Sampling
Non-probability Sampling
Probability Sampling
Sample
A smaller set of cases a researcher selects
from a larger pool and generalizes to the
population.
Quantitative vs. Qualitative Approach to
Sampling
1. Representative sample from a population
2. Generalizations
3. Sampling based on mathematical theories of probability
4. Random sampling (reduces time and cost and increases
precision or accuracy)
5. Collect specific cases, events or actions that can clarify
and deepen understanding.
6. Learn about the processes of social life in a specific
context.
Non-Probability Sampling
Non-Probability Sampling
A sample in which the sampling elements
are selected using something other than a
mathematically random process.
Used by qualitative researchers
Sample size is not the question
Selects cases gradually with focus on
specific content of the case
Relevance of research topic is more
important than representativeness.
Types of Non-Probability
Samples
Deviant
Haphazard Quota Purposive Snowball Sequential Theoretical
Case
1. Haphazard, Accidental Or Convenient
• Ineffective, unrepresentative and not
recommended
• Get any cases in any manner that is
convenient
• High systematic errors
• E.g., Television interviews, newspaper
surveys
2. Quota Sampling
• A researcher identifies relevant categories
of people and then decides how many to
get in each category.
• Number of people in various categories of
the sample is fixed.
• Once quotas or categories are fixed,
haphazard sampling is used to select
people from the categories.
3. Purposive or Judgmental Sampling
• Used in exploratory research or field research.
• Uses the judgement of an expert in selecting
cases.
• Selection of cases is done with a specific
purpose in the mind.
• Important for selecting unique cases that are
informative.
• Also good for selecting people of a difficult-to-
reach, specialized population.
• An attempt is made to locate all possible cases
of a highly specific population.
4. Snowball Sampling (Network, Chain referral or
reputational sampling)
• Researcher begins with one case, and then
based on information about
interrelationships from that case, identifies
other cases, and repeats the process again
and again.
• Interconnected web of linkages
5. Deviant Case Sampling (Extreme case
sampling)
• A researcher seeks cases that differ from
dominant pattern or that differ from the
predominant characteristics of other cases.
• The goal is to locate a collection of
unusual, different or peculiar cases that are
not representative of the whole.
• A greater insight is obtained into social
processes or setting.
6. Sequential Sampling
• A researcher wants to find as many relevant
cases as possible, until time, financial
resources or his or her energy is exhausted,
or until there is no new information or
diversity from the cases is generated.
• Very similar to purposive sampling, but,
with a difference
• In purposive, the principle is to get every
possible case
• In sequential, the principle is to gather cases
until a saturation point is reached.
7. Theoretical Sampling
• Researcher selects specific times, locations
or events to observe in order to develop a
social theory or evaluate theoretical ideas.
• Basis of Grounded theory
Probability Sampling
Probability Sampling
Population: The abstract idea of a large group of many cases from which a
researcher draws a sample and to which results from a sample are generalized.
(e.g., all people in a city)
Target Population: The concretely specified large group of many cases from which
a researcher draws a sample and to which results from a sample are generalized.
Sampling Ratio: The ratio of the size of the sample to the size of the target
population
Sampling Frame: Operationalization of population; Developing a specific list that
closely approximates all the elements in the population.
Conceptually defined population and sampling frame should match to reduce
errors
Sampling element: Unit of analysis or case in a population. (E.g., a person, a
group, an organization, a message, an action etc.)
Parameter and Statistics
Any characteristic of a population (e.g., city
dwellers who smoke cigarettes) is a
population parameter. For a large population,
parameter can never be accurately known. So
researchers estimate it on the bases of
samples. They use information from the
sample , called statistic, to estimate
population parameters.
Random Samples and Sampling
errors
Random sample is a sample in which the
researcher uses the random number table or
similar mathematical random process so that each
sampling element in the population will have an
equal probability of being selected.
Sampling Error: The extent to which a sample
deviates from its representative population. It is
the difference between sample results and a
population parameter due to random process.
Types of Probability Sampling
Simple
Systematic Stratified Cluster
random
Sampling Sampling Sampling
sampling
1. Simple Random Sampling
• Basis of all other types of random sampling
• A researcher creates a sampling frame and
uses a pure random process to select cases
that each sampling element in the
population will have an equal probability
of getting selected.
