Non-Probability Sampling
Non-probability samples have the distinguishing characteristic that
subjective judgments played a role in the selection of the sample(1).
In non-probability sampling, every participant has an unknown chance of
being selected(2,3). This is unlike probability sampling
which uses randomization to ensure selection without subjectivity and has
the characteristic that every element in the population has a known nonzero
probability of being included in the sample.
Convenience Sampling
Purposive Sampling
Snowball Sampling
Quota Sampling
Advantages of using Non-Probability Techniques
Disadvantages of using Non-Probability Techniques
References
Return to home
Convenience Sampling
The researcher selects cases based on their availability for the study.
In other words, the researcher surveys people who can be easily reached
or who are readily available to participate in the study.
-
example:A professor surveys a college class to find out about Americans'
attitudes towards gun control.
-
example:A classic example is a Literary Digest poll of a very large
sample, that predicted Franklin Roosevelt would lose the 1936 presidential
election when in fact he won by a landslide. The major problems with this
sample was it was "conveniently" selected from automobile registrations,
telephone directories and related sources which led to oversampling of
affluent individuals who were not representative of the voting public(2).
Purposive Sampling
Individuals, or cases, are selected by an experienced investigator based
on some appropriate characteristic required of the sample members. The
cases that are selected are considered to be most representative of the
population of interest as a whole(4). To increase credibility
of the design, cases that are considered unique or special are excluded
from the sample(1).
-
the researchers’ judgment and knowledge of the population is critical
-
there is much potential for selection bias
-
one might just "appear" to be similar to the population when in fact they
are very different
-
reliability of the results can not be tested mathematically
Example: If you wanted to estimate the total number of a service
performed during a given year in an outpatient clinic you might use purposive
samping and select a few typical days and review the records for these
days. This would be better than using a random sample of days if one only
had the resources for a few days since the ability to use judgment in this
approach would allow one to pick days that would be typical, rather than
risking the possibility of including unusually high or low days(4).
**Fetal Origin of Disease
Snowball Sampling
This non-probability sampling technique relies on previously identified
group members to identify other members of the population, hence the sample
grows like a "snowball". This is used when a population listing is unavailable
and cannot be compiled by researchers, such as illegal drug users or illegal
aliens(1).
Quota Sampling
Quota sampling is done to ensure that different subgroups in a population
are representative of the sample characteristics of interest. This technique
divides the population group being studied into subgroups (such as race,
gender, ethnicity) and then based on proportions of the subgroups needed
for the final sample, interviewers are given a number of units from each
subgroup that they are to select and interview(1). The
problem is interviewers tend to select individuals they find easy to talk
to, who are not intimidating or hostile, who are willing to participate,
and who are most like themselves(5). This form of sampling
is usually used in commercial studies, not academic.
Advantages of Non-Probability Sampling Techniques
-
useful and quick method in certain circumstances
-
might be only method available, such as if sampling illegal drug users
-
if researchers are truly interested in particular members of a population,
not the entire population
-
exploratory research attempting to determine whether a problem exists or
not, such as a pilot study
In summary, non-probability sampling techniques are useful when there
are limited resources, an inability to identify members of the population,
and a need to establish the existence of a problem(1).
Disadvantages of Non-Probability Sampling Techniques
-
the subjectivity of non-probability sampling prevents making inferences
to the entire population
-
validity and credibility questionable due to selection bias
-
the reliability of the resulting estimates can not be evaluated which results
in the user not knowing how much confidence can be placed in any interpretations
of the survey findings(1).
References
1. Henry GT. Practical Sampling.
Newbury Park: Sage Publications, 1990.
2. Morgan G, Harmon R. Sampling
and External Validity. Journal of the American Academy of Child and Adolescent
Psychiatry, 1999;38:1051-53.
3. Cox BG, Cohen S. Methodological
Issues for Health Care Surveys. New York: M. Dekker, 1985.
4. Levy P, Lemeshow
S. Sampling of Populations : Methods and Applications. New York: Wiley,
1991.
5. Weisberg HF. An Introduction
to Survey Research, Polling, Data Analysis, 3rd ed. California: Sage Publications,
1996.