Sampling Designs

Sampling Designs

            Sampling designs vary dependant upon the circumstances.  The most prominent methods are probability sampling and nonprobability sampling.  Probability sampling refers to the inclusion of the entire populace within the sampling based on the representation of each sector (Krosnick, 1999).  On the other hand, nonprobability sampling is defined as an inability to predict such inclusion (Leedy & Ormrod, 2010).  Of the two, probability sampling is typically deemed more accurate, yet depending upon the purpose of the research nonprobability may be equally or perhaps more efficient (Krosnick, 1999).

From these designs, 8 sampling approaches exist.  Under the umbrella of probability sampling lies simple random sampling, stratified random sampling, proportional stratified sampling, cluster sampling, and systematic sampling.  Approaches associated with nonprobability sampling include convenience sampling, quota sampling, and purposive sampling (Leedy & Ormrod, 2010).  The following describes these sampling approaches.

Probability Sampling:

1)      Simple Random Sampling – Participants are randomly selected, given equal opportunity of being selected.  This approach is geared towards small populations.

2)      Stratified Random Sampling – The sampling is separated into random layers or groups and utilizes equal representation from each group.

3)      Proportional Stratified Sampling – Is similar to the stratified random sampling yet instead of equal representations, this method utilizes proportional representations of each group.

4)      Cluster Sampling – Is most appropriately utilized to analyze larger populations or those consisting of individuals throughout an extended area (unlike simple random sampling which is most often used to examine small populations).  By observing subdivisions or smaller collections, a sampling may be gathered.  In addition, each cluster should demographically resemble one another.

5)      Systematic Sampling – Involves an orderly methodical approach to sampling.  For example, individuals may be separated into clusters, then every 5^th cluster may be selected to observe (Leedy & Ormrod, 2010).

Nonprobability Sampling:

1)      Convenience Sampling – Utilizes those accessible and willing to participate.  Whoever appears is considered.  This method may be appropriate to test customer-service or gather feedback concerning new equipment or programs.

2)      Quota Sampling – Refers to the sampling of a certain number of participants regardless of their make up.  The representation of society or the populace is disregarded.

3)      Purposive Sampling – Analyzes individuals with a particular purpose or perspective in mind.  An example would include a sampling of individuals who support same sex marriage (Leedy & Ormrod, 2010).

Now, in regard to sampling, one must consider bias as well.  The concept, sampling bias refers to any circumstance or persuasion that manipulates the research.  Considering such occurrences, personal bias or undetected factors as well as other influences may cause sampling bias (Leedy & Ormrod, 2010).  Additionally, sampling bias and representativeness of the sample have the potential to affect the generalizability of the research conclusions.  Particularly, by means of probability sampling, if the researcher selects a certain representation which is not accurately reflective of the populace, the data collected may be invalid (Leedy & Ormrod, 2010). 

References:

Leedy, P. D. & Ormrod, J. E. (2010). Practical research: Planning and design (9th ed.).

Upper Saddle River, N. J.: Pearson Education, Inc.
       

Krosnick, J. A. (1999). Survey research. Annual Review of Psychology, 50, 537-67.