of the criteria for dividing a population into strata are Sex (male,
female); Age (under 18, 18 to 28, 29 to 39); income, etc.

These subgroups are homogeneous and mutually exclusive groups

called strata

Abriham S.
 Divide the population into non-overlapping groups (i.e., strata)
N1 , N2 , … Ni such that N1 +N2 + ⋯ + Ni = N

A separate sample is taken independently from each stratum

depending on the type of allocation

Elements within each strata are homogeneous but are

heterogeneous across strata.

A simple random or a systematic sample is taken from each strata

Proportional allocation: allocating sampling proportional to the total

population of each strata using the formula:
n
ni = *Ni
N Abriham S.
Where n=total sample size to be selected

N=total population

Ni = total population of each strata

ni=sample size from each strata

Exercise: find proportional allocation for the following villages

Village A B C D Total

HHs 100 150 120 130 500

Sample size ? ? ? ? 60

Abriham S.
 Cluster Sampling
• The population is divided into non-overlapping groups called clusters.
• A simple random sample of groups or clusters of elements is chosen
and all the sampling units in the selected clusters will be surveyed.
• Clusters are formed in a way that elements within a cluster are
heterogeneous, i.e. observations in each cluster should be more or less
dissimilar.
• Cluster sampling is useful when it is difficult or costly to generate a
simple random sample.
Example: to estimate the average annual household income in a large
city we use cluster sampling
A less expensive way is to let each block within the city represent a
cluster. A sample of clusters could then be randomly selected, and every
household within these clusters could be interviewed to find the average
annual household income.
Abriham S.
Abriham S.
Abriham S.
 Non-Random Sampling
 It is a sampling technique in which the choice of individuals for a
sample depends based on convenience, personal choice, or interest.
 non-probability sampling does not involve random selection
 In non-probability sampling, every item has an unknown chance of
being selected
 They are quick, inexpensive and convenient
The most common types of non-probability sampling
1. Convenience sampling
[Link] sampling
3. Judgment or Purposive sampling
4. Quota sampling
5. Snowball sampling

Abriham S.
 Convenience Sampling
• is sometimes referred to as haphazard/accidental sampling.
• subjects are selected based on their availability and simplicity of access.
• Choosing participants who are readily available to the researcher.
• Easy to use, but that advantage is greatly offset by the presence of bias.
Example
A researcher wants to study the opinions of college students about a new
campus policy. They might distribute surveys to students in their own
classes or in the student union.
Volunteer Sampling
• individuals volunteer to participate in a study.
Example
A researcher studying online learning experiences might post a survey on
social media and invite people to participate voluntarily.
Abriham S.
 Judgment Sampling

• Researchers select participants based on their judgment and expertise.

• Researcher believes that the selected individuals are most likely to

provide the desired information.
• Used exploratory studies like pre-testing of questionnaires and focus groups.
• One advantage of judgment sampling is the reduced cost and time
involved in acquiring the sample.
Example

A researcher studying the impact of a new educational policy might select a few
experienced educators and policymakers as participants in an interview study.

Abriham S.
 Quota Sampling

• Researchers divide the population into subgroups or strata based on

specific characteristics (e.g., age, gender, race, income).
• Quotas for each subgroup and select participants until the quota is filled.

• the population has no suitable frame

Example
A researcher wants to conduct a survey on college students' attitudes
towards online learning. They might divide the population into four strata
based on year of study (freshman, sophomore, junior, senior) and set
quotas for each stratum.

Abriham S.
 Snowball Sampling

• The participants are recruited through referrals from existing

participants.

• Useful when the target population is hard to reach or identify.

Example:

A researcher studying a rare disease might start by identifying a few

individuals with the condition. These individuals can then refer other
people they know who have the same condition.

