3. Concepts to Know
Research Population The whole set of units (people, groups,
events, settings, etc.) on which the research
is focused and the findings are expected to
be generalized.
Sample Representative subset of the population the
researcher studies when the whole population
is not available
Sampling Scheme Specific strategies to select samples
Sampling Design Umbrella term that includes both selecting
strategies and sample size
4. Relationships to Remember
Research
Problem
• What needs to be studied under a specific
situation
Research
Question
• What to answer or test
Methodology • How to gather and
analyze valid data
5. Elements to Gather Valid Information
Valid Information to Answer or
Test Research Question
Data
Analysis
Methods
Sampling
Design
Researc
h Design
6. Next slides will present a basic
summary of some of the sampling
schemes that are widely used in
research.
However, be aware that the application
of probabilistic (random) schemes to
quantitative research and non-
probabilistic (non-random) ones to
qualitative studies does not imply that
these are the only approaches
available.
7. Onwuegbuzie, A. & Collins, K. (2007). A typology of mixed methods sampling designs in social science research.
The Qualitative Report, 12 (2), 281-316
Some form of non-random scheme is the most
common choice used in both quantitative and
qualitative studies and random ones for quantitative
research and non-random for qualitative studies is
the second most common combination
(Onwuegbuzie & Collins, 2007).
Such predominance of non-random samples
seems to be related to the fact that most studies in
social sciences cannot be done under experimental
conditions in which pure random selection is
expected.
8. Quantitative Research
Simple random The whole population is available and any unit has
the same chance of being chosen
Stratified random The whole population is divided into subpopulations
(strata) with respect to one or more characteristics
that interest and units are selected from each
stratum at random. Allocation of units can be done
equal or proportional to the population
Systematic Population is ordered according to a criterion and
units are chosen from the list by selecting every nth
one
Cluster Population consists of limited groups (clusters) and
sampling is focused on selecting clusters instead of
individual units. Not all clusters are included in the
sample
Multistage random Very large populations are divided into clusters and
then sub-clusters and units are selected at random
following a general to specific direction
9. Qualitative Research
Convenience Sample consists of units that are available and/ or willing to
participate
Purposeful Researcher is interested in studying specific groups. Selection
of units can be done at random, stratified, or using more than
one scheme (mixed)
Quota Units are selected in relation to pre-defined characteristics
either in proportion to population sub-groups or minimum
number from each sub-group
Snowball Similar units are required and access to them is done by
asking participants to recommend peers
Multistage purposeful Units are selected in more than one stage and always
applying a purposive scheme
10. Qualitative Research
Typical case Units are chosen because they represent the average element
of what is studied
Maximum variation When differences are the research target, dissimilar units are
chosen to form sample
Criterion Units are selected because each one of them represents one
or more desired criteria
Theoretical Units are chosen because they can provide input information
to build or test a theory
11. Mixed Methods Research
All previous schemes When choosing samples for mixed methods studies,
researchers need to take into account the purpose of the
study, the research questions, and the stages in which the
study will be developed.
All those elements help them select the most appropriate
sampling scheme for each stage. When generalization is the
main concern of a stage, the first five schemes presented are
the best options. When understanding is the focus of a stage,
then the rest of schemes should be considered.
13. Quantitative Research
The research director of an educational
city system, that includes forty-five
elementary schools, wants to answer the
following research question:
What’s the achievement in mathematics of
fifth-grade students of the school system,
measured by a standardized achievement
test?
15. Cluster Sampling
It is too expensive to administer the test to the whole population
and the same happens if the simple random sampling is chosen.
Stratified random sampling implies administering the test to some
of the students of a class and not to others and that may be
inconvenient from the students’ point of view.
Since students are organized in classes –which are in fact clusters,
the most appropriate option is cluster sampling and that will allow
to test all students in the chosen classes.
16. Qualitative Research
A researcher wants to know the reasons
why some employees of large companies
have strong pro-mentoring duties views.
18. Snowball
Since participants with similar views within organizations are
required and they could be difficult to locate, it is appropriate
to ask them to recommend potential subjects.
20. Criteria to Select Sample Size
• Costs (money, time, and effort) to get sample data.
• For quantitative studies:
• Population size (the larger the population, the larger the
sample)
• Confidence interval (how much error will be allowed)
• Confidence level (how much confident you want to be that
your results are within the selected confidence interval)
• Standard deviation (how much dispersion from the mean
you expect)
These criteria are used by calculators to determine the most
appropriate sample size for your study.
21. Criteria to Select Sample Size
For qualitative studies:
Samples are usually small, but the precise number is very
unlikely to be determined at the beginning of the study .
The main criterion is to have a sample as big as needed to
have all the information that might be important. Therefore,
when information becomes redundant, the sample size has
been reached.
22. Criteria to Select Sample Size
For mixed studies:
• Usually the sizes of the sample for each stage are
different
• The mixed methods variant used influences the
sample sizes. For example, in an explanatory design
the same participants must be included in all the
stages while in an exploratory design more
participants are needed in the quantitative stage.
23. Advanced Idea for Today:
More than one criterion is needed to select the
best sampling design to gather valid data.
24. Now You Know
1. The relationship between research
problem, research question, and
sampling.
2. Some of the basic sampling
schemes you may use to select
participants for your study