0% found this document useful (0 votes)
81 views36 pages

Sampling Method - CGS

The document discusses selecting study participants and describes inclusion and exclusion criteria to determine who will be included in a study example. It also defines terminology used in sampling such as sampling, population, sample, and sampling error. Different types of probability and non-probability sampling methods are described including simple random sampling, stratified sampling, cluster sampling, and convenience sampling.
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
0% found this document useful (0 votes)
81 views36 pages

Sampling Method - CGS

The document discusses selecting study participants and describes inclusion and exclusion criteria to determine who will be included in a study example. It also defines terminology used in sampling such as sampling, population, sample, and sampling error. Different types of probability and non-probability sampling methods are described including simple random sampling, stratified sampling, cluster sampling, and convenience sampling.
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/ 36

Selecting Study Participants

Identify those individuals to be included in your study. It can


be determined by having inclusion and exclusion criteria
Example: The participants of the study will include married
females of childbearing age who have never been pregnant,
have a household income of less than P50,000 per year, and
receive healthcare benefits from either their employer or the
employer of their partner.

Remember: the greater the number of inclusion criteria, the


more restrictive is the selection process, and the more
difficult it may become to identify the study participants
Terminology used in sampling
 Sampling – process of drawing a sample from a population
 Generalize– to make statements about a population based on
sample data
 Census – a study based on data from the whole population rather
than a sample
 Sample – a set of elements taken from a larger population
N – population size
n – sample size
 Statistic – a numerical characteristic of a sample
 Parameter – a numerical characteristic of a population
 Sampling error – the difference between the value of a sample
statistic and the population parameter
Sampling
Probability sampling – uses random selection of
participants and assures that each person in the
study has an equal chance and known probability
of being selected
 Random selection (randomization) of
participants is the common element of
probability sampling.
Sampling
Random assignment – another sampling procedure used in
research studies that require more than one group of
participants
 It is the assignment of participants to groups such that
each participant has an equal chance of being assigned
to any of the groups; the purpose of which to equalize
characteristics among the groups
 Example: 100 participants (60 males and 40 females –
each group will have 30 males and 20 females)
Advantages of random selection of participants:
1. It helps ensure the representativeness of the sample to
the study population about which generalizations will be
made
2. It shows that the researcher was unbiased as to which
members of the study population were selected for the study
Probability Sampling Procedures
Sample Primary Descriptive Elements
Simple random Each subject has an equal chance of being
selected if table of random numbers is
used
“Fishbowl” (or Approximates random sampling but not as
“out of a hat” precise. Can be done with or without
replacement
Systematic Using a list, subjects are selected at a
constant interval (N/n)
Nonproportional The population is divided into subgroups
stratified based on key characteristics (strata) and
subjects are selected from the subgroups
at random to ensure representation of the
characteristic
Probability Sampling Procedures
Sample Primary Descriptive Elements
Proportional Like the nonproportional stratified random
stratified sample, but subjects are selected in
proportion to the numerical strength of
strata in the study population
Cluster or area Random sampling of groups (e.g. teachers’
classes) or areas (e.g. city blocks) instead
of individuals
Matrix The responses of several randomly
selected subjects to different items are
combined to form the response of one
Simple random sampling
Researcher must have a list of all sampling units
in the study population known as sampling frame
Assigning a number with an equal number of
digits to each sampling unit in the frame (e.g 000
– 099)
Decide how many people will be selected in the
sample (e.g. 10 sample size)
Researcher closes his eyes, points to a number as
the starting point then proceeds to way he would
want to
Fishbowl or “Out of a Hat” Sampling
Everyone in the sampling frame is represented on a
piece of paper by a number, name or other identifying
characteristic.
Pieces of paper are placed in a container (e.g fishbowl,
hat) and someone draws them out one at a time until
the desired sample size is obtained
To ensure that all have an equal chance of being
selected, the pieces of paper should be placed back
into the container after being drawn so there is always
the same sample number from which to draw
Systematic sampling
The participants are selected from the sampling frame at a
constant interval determined by dividing the study
population size (N) divided by the study population size (n)
beginning with a randomly selected individual from the
sampling frame
Example:
Choose a sample of 10 health educators from the
professional association membership list of 50
 Randomly choose a number between 00 and 49, such as
24, using a table of random numbers.
 Choose every fifth (N/n=50/10=5); start with person 24,
then 29, 34, 39, 44, 04, 09,and 14) until there are ten
members for the sample
Stratified sampling
Researcher must divide the study population into groups (or
strata) then select a simple random sample from each group
Example: Collect data from school districts in a particular
state concerning the number of heath elective classes
offered in high schools
 Divide the school districts into groups by size such as
small (1-1000 students), medium (1001-8000 students)
and large (8001 +)
 Decide how many to select from each group
 Decide if proportional or nonproportional stratified
random sampling
Stratified sampling
1.Proportional – draw out the school districts in the same
proportions at which at which they are represented in the
study population
Example: 600 small school districts, 350 medium and 50
large districts; desired sample is 100 (divide by
ten); thus 60 small, 35 medium and 5 large
2. Nonproportional – equal representation from each group of
the study population
Example: Getting samples equally among a group of
students – 400 freshmen, 300 sophomores, 200
juniors and seniors 100
- Get 25 samples from which level regardless of the number
of population
Cluster (or Area) Sampling
Study population is composed of groups or cluster or
areas
 Create a sampling frame of all the clusters in a
study population and randomly select the desired
number of clusters from the frame
 Example: Survey a sample of students enrolled in a
general education course about their health
behavior
 - Create a list of all the sections of all general
education courses offered for the semester
- Select a simple random sample of course sections
(clusters)
Multi stage cluster (or area) sampling
 Combination of cluster sampling and simple random
sampling techniques in which a series of random
selections is made from units of progressively smaller
size
Example: The country’s department of education is
