Introduction to Sample Size

Calculation and Sampling Techniques

Dr. Muna Hassan Mustafa, MBBS,MD

 Contents
• Requirements for sample size calculation
• Sample size calculation formulae for different types of
studies
• Online- electronic sample size calculators
• Sample size for qualitative studies Vs quantitative studies
• When to use sampling instead of a census.
• Difference between probability and non probability
sampling.
• When and how to use various sampling techniques.
• Different types of error that can occur in a study.
Purpose of Sample size calculation

Sample size

How closely the sample values

approximate the population values
 Quantitative Studies
General rules:
• The expected variation in data
• Desired margin of error
• The level of confidence
• Sampling technique
• Non-Response rate
 Quantitative Studies
Estimating mean
n = sample size
z = confidence level (1.96 for 95%) n = z2 x s2 x deff
e = desired margin of error e2
p = percentage
s = standard deviation Estimating proportion
deff = design effect
n = z2 x p(1-p) x deff
e2
 Sample Size Calculators

• Open Epi

• https://www.openepi.com/Menu/
OE_Menu.htm
 Quantitative Studies
• Desirable sample size
• The eventual sample size is a compromise
between what is desirable and what is feasible
• It depends on:

Time
Money
Manpower
 Qualitative Studies
• No fixed rules
• Richness of data and analysis capabilities determine
validity and meaningfulness
• Depends on :
– The objectives of the study
– The study population

• Sample size is estimated before hand

Continue to collect
• Point of Saturation redundancy data until no new
information emerge
 Sampling Techniques

Dr. Muna Hassan Mustafa, MBBS, MD

 Technical Terminology

• Population: Collection of the elements which has characteristic

in common.
• An element : Object on which a measurement is taken.
• Population size: Number of elements in the population
• Sample : Collection of sampling units drawn from a sampling
frame
• Sampling: Selecting a number of study units from a defined
study population
• Parameter: Numerical characteristic of a population
• Statistic: Numerical characteristic of a sample
 Purpose of Research
Precision Generalizability
statistics are •
similar to Results Applicable •
population to the population
parameter
 Purpose of Research
Precision Generalizability
Results •
Results Applicable to •
approximate
population the population
Sampling •
parameter
Sample Size • Technique
Sample Versus census

Sample Census
Save money, • Population is very •
time effort small
Extensive •
Provide valid • resources
reliable, useful Expected low •
results response
 Purpose of Sampling

:Representative sample

Contains all the important characteristics of

population from which it is drawn
 Preparing for sampling
• Objectives of the study
• Definition of the population
• Nature of the population
• Availability of resources
• Research design
• Ethical and legal issues
 Sampling Techniques

Probability Non-Probability
Samples Samples

Simple Cluste Multist Convenien Purposiv Snowb

random
Systemati
c random
Stratified
random r age ce sample e sample Quota all
 PROBABILITY SAMPLING

A probability sampling:
• Every unit in the population has a chance (greater than zero) of
being selected in the sample

• Probability of selection can be accurately determined.

• Allows application of statistical sampling theory to results to:

– Generalise
– Test hypotheses

23
 Sampling Techniques

Probability Samples

Simple Systemati Stratified Multist

c random Cluster
random random age
 Simple Random
Sampling
Applicable when population is:
•Small
•Homogeneous
•Readily available

• Each element of the population has an equal chance

of being selected

25
 cl
hi
s
et
Simple Random Sampling
l
o
e
mf
e
a
n
lt
l
i
e
n
l
et
mh
e
e
n
Simple Random Sampling
Table of random numbers
684257954125632140
582032154785962024
362333254789120325
985263017424503686
 Simple Random Sampling

Advantages Disadvantages
Impracticable •
If sampling
frame large
Easy •
Rapid • Minority •
subgroups
unrepresente
d
29
 Systematic Sampling
• Applicable when population is:
•Small
•Readily available

• Selecting elements at regular intervals

30
 et
kh
=es
Systematic Random Sampling
a(
m f
pi
p
or P

psl
r
o
c
e
e
d
s

uti
w
i
t
h

t
h
e

nl
s
e
l
e
c
t
i

gat
o
n
o
f
e

o
v
e
r
y

t
k
t
h
e

fti
l
e
m
e
n
t
f

oher
r
o
m
t
h
e
n

na
o
n
w
a
r
d
s

m
.

eskt
hi
Systematic sampling
 Systematic Sampling

33
 Systematic Random
Sampling

Advantages Disadvantages
Sample may be •
biased if
Sample easy to • hidden
select periodicity in
Sample evenly •
spread over entire population
reference population coincides with
that of
34 .selection
 Stratified Sampling
• When we want to represent all subgroups of the
population
• Frame is divided into separate subgroups "strata."
• Individual elements are randomly selected from all
strata
• Every unit in a stratum has same chance of being
selected.
• Different sampling approaches can be applied to
different strata.

35
 ns
gt
r Stratified Random Sampling
af
tr
a
a
ma
ec
c
oi
ndr
ti
on
g
s
t
 Stratified Sampling

Draw a sample from each

stratum

37
 Stratified Random Sampling

Sampling frame
Disadvantages
•
Advantages of entire
More accurate •
of population population has to
representation be prepared
separately for
Potential for • each stratum
reducing Can potentially •
sampling error require a larger
38
sample
 Cluster Sampling
• When the population is widely dispersed
• Population divided into clusters of homogeneous
units, usually based on geographical contiguity.
• Sampling units are groups rather than
individuals.
• A sample of clusters is selected.
• All units from the selected clusters are studied.

39
 Cluster Sampling
Section 1 Section 2

Section 3

Section 5

Section 4
 nai
n
dr

m
oeg Cluster Sampling
fli
nyr
ac
m(l
p
ue
dr
oei
mdf
a
cr(
yl
ue
qs
uat
 Cluster Sampling
Two types of cluster sampling methods.
One-stage sampling. All of the elements within
selected clusters are included in the sample.
Two-stage sampling. A subset of elements within
selected clusters are randomly selected for
inclusion in the sample.

42
 Cluster Sampling

Advantages Disadvantages
Cuts down on the •
cost of preparing a Sampling error is •
.sampling frame higher for a simple
Reduces travel and • random sample of
other administrative
costs .same size
More convenient• Costs and problems •
for geographically of statistical analysis
dispersed are great
populations
43
 Multistage Sampling
• Complex form of cluster sampling in which two or
more levels of units are embedded one in the
other.
• Essentially the process of taking random samples of
preceding random samples.
• Used when a complete list of all members of the
population not exists and is inappropriate

44
 n eti
nhdess
goee Multistage Sampling
msll
el
ecf
ycl
rt
eet
aec(
mdpdt
eerr
deti
m
ooag
w hi
o
n fr
oy
u
nrs
ss
eea
m gh((
ost
peil
oderl
 What is the appropriate sampling
?technique
• Study the prevalence of diabetes in Riyadh
region
• Study the satisfaction of FMRP residents
regarding course organization
• Study association between hypertension and
CHD
• Study the average weight of FMRP residents
• Study behaviours of drug addicts