II.

SAMPLING & SAMPLING

DISTRIBUTION

MANAGERIAL
STATISTICS

UNITY UNIVERSITY; Management Department, TBH TBH: 2016 E.C 1st Semester CLASS
 GOALS
 Explain the rationale for a sampling, why it is the only
feasible way to learn about a population.
 Define sampling error
 Explain probability sampling & non-random sampling
methods
 Understand sampling distribution of the sample mean
& sample proportions
 Explain the central limit theorem.
 Use the concept of sampling distribution to find
probabilities of selecting possible sample means or
sample proportions from a specified population.
2
(

),
SAMPLING THEORY….
 Population is the total & Sample is part of the items under
study
 Parameter is any numerical summary measure calculated for
a population data. Some common parameters are like:
 Population mean (µ) e.g. the average age for getting
married for women in Ethiopia is 29;
 Population Proportion (P) e.g. 72% of all high school
students in Addis Ababa are female; Population Standard
deviation (σ) etc.
 Statistic is any numerical summary measure calculated for a
sample data. The measures include items like: sample mean
( ); sample proportion ( ); sample standard
3 deviation (s ), etc.
 MAJOR REASONS FOR SAMPLING
Why is sampling the only feasible way to learn about a
population.
 There are many reasons for sampling a population.
o The results of a sample may adequately estimate the
value of the population parameter, thus saving time and
money.
o It may be uneconomical too time-consuming & costly
to contact all members of the population.
o It may be impossible to check or locate all the members
of the population.
o Often testing destroys the sampled item and it cannot
be returned to the population.
4
 SAMPLING ERRORS
Sampling error is difference between a sample statistic and
its corresponding population parameter. It occurs because of
chance (only when the sample is taken randomly)
Non-sampling error is man made error during data
recording
Example: In a population of 5000 College students, 600 are
male. A sample of 120 college students selected from this
population randomly contains 18 male who possess the same
characteristic. What is the sampling error of the proportions?
Sampling error: /Statistic-Parameter /= /P- /=
/ 18/120 – 600/5000 /= /15%-12%/=3%
If the sample proportion is recorded as 13% wrongly, the non
sampling error is the difference between the correct & wrong
sample proportions i.e. 15%-13%=2%
5
 Random (Probability) Sampling Methods

 It is unbiased & representative, every members of the

population have a known non-zero chance of being selected
into the sample. There are several probability (random)
sampling methods.
 In a simple random sample, all members of the population
have the same chance of being selected for the sample.
 In a systematic sample, a random starting point is selected,
and then every kth item thereafter is selected for the sample.
 In a stratified sample, the population is divided into several
groups, called strata, and then a random sample is selected
from each stratum.
 In cluster sampling, the population is divided into primary units
(clusters), then samples are drawn from the primary units.
6
 How do you apply random sampling?
Generally, first get the full list of the population
(Sampling frame)
1. Simple random sampling:
 The lottery method (small population):
 Each item is numbered and a table of random
numbers (RNT) is used to select the members of the
sample, Blindfolding oneself point anywhere on the
table, using this point as starting #, pick the #s that
are in the list of the population until the desired
sample size is obtained (by moving on the table as per
the predetermined pattern)
7 computer programs generating random numbers
 How do you apply random sampling?

2. Systematic random sampling:

Select from term papers of 24 students 4 papers for
presentation using systematic random sampling.
 Items of the population are arranged in some way
(alphabetically, date received, etc.) are numbered;
 compute K (sampling interval), k= N/n = K=24/4=6
 select random number as starting point between 1
and K (from items in range 1-6), Starting with this
number select every Kth (6th)number until all the n
units (4) are selected. let’s say 2 was selected
randomly as a starting point; students 8, 14, 20 are
8 then selected; by picking at an interval of 6
 How do you apply random sampling?

3. Stratified Random Sampling is suitable when

the population is heterogeneous concerning the
variable under study. The population is divided into
homogeneous subgroups /strata, and the intended
sample is selected proportionally or non-
proportionally from each strata randomly.

9
 How do you apply random sampling?
Suppose a study involved College students, we might decide to
stratify the population (all college students) into freshman,
sophomore, and senior. Suppose the sample size to be taken is
n=50; and the population of college students N=1000. Let the
population of students in each strata be 500, 300, and 200
respectively. Select the sample using proportionate stratified
random sampling
N.B. If sample is proportionally taken from each strata it is done
as follows: n/N * Xi; Xi is the portion of the population in each
strata. Sample proportion=50/1000=0.05 or 5%
Freshman= 500*5%=25; Sophomore=300*5%=15;
Senior= 200*5%=10
The identified sizes of the sample will be selected from each
10 strata randomly
 How do you apply random sampling?

