Scribd
Upload
7 views27 pages

CLS 351 Lecture Note - Sampling

Epidemiology

Uploaded by

Abdelwahab Ali
Copyright
© © All Rights Reserved
We take content rights seriously. If you suspect this is your content, claim it here.
Available Formats
Download as PPTX, PDF, TXT or read online on Scribd
7 views27 pages

CLS 351 Lecture Note - Sampling

Epidemiology

Uploaded by

Abdelwahab Ali
Copyright
© © All Rights Reserved
We take content rights seriously. If you suspect this is your content, claim it here.
Available Formats
Download as PPTX, PDF, TXT or read online on Scribd
You are on page 1/ 27

Sampling in research

CLS 351
Outline:
• Recognizing different definitions concerning
sampling.
• Listing two basic approaches to sampling.
• Identifying various types of probability
sampling and their uses.
• Identifying various types of non probability
sampling and their uses.
Definitions:
A Sample
A subset or small part of the total number (population)
that could be studied.
Population
All people in a defined setting.
*The greater the diversity in the population, the larger the
sample size should be.
Sampling
Process of selecting the sample.
Generalizability:
Extending the findings from a sample to the larger
population.

Representativeness:
Extent to which the sample and population are alike.
• When the sample is representative, sample
results can be generalized.

Randomization:
Each unit within a sampling frame has an equal chance
of being selected.

Sample Size:
Number of elements in the obtained sample.
Sampling frame
• Group of individuals having a real chance of being
selected for the sample.
 For example, if a study has to be carried on
adolescent females, the researcher selected female
students from preparatory and secondary schools
as his sampling frame.
Can these results be generalized on the population
(all adolescent females in the country)?
NO; because the sample is biased, because not all
adolescent females are in schools
i.e. adolescent females with poor mental or
lower socio-economic status.
Purposes of sampling (why to sample)?
1. Economy: cheaper to observe a part rather
than whole population.
2. Timeline,
3. The large size of many populations,
4. Inaccessibility of some of the population (e.g.
people in prison),
5. Accuracy.
Two basic approaches to sampling:
I. Probability sampling
II. Nonprobability sampling.
I. PROBABILITY SAMPLING
Each person in the population has the same
probability/chance of being selected.

Used when the purpose of a research is to make


conclusions affecting the population as a whole.

Types of probability sampling:


1. Simple Random Sampling (SRS).
2. Systematic Sampling/Interval Random Sampling.
3. Stratified Sampling.
4. Cluster (area) Sampling.
I. Probability sampling

1. Simple Random Sampling


(SRS)
A. Selecting a sample using a table of
random numbers.
(Obtain a list of all residents,
Use a sequence of numbers from
this random numbers table.)
B. Selecting a sample by drawing
names or numbers out of a hat.
C. Using Random Number
Generator
(available at http://stattrek.com/Tables/Random.aspx)
I. Probability sampling
2. Systematic Sampling/Interval Random Sampling:
Selecting every 10th, 20th, or 30th person from a list, starting
with a person randomly selected from among the first 10
persons.
• referred to as a systematic sample with a random
start.
I. Probability sampling
2. Systematic Sampling/Interval Random Sampling:
Sampling Interval
Standard distance between elements in the sample frame (=
total population/sample size).
Example, if you have a population of 1000 persons and the
sample size is 100
Sampling interval:
= total population/ sample size
1000/100=10 persons, Selecting a person every 10
persons.
In case of systematic sample with a random start, a
random number will be selected from the first 10
persons (say the 3rd person) then 13th, 23rd, 33rd..etc
persons will be selected.
I. Probability sampling
3. Stratified Sampling:
Populations are classified into strata, Then separate
samples selected from each strata. To assure that certain
variable (gender, age…etc) will have sufficient sample (not
over-sampled or under-sampled).
I. Probability sampling
For example,
1000 employees in a hospital; 900 males & 100
females
If you randomly select 100 persons, Males can
oversample the females. So, can classify them into
two strata (males and females)
• Select 100 employees (10% of each
stratum; 90 males and 10 females)
• Using systematic random sampling with
a random start.
I. Probability sampling
4. Cluster (area) Sampling
If the target population is spread across a wide
geographic region, so divide this population into clusters
(states, villages, schools, etc), then sample everyone in
the selected clusters.
I. PROBABILITY SAMPLING
4. Cluster (area) Sampling
Example:
10 schools must be selected from all elementary schools
in a city (500 schools), To study anemia among their
students
Each elementary school will be a cluster
• Have a list with the 500 elementary schools
• Either simple or systematic random sampling to select
the schools,
Sampling interval= 500/10= 50 schools
So, every 50 schools you can select a school
Select the first school randomly from the first 50
schools (say 14th then select the 64th, 114th,164th ..etc)
All students in a cluster (school) must be interviewed.
II. Non-probability sampling
The probability of inclusion in the sample is unknown.

Types of Non-probability Sampling:


1. Convenience Sampling.
2. Proportional quota sampling.
3. Criterion sampling.
4. Snowball/chain/network Sampling.
5. Extreme and deviant case sampling.
6. Maximum variation/heterogeneity/diversity sampling.
7. Homogeneous sampling.
II. Non-probability sampling
1. Convenience Sampling
Choosing Persons because they are:
• “easy to get to”
• interested in your research.
Example:
-In mall samples, interviewing the first 10 people met in a
mall.
-Use of volunteers is a convenience sampling
II. Non-probability sampling
2.Proportional quota sampling
Sampling until sufficient units are obtained.
Example:
Selection of 100 students from college of Nursing,
Make two groups (male and female)
Begin to select the required sample size
until getting 50 males and 50 females
II. Non-probability sampling
3. Criterion sampling
Used in quality assurance
Set criteria and pick cases that meet these criteria,
Example,
• Age requirements,
• Service users,
• Twitter users.
II. Non-probability sampling
4. Snowball/ Chain /network Sampling
Initial individuals are sampled, these individuals
identify more individuals to be sampled.
II. Non-probability sampling
5. Extreme and deviant case sampling
Selecting samples from the extremities:
• top of the class or drop-out,

Used to learn from highly unusual manifestations


of the phenomenon of interest.
II. NON-PROBABILITY SAMPLING
6. Maximum variation/heterogeneity/diversity
sampling
Selecting a sample with a wide range of variations,
For example, interviewing University students, to
get students of different nationalities, cultures,...etc

7. Homogeneous sampling
In the previous example, instead of having
maximum number of nationalities, focusing on one
nationality (say Saudis only).
Conclusion
Sampling yields significant research result. However,
with the differences that can be present between a
population and a sample, sample errors can occur.
Therefore, it is essential to use the most relevant and
useful sampling method.

You might also like