Abriham S.
 Sample Size Determination
• Sample size is the number of study subjects selected to represent a
given study population

• Should be sufficient to represent the characteristics of interest of the

study population

Abriham S.
 Sample size depends on:

• Objective of the study

• Design of the study

• Plan for statistical analysis

• Accuracy of the measurement to be made

• Degree of precision required for generalization

• Degree of confidence

Abriham S.
 Sample Size for Single Proportion
For making a confidence limit statement the following formula can be
used to estimate the minimum sample size:

2
Zα pq
2
n=
d2
Where
Z refers reliability coefficient (confidence level)
P the proportion in the population possessing the characteristic of
interest,
d margin of error
Abriham S.
To get P we have three options

o Doing pilot study

o Literature

o Take a value of 0.5

The margin of error (d) measures the precision of the estimate

o Small value of d indicates high precision

o For P close to 50%, d is assumed to be close to 5%

o For smaller value of p, d is assumed to be lower than 5%

Abriham S.
Abriham S.
 Parameters in the formula
o n is the minimum sample size
o P is an estimate of the prevalence rate for the population
• From available data or Pilot study results, or 0.5 should be used to
get the possible minimum large sample size; if given in range, take
the value closest to 0.5.
o d is the margin of sampling error tolerated
o Zα/2 is the standard normal variable at (1-α )% confidence
level and α is mostly taken to be 5%
• Usually 95% confidence level is used = 1.96
o N population size

Abriham S.
 Sampling Distribution
A sampling distribution is a distribution of all possible values of a statistic
computed from samples of the same size randomly selected from the
same population.

Let N is Population size, n is sample size, suppose simple random

sampling is used
Sampling is with replacement
1st draw = N, 2nd draw = N, 3rd draw = N, … 𝑛𝑡ℎ draw = N

We have 𝐍 𝐧 possible samples

𝝈𝟐
𝝈𝒙 𝟐 =
𝒏
Abriham S.
 Sampling is without replacement

1st draw = N, 2nd draw = N-1, 3rd draw = N-2, …,

𝑛𝑡ℎ draw = (N-n+1)

𝐍
We have 𝐧
possible samples

𝛔𝟐 𝐍 − 𝐧
𝛔𝐱 𝟐 =
𝐧 𝐍−𝟏

Abriham S.
 Sampling Distribution of the sample mean
Sampling distribution of the sample mean is a theoretical probability
distribution that shows the functional relation ship between the possible
values of a given sample mean based on samples of size and the
probability associated with each value, for all possible samples of size
drawn from that particular population.

Example: Suppose we have a population of size N=4, constituting the

ages in years of four outpatients are, 18, 20, 22, 24

Abriham S.
 Xi 18 + 20 + 22 + 24
μ= = = 21
N 4

Xi −μ 2 (18−21)2 +(20−21)2 +(22−21)2 +(24−21)2

σ= = = 2.236
N 4

Construct sampling distribution of the sample mean of size 2 using sampling

without replacement and with replacement.

SWOR N n =42 =16 possible samples 16 sample means

(18,18) (18,20) (18,22) (18,24) 18 19 20 21

(20,18) (20,20) (20,22) (20,24) 19 20 21 22

(22,18) (22,20) (22,22) (22,24) 20 21 22 23

(24,18) (24,20) (24,22) (24,24) 21 22 23 24

Abriham S.
 𝑥𝑖 18 + 19 + … + 24
𝜇𝑥 = = = 21
𝑛 16

𝜎 2 2.236 2
𝜎𝑥 2 = = = 2.499848
𝑛 2
Table xx sampling distribution of the sample mean

Sampling is without replacement.

N 4
We have n
= 2
= 6 possible samples

Abriham S.
Exercise: Suppose we have a hypothetical population of size 3,
consisting of three children age in year: 3, 6 and 9. Construct sampling
distribution of the sample mean of size 2 using sampling without
replacement and with replacement.

a) Show population mean equal Sample mean

b) Find the population variance and Sample variance

Abriham S.
 Central Limit Theorem (CLT)

If X1 , X2 , … Xn is a random sample from a population with mean µ and

variance 𝜎 2 , then as n goes to infinity the distribution of the sample mean,
𝜎2
X, approximates normal distribution with mean µ and variance .
𝑛
𝜎2
In short as n gets large number, X ~𝑁 𝜇,
𝑛

Abriham S.
Abriham S.