Sampling and sample

size determination

Introduction
Error in sampling
Samepling method
Sample size calculation
♣ Researchers often use sample survey methodology to
obtain information about a larger population by
selecting and measuring a sample from that population.

♣ Since population is too large, we rely on the information

collected from the sample.

♣ Inferences about the population are based on the

information from the sample drawn for the
population.
♣ However, due to the variability in the characteristics of
the population, scientific sampling designs should be
applied to select a representative sample.

♣ If not, there is a high risk of distorting the view of the

population.

♣ Sampling is a procedure by which some

members of the given population are selected as
representative of the entire population.
Sampling involves the selection of a number of study unites
from defined population.

The population is too large for us to considered collecting the

information from all its members.

If the whole population is taken there is no need of statically

inference.

Usually, a representative subgroup of the population(sample)

is include in the sample.

A representative sample has all the important

characteristics of the population from which it is drown.
 Advantage of sampling
♣ Reduced cost sampling reduces demands on resource
such as finance, labour and material,

♣ Greater speed data can be collected and summarized more

quickly,

♣ Quality data: due to

The use of better trained personnel and more time and
efforts can be spent on getting reliable data on each
individual sampled.

♣ It cover many important variables can not covered in DHS

or censes
 Disadvantage of sampling

There is always sampling error

Sampling may create a feeling of
discrimination within the population
It may be inadvisable where every unit in the
population is legally required to have a record
 Definition of terms used in sampling

1. Reference population (also called source population or

target population): Is the population about which an
investigator wishes to draw conclusion.

2. Study or sample population: Population from which the

sample actually was drawn and about which a conclusion
can be made. Or the subset of the target population from
which a sample will be drawn.

3. Sampling unit : The unit of selection in the sampling

process. For example sample of district, the sampling unite is
district.
4. Study unit: The unit on which information is collected.
– if the objective is to determine the availability of latrine,
then the study unit would be the household; if the
objective is to determine the prevalence of trachoma,
then the study unit would be the individual.

5.Sampling frame: The list of all the units in the

reference population, from which a sample is to be
picked.
Researchers are interested to know about factors
associated with ART use among HIV/AIDS patients
attending certain hospitals in a given Region

Target population = All ART

patients in the Region

Sampling population = All

ART patients in, e.g. 3,
hospitals in the Region

Sample
What are the error to be confronted with when
taking a random sample?

♣ When we take a sample, our results will not exactly

equal the correct results for the whole population. That
is , our result will be subject errors. Those errors has
two components.

a) Sampling errors(Random error)

b) Non sampling error(bias)
 Causes of sampling error
♣ One is chance: That is the error that occurs just
because of bad luck

♣ Design error : Un representativeness of the sample

Sampling errors (random error ) can be minimized by

increasing the size of the sample. But can not void it
completely .
 Non-sampling error/BIAS
Selection Bias Information Bias
-Dx bias - Observational error
- non response bias - Respondent error/recall bias
- Accessibility bias , - social desirability bias
-Voluntary bias

It is possible to eliminated or reduce the non sampling

Errors(bias) by careful design of the sampling procedure
and by taking care of the errors that may be a rise during
data analysis.
♠ The best source of non sampling bias is non response.
It is failure to obtain information on some of the
subjects include in the sample to be studied.

♠ Non response bias is significant bias when the

following two conditions are both fulfilled

 When non respondents constitute a significant proportion of

the sample(about 15% and more)
 When non respondents differ significantly from respondents.
Sampling Methods

Two broad divisions:

A. Non Probability sampling methods

B. Probability sampling methods

 A. Non-probability sampling
• In non-probability sampling, every item has an
unknown chance of being selected.

• In non-probability sampling, there is an assumption

that there is an even distribution of a characteristic of
interest within the population.

• For probability sampling, random is a feature of the

selection process.
• In non-probability sampling, since elements are
chosen arbitrarily, there is no way to estimate the
probability of any one element being included in the
sample.

• Also, no assurance is given that each item has a

chance of being included, making it impossible either
to estimate sampling variability.
• Despite these drawbacks, non-probability sampling
methods can be useful when descriptive comments
about the sample itself are desired.

• Secondly, they are quick, inexpensive and

convenient.
 The most common types of non-
probability sampling

1. Convenience or haphazard sampling

2. Volunteer sampling
3. Judgment sampling
4. Quota sampling
5. Snowball sampling technique
 1. Convenience or haphazard sampling
Convenience sampling is sometimes referred to as
haphazard or accidental sampling.

• It is not normally representative of the target population because

sample units are only selected if they can be accessed easily
and conveniently.

• The obvious advantage is that the method is easy to

use, but that advantage is greatly offset by the presence
of bias.
 2. Volunteer sampling
• As the term implies, this type of sampling occurs when
people volunteer to be involved in the study.

• Sampling voluntary participants as opposed to the

general population may introduce strong biases.

