Sampling and Sampling Distributions
Edwin Moyo (emoyo@[Link])
Edwin Moyo (emoyo@[Link]) Sampling and Sampling Distributions 1 / 20
Learning outcomes
This lecture develops work on sampling and sampling distributions. At the
end of this session, students will be able to;
Outcomes
1 Discuss different sampling methods
2 The concept of the sampling distribution
3 To compute probabilities related to the sample mean and the sample
proportion
4 The importance of the Central Limit Theorem
Edwin Moyo (emoyo@[Link]) Sampling and Sampling Distributions 2 / 20
Introduction
Sampling is a process of selecting items from a population so we can use
this information to make judgments or inferences about the population.
Recall: a sample is defined as the portion of a population that has been
selected for analysis.
Note: The sampling process begins by defining the frame, a listing of
items that make up the population. Frames are data sources such as
population lists, directories, or maps. Samples are drawn from frames.
Edwin Moyo (emoyo@[Link]) Sampling and Sampling Distributions 3 / 20
Introduction
Sampling is a process of selecting items from a population so we can use
this information to make judgments or inferences about the population.
Recall: a sample is defined as the portion of a population that has been
selected for analysis.
Note: The sampling process begins by defining the frame, a listing of
items that make up the population. Frames are data sources such as
population lists, directories, or maps. Samples are drawn from frames.
Reasons to Sample
To contact the whole population would be time-consuming.
The cost of studying all the items in a population may be prohibitive.
The physical impossibility of checking all items in the population.
The destructive nature of some tests.
The sample results are adequate.
Edwin Moyo (emoyo@[Link]) Sampling and Sampling Distributions 3 / 20
There are two types of samples: nonprobability samples and probability
samples.
In a nonprobability sample, you select the items or individuals without
knowing their probabilities of selection. It cannot be used for statistical
inference
In a probability sample, you select items based on known probabilities.
Probability samples allow you to make inferences about the population of
interest.
Edwin Moyo (emoyo@[Link]) Sampling and Sampling Distributions 4 / 20
There are two types of samples: nonprobability samples and probability
samples.
In a nonprobability sample, you select the items or individuals without
knowing their probabilities of selection. It cannot be used for statistical
inference
In a probability sample, you select items based on known probabilities.
Probability samples allow you to make inferences about the population of
interest.
Edwin Moyo (emoyo@[Link]) Sampling and Sampling Distributions 4 / 20
Methods of data collection
There are several methods of data collection but the most common are
questionnaires, personal interviews, telephone and computer technology
methods e.g CAPI, CATI, ACASI
Non probability Sampling
1 Convenience sampling, items selected are easy, inexpensive, or
convenient to sample. For example, if you were sampling tires stacked
in a warehouse, it would be much more convenient to sample tires at
the top of a stack than tires at the bottom of a stack.
Expert sampling involves the assembling of a sample of persons with
known or demonstrable experience and expertise in some area. Often,
we convene such a sample under the auspices of a ”panel of experts.”
In quota sampling, you select people nonrandomly according to some
fixed quota. There are two types of quota sampling: proportional and
non proportional.
2 Judgmental sample, you get the opinions of preselected experts in
the subject matter. Although the experts may be well informed, you
cannot generalize their results to the population.
Edwin Moyo (emoyo@[Link]) Sampling and Sampling Distributions 5 / 20
Probability Sampling
1 Simple random sampling:- When a sample of size n is drawn from a
population of size N in such a way that every item in the population
has the same probability of being selected e.g lottery.
2 Stratified random sampling:- Here the population is divided into
subgroups or strata (eg. male/female). Random samples are then
extracted from each stratum. This technique is used to reduce bias
and sampling error.
3 Systematic sampling:- This method is often used for convenience.
The technique involves ordering the population from 1 to Nth number
and selecting every ith number to be included in the sample.
4 Cluster (Area) Random Sampling:- This is used when we sample
data that is spread over a geographical area. In cluster sampling, we
follow these steps:
divide population into clusters (usually along geographic boundaries)
randomly sample clusters
measure all units within sampled clusters
Edwin Moyo (emoyo@[Link]) Sampling and Sampling Distributions 6 / 20
Survey Errors
Even when surveys use random probability sampling methods, they are
subject to potential [Link] are four types of survey errors:
1 Coverage error occurs if certain groups of items are excluded from
the frame so that they have no chance of being selected in the sample.
2 Nonresponse error arises from failure to collect data on all items in
the sample and results in a nonresponse bias.
3 Sampling error reflects the variation, or “chance differences,” from
sample to sample, based on the probability of particular individuals or
items being selected in the particular samples.
4 measurement error. Three sources of this error are ambiguous
wording of questions, the Hawthorne effect(pleasing interviewer), and
respondent error.