interested in measuring the impact of the of the
country’s new nutrition education program on fourth
graders
- from a list of provinces in each region, select the
provinces, then towns, school districts selected from
each town, Elementary schools selected, then
classrooms selected
Matrix sampling
 Used when data collection involves having a large number of
participants respond to a large number of questions and when
researchers are only interested in making inferences about the
study population as opposed to individuals
Example: Instrument is lengthy (200 items)
- Divide the 200-item instrument into 10 smaller instruments
each containing 20 questions from the original instrument
- Instead of only ne participant answering all 200 questions, 10
randomly selected participants will answer 20 questions each
- Answers from the 10 participants will comprise one
completed response
Non-probability Sampling
 Participants can be included on the basis of
convenience (because they have volunteered or
because they are available or can be easily
contacted) or because they have a certain
characteristic
 Uses human judgment in selecting respondents
thus there is an increased chance of sampling bias
and there is no theoretical basis for estimating
study population characteristics
 Often used in qualitative studies
Nonprobability Sampling Procedures
Sample Primary Descrptive Elements
Convenience Includes any available subject meeting some minimum
criterion, usually part of an accessible intact group
Including people who are available, volunteer or can be
easily be recruited in the sample
Volunteer Includes any subject motivated enough to self-select
for a study
Grab Includes whomever investigators can access through
direct contact, usually for interviews
Homogeneous Includes individuals chosen because of a unique trait or
factor they possess
Nonprobability Sampling Procedures
Sample Primary Descriptive Elements
Snowball Includes subjects identified by investigators, and any
other persons referred by initial subjects
Each research participant is asked to identify other
potential research participants
Judgmental Includes subjects whom the investigator judges to be
(Purposive ) “typical” of individuals possessing a given trait
Researcher specifies the characteristics of the total
population of interest and locates those individuals who
have those characteristics
Quota Includes subjects chosen in approximate proportion to
the population traits they are to “represent”
Researcher determines the appropriate sample size or
quotas for the groups identified as important and takes
convenience samples from those groups
Convenience Sampling
 Selects participants based on certain inclusion criteria and their
accessibility and proximity to the researcher
 Most frequently used because of time and money saving
 Collect large amounts of data in a short period of time
 Weakest of all sampling methods
 Drawn by selecting any available participant from an intact group
 Drawn by selecting any available participant from an intact group
Examples of intact groups:
Students in a classroom, employees of a certain company, or
people who voluntarily show up at certain place for a specific reason
(e.g. members of a support group or those receiving free flu shots
from the health department)
Volunteer Sampling
 Volunteer samples – samples that include participants
motivated enough to self-select for the study
 Researcher is unable to identify the study population, thus
volunteers are selected
Example:
Studying the exercise behaviors of people when they
have colds would have a hard time identifying a suitable
population. Hence researcher puts in the ads of a
newspaper asking for volunteers to participate; testing
the effects of a new drug on insomnia or indigestion
participants maybe eager for the treatment (may need
incentive)
 May not be a representative of those who do not volunteer
Grab Sampling (Chunk sampling)
 Includes
whomever researchers can access
through direct contact and is
representative of no study population
Example:
Individuals asking people questions about
their shopping habits as part of some
marketing research
Homogeneous Sampling
 Used when certain research problems pertain only
to a very specific group of individuals who
possess a unique trait or factor thus individuals
are selected for this trait
Example:
In-depth interviews with individuals who have
been named outstanding mentor/outstanding health
educator in their organization for the past ten years
Judgmental Sampling (Purposive Sampling)
 Researchers select participants that they judge to be
“typical” of individuals possessing a given trait
Example: Conducting evaluation research to determine
why employees have enrolled in the worksite health
promotion program would want to interview “typical
enrollees” to find the answers
 Of greatest concern is to determine what is typical
 Try to achieve broad consensus about what is typical so
that the selection of participants reflects the criteria
Snowball Sampling (Chain sampling)
 Multistage sampling method that literally “snowballs”
 Used when it is difficult to identify the study population
because of the sensitive nature of the research problem (e.g.
issues dealing with sexuality or diseases that carry a social
stigma)
 Begins when participants with certain characteristics are
identified and interviewed; they are asked if they know others
with the same characteristics who could be included in the
same sample
 Second group of people is sought out, interviewed, and asked if
they know about others.
 The process continues until the desired number of participants
are contacted
Quota Sampling (equivalent of stratified sampling)
 Used when researchers want a ample to be
representative of certain characteristics of a study
population and are unable to select a probability sample
 Includes those participants who are “chosen in
approximate proportion to the population traits they are
to represent
 Used when the parameters are known and a probability
sample cannot be selected
Example:
Selecting a quota sample of 200 students in a school
having an enrollment in grades 9-12 of 55% female and 45%
male (thus 110 females – 55% and 90 males – 45%)
Determining Sample Size for
Quantitative Studies
Reasons why sample size is necessary :
1. The sample must represent the study population so that
reliable statistical inferences can be made about the
study population.
- The characteristics identified in the sample must be
the same/very similar to those of the study population
2. To achieve statistically significant results
- The larger the sample, the smaller the sampling
error and the greater the chance of getting statistically
significant results
Times when larger samples are needed
1. Whenthe sampling units are highly heterogeneous
on variables being studied
2. When
researchers want the sample estimates to be
more precise
3. When small differences in results are anticipated
4. When using survey research methods
5. When
high attrition (e.g. mortality, dropouts)
among the sampling units is expected
Times when larger samples are needed
6. When a large number of uncontrolled variables
(extraneous/confounding variables) are present
7. When the samples being studied must be broken into
subsamples
8. When the level of statistical significance, statistical power,
or both must be high
9. When the reliable measures of the dependent variable are
not available
10. When there are a large number of variables being
analyzed
Sample size requirements for different research traditions
Condition Sample requirement
Pilot studies At least 10 events