• In psychological experiments or pharmaceutical trials

(drug testing), for example, it would be difficult and
unethical to enlist random participants from the general
public.
 3. Judgment sampling
• This approach is used when a sample is taken based
on certain judgments about the overall population.

• The underlying assumption is that the investigator

will select units that are characteristic of the
population.

• The critical issue here is objectivity: how much can

judgment be relied upon to arrive at a typical
sample?
• Judgment sampling is subject to the researcher's
biases and is perhaps even more biased than
haphazard sampling.

• Since any preconceptions the researcher may

reflected in the sample, large biases can be introduced
if these preconceptions are inaccurate.

• One advantage of judgment sampling is the reduced

cost and time involved in acquiring the sample.
 4. Quota sampling
• This is one of the most common forms of non-
probability sampling.

• Sampling is done until a specific number of units

(quotas) for various sub-populations have been
selected.

• Since there are no rules as to how these quotas are to be

filled, quota sampling is really a means for satisfying
sample size objectives for certain sub-populations.
 5. Snowball sampling
• A technique for selecting a research sample where
existing study subjects recruit future subjects among
their friends.

• Thus the sample group appears to grow like a rolling

snowball.
• This sampling technique is often used in hidden
populations which are difficult for researchers to
access; example populations would be drug users or
commercial sex workers.

• Because sample members are not selected from a

sampling frame, snowball samples are subject to
numerous biases. For example, people who have
many friends are more likely to be recruited into the
sample.
 B. Probability sampling
♣ Involves random selection

♣ Procedures to ensure that each unit of the sample is

chosen on the basis of chance.

♣ Every sampling unit has a known and non-zero

probability of selection into the sample.

♣ Sample finding can be generalizable.

– More complex,
– More time-consuming and
– Usually more costly than non-probability sampling.
There are several different ways in which a probability
sample can be selected.

The method chosen depends on a number of factors,

such as
The available sampling frame,
How spread out the population is,
How costly it is to survey members of the
population
 Most common probability
sampling methods
1. Simple random sampling
2. Systematic random sampling
3. Stratified random sampling
4. Cluster sampling
5. Multi-stage sampling
 1. Simple random sampling
♣ The required number of individuals are selected at
random from the sampling frame, a list or a database of
all individuals in the population

♣ Each member of a population has an equal chance of

being included in the sample.

♣ Representativeness of the sample is ensured

♣ However, it is costly to conduct SRS. Moreover,

minority subgroups of interest in the population may
not present in the sample sufficient numbers for study.
To use a SRS method:
– Make a numbered list of all the units in the population
– Each unit should be numbered from 1 to N (where N
is the size of the population)
– Select the required number.

The randomness of the sample is ensured by:

Use of “lottery’ methods
Table of random numbers
Computer programs

Difficult if the reference population is dispersed or to

mach large in case of lottery method.
 2. Systematic random sampling
♣ Sometimes called interval sampling

♣ Selection of individuals from the sampling frame

systematically rather than randomly.

♣ Individuals are taken at regular intervals down the

list(For example every Kth )

♣ The first unit to be selected is taken at random from

among the first K units.
Important if the reference population is arranged in
some order:
– Order of registration of patients
– Numerical number of house numbers
– Student’s registration books
 Steps in systematic random
sampling
1. Number the units on your frame from 1 to N (where N
is the total population size).

2. Determine the sampling interval (K) by dividing the

number of units in the population by the desired
sample size.
3. Select a number between one and K at random. This
number is called the random start and would be the
first number included in your sample.

4. Select every Kth unit after that first number

 Example
To select a sample of 100 from a population of 400,
you would need a sampling interval of 400 ÷ 100 = 4.
Therefore, K = 4.
You will need to select one unit out of every four
units to end up with a total of 100 units in your
sample.
Select a number between 1 and 4 by using SRS
methods like lottery methods.
If you choose 3, the sample might consist of the
following units to make up a sample of 100: 3 (the
random start), 7, 11, 15, 19...395, 399 (up to N, which
is 400 in this case).
Merits
 Systematic sampling usually less time consuming and
easier to perform than SRS. It provide good
approximation to SRS

 Unlike SRS, systematic sampling can be conducted

with out sampling frame (usually in some situation
sampling frame not readily available )
Demerits
 If there is any sort of cyclic pattern in the ordering of
the subjects which coincids with the sampling interval ,
the sample will not be representative of the population
 3. Stratified random sampling
 It is done when the population is known to be have
heterogeneity with regard to some factors and those
factors are used for stratification

 Using stratified sampling, the population is divided

into homogeneous, mutually exclusive groups called
strata, and

 A population can be stratified by any variable that is

available for all units prior to sampling (e.g., age, sex,
province of residence, income, etc.).
 A separate sample is taken independently from each
stratum
 The population is first divide into group(strata) according
to characteristics of interest(e.g. age, geographical area,
income)

 A separate sample is taken independently from each

stratum, by using SRS or SS

 Proportional allocation: If the same sampling fraction is

used for each stratum.
 Non proportional allocation: If different sampling
fraction is used for each stratum or if the strata are
unequal in the size and a fixed number of units is selected
from each stratum
• Equal allocation:
– Allocate equal sample size to each stratum
• Proportionate allocation:
n
nj  Nj
N
– nj is sample size of the jth stratum
– Nj is population size of the jth stratum
– n = n1 + n2 + ...+ nk is the total sample size
– N = N1 + N2 + ...+ Nk is the total population
size
 Example: Proportionate Allocation

Village A B C D Total
 HHs 100 150 120 130 500
 S. size ? ? ? ? 60
Merits
 The representativeness of the sample is improved . That
is, adequate representation from each group.