Edwin Moyo (emoyo@[Link]) Sampling and Sampling Distributions 7 / 20
Task
A simple random sample of n = 300 full-time employees is selected from a
company list containing the names of all N = 5, 000 full-time employees in
order to evaluate job satisfaction.
(a) Give an example of possible coverage error.
(b) Give an example of possible nonresponse error.
(c) Give an example of possible sampling error.
(d) Give an example of possible measurement error.
Edwin Moyo (emoyo@[Link]) Sampling and Sampling Distributions 8 / 20
Sampling distributions
A sampling distribution is the distribution of the results if you actually
selected all possible samples.
Sampling Distribution of the Mean
The sampling distribution of the mean is the distribution of all possible
sample means if you select all possible samples of a given size either with
replacement or without replacement.
Edwin Moyo (emoyo@[Link]) Sampling and Sampling Distributions 9 / 20
Sampling distributions
A sampling distribution is the distribution of the results if you actually
selected all possible samples.
Sampling Distribution of the Mean
The sampling distribution of the mean is the distribution of all possible
sample means if you select all possible samples of a given size either with
replacement or without replacement.
Expected value of x̄
E (x̄ ) = µ
where
E (x̄ ) = the expected value of x̄
µ = the population mean
Note: This result shows that with simple random sampling, the expected
value or mean of the sampling distribution of x̄ is equal to the mean of the
population.
Edwin Moyo (emoyo@[Link]) Sampling and Sampling Distributions 9 / 20
Note
When the expected value of a point estimator equals the population
parameter, we say the point estimator is unbiased.
Edwin Moyo (emoyo@[Link]) Sampling and Sampling Distributions 10 / 20
Note
When the expected value of a point estimator equals the population
parameter, we say the point estimator is unbiased.
Standard deviation of x̄
Finite Population r
N −n
σ
σx̄ = √
N −1 n
Infinite Population
σ
σx̄ = √
n
where:
σx̄ = the standard deviation of x̄
σ = the standard deviation of the population
n = the sample size
N = the population size
Edwin Moyo (emoyo@[Link]) Sampling and Sampling Distributions 10 / 20
The term
N −n
N −1
is called the finite population correction factor.
Note:
Use the following expression to compute the standard deviation of x̄
σ
σx̄ = √
n
whenever;
(1.) The population is infinite;
(2.) The population is finite and the sample size is less than or equal to
5% of the population size; that is, n/N ≤ .05.
Edwin Moyo (emoyo@[Link]) Sampling and Sampling Distributions 11 / 20
The term
N −n
N −1
is called the finite population correction factor.
Note:
Use the following expression to compute the standard deviation of x̄
σ
σx̄ = √
n
whenever;
(1.) The population is infinite;
(2.) The population is finite and the sample size is less than or equal to
5% of the population size; that is, n/N ≤ .05.
When choosing a sample with replacement, the total number of samples
is equal to N n and the total number of samples is equal to (Nn ) for
without replacement.
Edwin Moyo (emoyo@[Link]) Sampling and Sampling Distributions 11 / 20
Example
Consider the number of errors made by each of the administrative
assistants. This population distribution is shown below,
Administrative Assistant Number of Errors
Ann X1 = 3
Bob X2 = 2
Carla X3 = 1
Dave X4 = 4
Compute the mean and variance of the population. Take samples of size
two (2) from this population and calculate the mean and variance of the
sample means.
1 with replacement.
2 without replacement.
Edwin Moyo (emoyo@[Link]) Sampling and Sampling Distributions 12 / 20
Solution (1)
Thus, for the data,
3+2+1+4
µ= = 2.5 errors
4
and
r
(3 − 2.5)2 + (2 − 2.5)2 + (1 − 2.5)2 + (4 − 2.5)2
σ= = 1.12 errors
4
There are 16 possible samples i.e. N n = 42 = 16. The table below lists
the 16 possible sample outcomes. If you average all 16 of these sample
means, the mean of these values, is equal to 2.5, which is also the mean of
the population, µ.