• Use of Random number table
• Simple random sampling with
replacement (Unrestricted) and without
replacement (restricted).
Sampling distribution and Central Limit
Theorem
Sampling distribution is a frequency
distribution created by drawing many random
samples from the same population.
When many different random samples are
plotted in the graph, the sampling distribution
looks like a normal or bell-shaped curve and
the mid-point of the curve approaches the
population parameter :Central Limit theorem.
2. Systematic Sampling
• A random sample in which a researcher selects
every Kth (4th, 12th etc.) case in the sampling frame
using a sampling interval.
• Sampling interval is the inverse of the sampling
ratio calculated by dividing population size with
sample size.
• Systematic sampling cannot be substituted for
simple random sampling when the elements in a
sample are organized in some kind of cycle or
pattern.
3. Stratified Sampling
• Sampling in which the researcher first
identifies a set of mutually exclusive and
exhaustive categories, dividing the
sampling frame by the categories (strata),
and then uses random selection to select
cases from each category.
• Samples generated are more representative
of population than simple random
sampling.
Stratified sampling
Proportionate: proportion of a stratum
in a sample is true proportion in the
population.
Disproportionate: proportion of a
stratum in a sample to differ from its true
proportion in the population.
4. Cluster Sampling
• A type of random sampling
• Uses multiple stages
• Dispersed sampling frame spread acrooss
wide geographical areas
• Aggregated units are randomly selected
• Samples are drawn from the sampled
aggregated units, also called clusters.
• A cluster is a unit that contains final
sampling elements but can be treated
temporarily as a sampling element itself.
Example: Three-stage cluster sampling
• Stage 1: Random selection of big clusters
• Stage 2: Random sampling of small clusters
within each selected big cluster
• Final Stage: Sampling of elements from within the
sampled clusters.
• Less expensive but less accurate too.
• Each stage added increases sampling errors.
• A design with more clusters is better because
elements within clusters tend to be similar to each
other, thus, raising the issue of representativeness
of sample to the population.
Cluster Sampling
Within Household Sampling: randomly selecting a person
within a household in several ways. This is to avoid
selecting the first person to answer the call or open the
door or selecting the person who appears to be friendliest.
Proportionate or unweighted cluster sampling: Size of
each cluster (or number of elements at each stage) is the
same
Probability Proportionate to size (PPS): An adjustment
made in cluster sampling when the each cluster does not
have the same number of sampling elements.
Supplementary Techniques of Sampling
Within Household Sampling: randomly selecting a person
within a household in several ways. This is to avoid selecting
the first person to answer the call or open the door or selecting
the person who appears to be friendliest.
Random-digit dialing: A method of randomly selecting cases
for telephone interviews that uses all possible telephone
numbers as a sampling frame.
Combination of non-probability and probability sampling for
special situations like hidden population who engage in
clandestine, socially disapproved or concealed activities and
who are difficult to locate and study.
Issue of Sample Size
Sample size depends on…
• 1. The degree of accuracy required: large
sample sizes, more accuracy.
• 2. The degree of variability / diversity in
population: again large sample size is
required.
• 3. The number of different variables
examined simultaneously: Large sample
size is necessary.
A rule of thumb
The smaller the population, the bigger the
sampling ratio has to be for an accurate
sample
Larger the population, smaller sampling ratio
yields good results. This is because as the
population size grows, the returns in
accuracy for sample size shrink.
A rule of thumb
Under 1000 population size, approximately 30%
sampling ratio is required. {1000 / 300; 30 %}.
For moderately large population, a small sampling
ratio of about 10 % is required {10,000 / 1,000 = 10 %}
For large populations (over 150,000), smaller
sampling ratio of 1 % is required .{150,000 / 1500 = 1
%}
Over 10 million, 0.025 %; sample size of 2500
Reducing Sampling Error
LARGER THE SAMPLE SIZE, SMALLER IS THE
SAMPLING ERROR.
GREATER THE HOMOGENEITY IN A SAMPLE.
THE SMALLER ITS SAMPLING ERROR.
A NARROWER CONFIDENCE INTERVAL
MEANS MORE PRECISE ESTIMATES OF THE
POPULATION PARAMETER FOR A GIVEN
LEVEL OF CONFIDENCE>