SS3290
Social Work Research
Topic 4:
Sampling
❑ Introduction
❖ Sampling is a very important and pivotal
research procedure to influence the
trustworthiness and quality of a research.
❖ It may lead to reliability, validity, accuracy and
generalizability of the results of a study.
❖ When sampling people to take in your study,
rationality, reasons, approaches and steps must
be carefully considered.
❑ Introduction (cont’d)
❖ In sampling, the first thing you need to consider
is who are your study population.
❖ For example, if you want to study whether the
Hong Kong citizens would support for a whole-
citizen retirement scheme policy, thereby all
Hong Kong people who have a right of abode
and are aged 18 or above should vote for the
policy.
❖ Hence all Hong Kong citizens are your study
population.
❑ Introduction (cont’d)
❖ In a practical term, you cannot survey all the
Hong Kong citizens for your study due to time
and resource reasons; therefore you must select
a subgroup of the Hong Kong citizens to take
part in your study to represent your study
population.
❖ In this sense, the selected participants are the
sample of your study; and the method and
procedure that you employed to select your
sample is called sampling.
❖ Before moving on, have you ever heard the story
of President Alf Landon?
❑ Definitions of sample components
and the population
❖ Population- the entire set of individuals or other
entities to which study findings are to be
generalized, e.g. secondary school students in
Hong Kong; local elderly people; RSWs in HK,
China, Taiwan.
❖ Elements- the individual members of the
population whose characteristics are to be
measured, e.g. the secondary school students
who are selected and interviewed are elements.
❑ Definitions of sample components
and the population (cont’d)
❖ Sample- a subset of a population that is used to
study the population as a whole, e.g. a subset of
secondary school students that would take part
in your study.
❖ Sampling frame- a list of all elements or other
units containing the elements in a population,
e.g. how many schools, which classes and
students are to be selected in the study are the
sampling frame.
❑ Definitions of sample components
and the population (cont’d)
❖ Enumeration units- units that contain one or
more elements and are to be listed in a sampling
frame, e.g. classes selected are enumeration
units and students who are invited are elements.
❖ Sampling units- refer to units listed at each
stage of a multistage sampling design, e.g.
districts, schools and classes being chosen in
the process of randomly selecting students as
study sample are sampling units.
❑ Sampling Methods
❖ One main concern of sampling is to locate
participants in the study that can represent the
characteristics of the population under
investigation.
❖ There are two types of sampling methods in
social science/ social work research: probability
sampling method and nonprobability
sampling method.
❖ However, any of these two types of sampling
methods need to strive for maximizing diversity
of the study population.
❑ Sampling Methods (cont’d)
❖ Probability sampling refers to the general term
for study samples that are selected in accord
with probability theory, typically involving some
random-selection mechanism, e.g. simple
random sampling and systematic sampling.
❖ Nonprobability sampling refers to any
technique in which samples are selected in
some way not suggested by probability theory,
i.e. reliance on available subjects (participants)
and snowball sampling.
❑ Sampling Methods (cont’d)
❖ Probability sampling method refers to the
randomly sampling procedures in which the
probability of selection is known and contain no
zero of some chance of being selected as an
element.
❖ This sampling method can randomly select
elements from the population and therefore has
no systematic bias.
❑ Sampling Methods (cont’d)
❖ Several different probability sampling method
approaches include:
➢ Simple random sampling- it involves assigning
a single number to each element (i.e. each
participant) in a sampling frame and goes to
randomly select the elements according to
certain assigned numbers by probability. In
practice, a table of random numbers is used to
help select elements for the sample.
❑ Sampling Methods (cont’d)
➢ Systematic random sampling- it is a variant of
simple random sampling, in which the first
element is selected randomly from a list and
then every nth elements is selected. This is a
systematic way for drawing a random sample
when the population elements are arranged
sequentially. Three steps should be noted when
using this method:
❑ Sampling Methods (cont’d)
1) The cases in the population are divided by the
number of cases required for sampling, which
yields the sampling interval, e.g. to select 100
cases from 1000 people the sampling interval is
10, meaning every 10th case is selected.
2) A number from 1 to 10 is randomly drawn, e.g.
7; and the selected number 7 will be the
identifier to select cases, e.g. 7, 17, 27 and so
on.
3) The final sample is selected based on this
sequence to complete the sampling procedure,
e.g. N= 100.