The standard error of the mean, σx̄ is
√
σ 1.12 14 2
σx̄ = √ = √ =
n 2 25
Edwin Moyo (emoyo@[Link]) Sampling and Sampling Distributions 13 / 20
Sample Administrative Assistants Sample Outcomes Sample Mean
1 Ann, Ann 3,3 X̄1 = 3
2 Ann, Bob 3,2 X̄2 = 2.5
3 Ann, Carla 3,1 X̄3 = 2
4 Ann, Dave 3,4 X̄4 = 3.5
5 Bob, Ann 2,3 X̄5 = 2.5
6 Bob, Bob 2,2 X̄6 = 2
7 Bob, Carla 2,1 X̄7 = 1.5
8 Bob, Dave 2,4 X̄8 = 3
9 Carla, Ann 1,3 X̄9 = 2
10 Carla, Bob 1,2 X̄10 = 1.5
11 Carla, Carla 1,1 X̄11 = 1
12 Carla, Dave 1,4 X̄12 = 2.5
13 Dave, Ann 4,3 X̄13 = 3.5
14 Dave, Bob 4,2 X̄14 = 3
15 Dave, Carla 4,1 X̄15 = 2.5
16 Dave, Dave 4,4 X̄16 = 4
µx̄ = 2.5
Edwin Moyo (emoyo@[Link]) Sampling and Sampling Distributions 14 / 20
Solution (2)
Sample Administrative Assistants Sample Outcomes Sample Mean
1 Ann, Bob 3,2 X̄1 = 2.5
2 Ann, Carla 3,1 X̄2 = 2.0
3 Ann, Dave 3,4 X̄3 = 3.5
4 Bob, Carla 2,1 X̄4 = 1.5
5 Bob, Dave 2,4 X̄5 = 3.0
6 Carla, Dave 1,4 X̄6 = 2.5
µx̄ = 2.5
Therefore,
r r √
σ N −n 1.12 4−2 28 3
σx̄ = √ = √ =
n N −1 2 4−1 75
Edwin Moyo (emoyo@[Link]) Sampling and Sampling Distributions 15 / 20
Task (Done in class)
Find the mean µ and the variance σ2 of the population 1, 4, 7, 8. Draw up
a frequency distribution of the means of all possible samples of size 3,
taken without replacement. Find the mean and variance of this
distribution and verify that
σ2 N − n
Var (X̄ ) =
n N −1
where X̄ is the random variable ’the sample mean’, N is the number of
elements in the population and n is the sample size.
Edwin Moyo (emoyo@[Link]) Sampling and Sampling Distributions 16 / 20
Task (Done in class)
Find the mean µ and the variance σ2 of the population 1, 4, 7, 8. Draw up
a frequency distribution of the means of all possible samples of size 3,
taken without replacement. Find the mean and variance of this
distribution and verify that
σ2 N − n
Var (X̄ ) =
n N −1
where X̄ is the random variable ’the sample mean’, N is the number of
elements in the population and n is the sample size.
Sampling from Non-Normally Distributed Populations—The Central
Limit Theorem
The Central Limit Theorem states that as the sample size (i.e., the number
of values in each sample) gets large enough, the sampling distribution of
the mean is approximately normally distributed. This is true regardless of
the shape of the distribution of the individual values in the population.
Edwin Moyo (emoyo@[Link]) Sampling and Sampling Distributions 16 / 20
Sampling Distribution of the Proportion
The sampling distribution of p̄ is the probability distribution of all possible
values of the sample proportion p̄. For computing the sample proportion is
x
p̄ =
n
where
x = the number of elements in the sample that possess the characteristic
of interest
n = sample size
Edwin Moyo (emoyo@[Link]) Sampling and Sampling Distributions 17 / 20
Sampling Distribution of the Proportion
The sampling distribution of p̄ is the probability distribution of all possible
values of the sample proportion p̄. For computing the sample proportion is
x
p̄ =
n
where
x = the number of elements in the sample that possess the characteristic
of interest
n = sample size
Expected value of p̄
E (p̄ ) = p
where
E (p̄ ) = the expected value of p̄
p = the population proportion
Edwin Moyo (emoyo@[Link]) Sampling and Sampling Distributions 17 / 20
Standard Deviation of p̄
Finite Population
r r
N −n p (1 − p )
σp̄ =
N −1 n
Infinite Population r
p (1 − p )
σp̄ =
n
Note: The sampling distribution of p̄ can be approximated by a normal
distribution whenever np ≥ 5 and n (1 − p ) ≥ 5.
Edwin Moyo (emoyo@[Link]) Sampling and Sampling Distributions 18 / 20
Standard Deviation of p̄
Finite Population
r r
N −n p (1 − p )
σp̄ =
N −1 n
Infinite Population r
p (1 − p )
σp̄ =
n
Note: The sampling distribution of p̄ can be approximated by a normal
distribution whenever np ≥ 5 and n (1 − p ) ≥ 5.
Task
Assume that the population proportion is .55. Compute the standard error
of the proportion, σp̄ , for sample sizes of 100, 200, 500, and 1000 . What
can you say about the size of the standard error of the proportion as the
sample size is increased?
Edwin Moyo (emoyo@[Link]) Sampling and Sampling Distributions 18 / 20
Task
The following data represent the responses ( Y for yes and N for no) from
a sample of 40 college students to the question ”Do you currently own
shares in any stocks?” N N Y N N Y N Y N Y N N Y N Y Y N N N Y N
YNNNNYNNYYNNNYNNYNN
(a) Determine the sample proportion, p, of college students who own
shares of stock.
(b) If the population proportion is 0.30, determine the standard error of
the proportion.
Edwin Moyo (emoyo@[Link]) Sampling and Sampling Distributions 19 / 20
Edwin Moyo (emoyo@[Link]) Sampling and Sampling Distributions 20 / 20