Studies of physiological At least 8 events randomly selected (if individual


measures response ratings are involved, for experiments it must be
increased

Validation or cross validation At 200 events if standard methods of validation are uses
(such as multiple correlation or factor analysis); with
other tools “N should be at least about 60, and
preferably 100 or more.

Studies that use multiple Fifteen subjects to every predictor or independent


correlation methods variable

Studies that use factor At least ten events for each variable item to be
analysis included in the factor analysis
Sample size at 95% confidence interval
Population Sample size
Under 10 NA
10 10
15 14
20 19
30 28
40 36
50 44
75 63
100 80
250 152
500 217
1000 278
Sample size
 Sample size and precision of population estimates (95%
confidence level (by Paul C. Cozby, Methods in Behavioral
Research p. 139)

Size of 3% 5% 10%
Population
2000 696 322 92
5000 879 357 94
10000 964 370 95
50000 1,045 381 96
100000 1,056 383 96
Over 100000 1,067 384 96
Determining Sample Size for Qualitative
Studies
 No rules for sample size in qualitative inquiry
 Less structured and is based more upon the research approaches
 Qualitative studies collect much more detailed data from a smaller
number of people
 Samples are judged by the extent to which they are informationally
representative rather than statistically representative
 Sample size is left to the judgement of the researcher
Example:
Knowing why senior citizens from a specific community are not
obtaining free vaccines from the health department (data collected
reaches a point of redundancy/saturation
Sampling in Qualitative Research
 Criterion-based selection - researcher develops inclusion criteria to be used
in selecting people or other units (similar to purposeful/purposive sampling)
1. Comprehensive sampling – all relevant cases (individuals, groups, settings,
or other phenomena) are examined in the research study
2. Maximum variation sampling – a wide range of cases are purposely selected
so that all types of cases along one or more dimensions are included in the
research
3. Homogeneous sample selection – selecting a small and homogeneous case or
set of cases for intensive study (Used by focus group researchers)
4. Extreme-case sampling - the extremes or poles of some characteristics are
identified and then cases representing only the extremes are selected for
examination (example: compare the teaching environment created by an
outstanding teacher with that created by a notably ineffective teacher
Sampling in Qualitative Research
5. Typical-case sampling – researcher lists the criteria that describe a
typical or average case and then finds one or several to study
(e.g., a hypothetical teacher of interest might be described on
characteristics such as age, gender, teaching style and number of
years of experience)
6. Critical-case sampling – cases that can be sued to make a
previously justified point particularly well or are known to be
particularly important are selected for in-depth study (e.g., a
school superintendent wants to change a policy and expects that
change to face resistance in the local schools. He might decide to
select a school where he/she expects the greatest resistance to
determine whether enacting the policy is feasible in practice)
Sampling in Qualitative Research
7. Negative-case sampling – cases that are expected to disconfirm
the researcher’s expectations are purposely selected (e.g.,
researcher typically explores a phenomenon and attempts to
build a theory inductively about it)
8. Opportunistic sampling – researcher takes advantage of
opportunities during data collection to select important cases
(cases might be critical, negative, extreme or even typical;
researcher is quick to discern whom to talk to, and what to
focus on while collecting the data in the field
9. Mixed purposeful sampling – mixing of more than one strategy
(e.g. start with maximum variation sampling, then use negative
-case sampling)
Sample size in qualitative studies depend
on:
1. What the researchers want to know
2. The purpose of the study
3. What is at stake based upon the outcomes of the study
4. What will be useful data
5. What will have credibility with others interested in the
study
6. What can be accomplished with the available time and
resources

You might also like