 Minority subgroups of interest can be ensured by

stratification and by varying the sample fraction
between strata as required.

Demerit
 Sampling frame for the entire population has to be
prepared separately for each stratum.
 4. Cluster sampling
♠ Sometimes it is too expensive to carry out SRS
– Population may be large and scattered.
– Complete list of the study population unavailable
– Travel costs can become expensive if interviewers have
to survey people from one end of the country to the other.
♣ Cluster sampling is the most widely used to reduce
the cost
♣ The clusters should be homogeneous, unlike
stratified sampling where the strata are
heterogeneous
 Steps in cluster sampling
♠ Cluster sampling divides the population into groups or
clusters.
♠ A number of clusters are selected randomly to represent the
total population, and then all units within selected clusters
are included in the sample.

♠ No units from non-selected clusters are included in the

sample—they are represented by those from selected
clusters.

♠ This differs from stratified sampling, where some units are

selected from each group.
Merit
A list of all the individual study unites in the reference
population is not required . It is sufficient to have a list of
clusters.

Demerit
It is based on the assumption that the characteristics to be
studied is uniformly distributed through the reference
population, which may not always be the case. Hence,
sampling error is usually higher than SRS with the same
sample size.
 5. Multi-stage sampling
This method is appropriate when the reference
population is large or widely scattered

The selection done in in stage until the final sampling

unit(e.g. household, person) are arrived at.

The primary sampling unit (PSU) is the sampling unit in

the first sampling stage.

The secondary sampling unit (SSU) is the sampling unit

in the second sampling stage, etc.
Woreda PSU

Kebele SSU

Sub-Kebele TSU

HH
In the first stage, large groups or clusters are identified
and selected. These clusters contain more population
units than are needed for the final sample.

In the second stage, population units are picked from

within the selected clusters (using any of the possible
probability sampling methods) for a final sample.

If more than two stages are used, the process of

choosing population units within clusters continues
until there is a final sample.
With multi-stage sampling, you still have the benefit of
a more concentrated sample for cost reduction.

However, the sample is not as concentrated as other

clusters and the sample size is still bigger than for a
simple random sample size
Also, you do not need to have a list of all of the units in
the population. All you need is a list of clusters and list
of the units in the selected clusters.

Admittedly, more information is needed in this type of

sample than what is required in cluster sampling.

However, multi-stage sampling still saves a great

amount of time and effort by not having to create a
list of all the units in a population.

Since MSS enquire large sampling error , it need some

adjustment that is design effect(multiplying the final sample size
by some factors(1.5 or 2) to get the final sample size.
 Sample size determination
o In planning of any investigation we must decide how
many people need to be studied to answer the study
objectives .

o If the study is too small, we may fail to detect important

effects or if the study is too large we will waste
resources.

o In general, it is must better increase the accuracy of the

data collection(by improving the training of data
collection and data collection tools) than to increase
sample size after a certain point. That is called the
minimal sample size required.
In order to calculate the required sample
size , you need to know the following facts
The reasonable estimate of the key proportion to be
studied . If you can not get, guess the proportion , take it
as it has 50%.

The degree of accuracy required. This is the deviation

from the true proportion in the population as the whole.
It take with in 1% to5%

Confidence level required, usually take as 95%.

• The size of the population that the sample is to
represent . If it is more than 10, 000 the precise
magnitude is not likely to be very important.

• But if the population less than 10, 000 and

sampled population greater than study
population, we need adjustment.
 Estimation proportion
• The minimum sample size required for every
large population(N> 10,000)is

n= (za/2 2)P(1-p)/w2

Sample less than 10,000 and sample population

greater than source population the final sample
nf= n/(1+(n/N))
Example1

P= 0.28, W, 0.03, Z, 1.96(for a 95%CI)

n= 821.25 =822

Example 2
If the above sample is to be taken from s relatively
small population (N=3000), the required minimum
sample will get by adjusting the above sample
821.25/(1+(821.25/3000)) =644.7=645
Example 3
• A hospital administrator wish to know what proportion of
discharge patients are unhappy with the care received
during hospitalization. If 95% confidence interval is
desired to estimate the proportion within 5%, how large a
sample should be drawn?
n= z2p(1-p)/w2

= (1.96)2(0.5*.5)/(0.05)2

= 385

If you
do not have any information about P , take it
50% and get the maximum value pq= 0.25
58