3. Cluster (Area) Sampling It is helpful when the elements in the

population that are studied are geographically scattered over large
area and generally have related characteristics concerning the variable
under study. The population is divided into clusters; each cluster is a
primary unit that is a small representation of the big population.
Suppose the objective is to study household waste collection in
Bole subcity. The city woredas or kebeles in the subcity can be
taken as clusters. Then the woredas are numbered and some
clusters are selected randomly. Finally, households in the
selected woredas are entirely surveyed or only randomly
selected households are surveyed. N.B.: The researcher may
go for multistage sampling.

11
 Non-probability (non-random) Sampling methods

 They are not representative; they are rather opportunistic.

Non-probability sampling methods:
o Convenience Sampling: The criteria to select sample is
accessibility and ease
o Judgment (Purposive) Sampling: Expert opinion is the basis
to select sample.
o Quota Sampling: sample quota will be set for each group
in the population
o Snowball Sampling: sample elements are identified via
referral or networks
 It is mandatory in inferential statistics to use only probability
sampling for valid conclusions. Non-probability sampling is
suitable for pilot studies and exploratory research.
12
 Sampling Distributions
 Sampling distribution of the sample mean is a probability
of distribution of (all possible) sample means of a given sample
size selected from a population randomly
 The sampling distribution is normal if the parent population is
normal for any sample size; but if the parent population is not
normal the sampling distribution will be approximately normal
for n≥30, according to the central limit theorem.
 Properties of Sampling distribution of sample mean (if we take
repeated random samples of size n from a population), Given:
o The mean and standard deviation (standard error) of the sampling
distribution of the sample mean respectively µx and σx
o Parent Population mean and standard deviation respectively: µ, σ
o N=Parent population size; n=random sample size
=µ ; =
13
 Sampling Distributions of Sample Means

Example: Suppose 5 students in a class missed test out of 10

and have sat for make-up test, the following is the summary of
their result. As the data set represents all students that sat for
make-up test, it is a population distribution (refer to Table
below). If a random sample of size (n=2) is taken without
replacement from population of size (N=5); answer the
following:

Students Abas Betty Chaltu Dan Eleni

(M) (F) (F) (M) (F)
Make up score 2 4 6 8 10
/10

14
 Sampling Distributions of Sample Means

A. What is the number of all possible random samples that can be

taken from the population without replacement?
 One can take (all in all) (Ncn random samples) without
replacement, thus, 5c2 samples i.e. 10 samples are taken
exhaustively.
B. List all possible samples of two students from the population
and compute their means.
Sample Abas, Abas, Abas, Abas, Betty, Betty, Betty, Chaltu, Chaltu, Dan,
size 2 Betty Chaltu Dan Eleni Chaltu Dan Eleni Dan Eleni Eleni
Elements 2,4 2,6 2,8 2,10 4,6 4,8 4,10 6,8 6,10 8,10
Sample 3 4 5 6 5 6 7 7 8 9
Means

15
 Sampling Distribution of Sample proportions

 It is a probability of distribution of (all possible) sample

proportions( ) of a given sample size selected from a
population randomly
 Properties of Sampling distribution of sample proportion (if we
take repeated random samples of size n from a population)
E ( ) =P and =
When n/N≥5%,
= *

 The sampling distribution of the proportion is

approximately normally distributed for large samples
16
i.e. when (n.P & n.Q ≥5),
 Sampling Distribution of Sample proportions

 Example: Develop sampling distribution of sample proportions

for female (gender) by taking a random sample of 4 students
from the population of 5 without replacement.
 Ncn=5c4=5, thus there are a total of 5 possible random samples
Samples A,B,C,D A,B,C,E A,B,D,E A,C,D,E B,C,D,E
Gender M,F,F,M M,F,F,F M,F,M,F M,F,M,F F,F,M,F
Sample 2/4=0.5 1/4=0.25 2/4=0.5 2/4=0.5 1/4=0.25
Proportions (𝐩)
Sample 0.25 0.5
Proportions (𝐩)

Probability 2/5=0.4 3/5=0.6

P(𝐩)

E( )=P=0.4; = * =0.2449*0.5=0.1224
17
 Sampling Distribution Examples

1. The grade of all students on a test for a given course

i.e. marked out of 10 is normally distributed with a
mean of 6 and a standard deviation of 2. For the
sampling distribution of mean based on a random
sample of 4 students; what is the probability that a
random sample mean of students score is at least 8 or
what proportion of the random samples show mean
test score of at least 8? P(Sample mean ≥ 8)
Z= = (8-6) / (2 / ) =2;
Z ≥ 2=0.5-0.4772=0.0228
18
 Sampling Distribution Examples

2. It is known that 65% of all American voters favor the

democrats. 100 American voters were randomly
sampled. What is the probability that 60% or less of
the sampled voters favor the democrats?
Z=0.6-0.65/ 𝟎. 𝟔𝟓 ∗ 𝟎. 𝟑𝟓 𝟏𝟎𝟎 =-.05/.0477=-1.05;

P(𝐩)<0.6)=P(Z<-1.05)=.5-.3531=0.1469

-1.05 0
19
 END OF SYNOPSIS

Question
and
Answers

20