❑ Sampling Methods (cont’d)
➢ Stratified random sampling- this method is to
grouping the units composing a population into
homogenous subsets (groups) before sampling;
and then it may be used in conjunction with
simple random, systematic random sampling or
cluster sampling to select sample for enhancing
its representativeness
➢ The main function of this sampling method is to
organize the population into homogeneous
subgroups/ subsets (with heterogeneity between
subsets)
❑ Sampling Methods (cont’d)
➢ After designing the homogeneous subgroups/
subsets, you need to select the appropriate
number of elements from each subgroup to form
your a final sample who can represent the
population.
➢ Samples are drawn from this method can be
more heterogeneous and diverse for reducing
sampling error, but more costly and time
consuming.
❑ Sampling Methods (cont’d)
➢ For example, by using stratified random
sampling method, a researcher should ensure
the sample that contains the proper numbers of
females and males (gender-stratifying) and
minority groups (ethnicity-stratifying). Three
steps should be noted when using this method:
1) The available variables that can most represent
the population should be selected as
stratification variables, e.g. gender, age, race/
ethnicity, localities, which are sued to form
strata.
❑ Sampling Methods (cont’d)
2) Random selection of a subsample from each
stratum is needed.
3) Pooling all the selected subsamples together to
form a final stratified representative sample that
can be well representative of the population.
❑ Sampling Methods (cont’d)
➢ Multistage cluster sampling- In this method,
natural groups (clusters) are sampled initially,
with the members of each selected natural
group being subsamples afterward. It involves
the repetition of two basic steps: Listing and
Sampling.
➢ The list of primary sampling units (e.g. schools
and public housing estates) is compiled and,
perhaps stratified for sampling.
➢ A sample of these compiled units is then
randomly selected.
❑ Sampling Methods (cont’d)
➢ The randomly selected primary sampling units
are again listed and perhaps stratified; and this
list of secondary sampling units is then again
sampled and so forth.
➢ The final representative sample can be achieved
after pooling individual elements in each
subsamples embedded in the finest sampling
units.
❑ Sampling Methods (cont’d)
➢ For example, a representative secondary
student sample is randomly selected by first
listing and randomly drawing districts/
neighborhoods in Hong Kong, and then random
selection of schools in each sampled districts/
neighborhoods is applied, and again random
selection of classes in each sampled schools is
applied, and finally random selection of students
in each sampled classes is applied to achieve a
final representative sample.
❑ Sampling Methods (cont’d)
❖ Non-probability sampling methods refer to
availability/ convenience sampling procedures,
in which researchers go to select elements from
an available and convenient “population” that is
easily approachable, e.g. elderly members in
DECCs or students in three available schools.
❖ These sampling methods do not randomly select
elements from the population and thereby may
contain systematic bias.
❑ Sampling Methods (cont’d)
❖ Several different non-probability sampling
methods include
➢ Quota sampling- it involves the quotas that are
set to ensure the sample that represents certain
characteristics in proportion to their prevalence in
the “population”, e.g. gender.
➢ For example, if there are 10,000 elderly people
with a gender combination of 60% females and
40% males known in a community and you decide
a sample of 1,000, you have to sample 600
females and 400 males by quotas to represent the
characteristic.
❑ Sampling Methods (cont’d)
➢ The problems of this method are that the
accurate characteristics of quota frame are hard
to be guaranteed, the sample is not selected
randomly by chance; and many human biases
may involve.
❑ Sampling Methods (cont’d)
➢ Purposive sampling- this method is sometimes
called judgmental sampling, in which a
researcher goes to select her sample on the
basis of her own knowledge of the population, its
elements, and the nature of her research based
on the study purpose and needs.
➢ The use of this method is because you are hard
to have a complete list of the target population,
e.g. homeless people, but a subset of them is
easily identified.
❑ Sampling Methods (cont’d)
➢ Snowball sampling- this method is useful when
the members if a specific population are difficult
to locate, e.g. compensated dating girls, male
sex workers, and adolescent gangsters.
➢ The sampling procedure is implemented by
collecting data on the few members of the target
population for whom you are able to approach,
and then asking those approached participants
to provide needed information to locate other
members of that population they would be more
to be reached.
Key Terms
Sampling and sample
Population
Elements
Sampling frame
Enumeration units
Sampling units
Probability sampling
Simple random sampling
Systematic random sampling
Stratified random sampling
Multistage cluster sampling
Key Terms
✓ Non-probability sampling
✓ Quota sampling
✓ Purposive sampling
✓ Snowball sampling