Desktop version

Home arrow Engineering arrow Behavioral Intervention Research: Designing, Evaluating, and Implementing

Source

SAMPLING METHODS

The process of selecting a representative sample for a study is called “sampling.” There are two main categories of sampling approaches: probability sampling and nonprobability sampling. Within these categories, there are various methods or procedures that are used to select samples for intervention studies, particularly at the trial phase. These approaches are summarized in Table 9.1. The choice of sampling approach and procedure should be based on the intent of the intervention, specific research goals and questions, the study design, information available about the target population, resources available, and the stage along the pipeline. For example, one typically does not use a probability sampling approach when conducting focus groups early on in the pipeline to gather initial information about perceptions of the need for an intervention. As discussed later in this chapter, this type of approach is more likely to be used when engaged at different stages in the pipeline or in survey research. Before we discuss these strategies, we define the concept of “sampling frame”—a key concept in sampling.

A sampling frame is the “list” or source material that is used to select a sample from a population. For example, if you were conducting a study that was evaluating an intervention to foster safe sex practices among high school students in a particular school region, the sampling frame would be a list of all registered high school students in that region. Examples of sampling frames include an electoral register, telephone directories, employment records, school class lists, patient files in a clinic or hospital, organizational lists, and so on. A sampling frame must be representative of the target population. Oftentimes, a complete sampling frame does not exist. For example, assume that one is relying on use of a telephone directory to conduct a survey about the prevalence of family caregiving within a particular geographic region. This sampling frame would be incomplete, as it would not include people with unlisted numbers or those who have temporary cell phones.

A sampling frame may also be unavailable. A work organization, for example, may not be willing to provide a list of employees; a clinic may not be willing to share the names of patients. In other cases, a sampling frame may not exist because the target population is challenging to identify or reach or may remain hidden—for example, those whose behaviors are illegal (e.g., drug abusers) or individuals who are reluctant to be identified as having a particular characteristic (e.g., persons affected with a specific illness or condition) (Magnani, Sabin, Saidel, & Heckathorn, 2005). As described later, in these cases, other methods such as “snowball sampling techniques” (an initial number of the sample is identified and recruited and helps to identify other individuals to be included in the sample) are used to identify and recruit research samples. In these cases, it is difficult to recruit a representative sample.

TABLE 9.1 Summary of Sampling Methods

Sampling Method

Summary Description

Probability sampling

All elements (e.g., individuals, skilled living facilities) in the target population have some opportunity of being included in a sample, and the probability of being included in the sample is known for each element in the population.

Simple random sampling

Each member of the population has an equal probability of being included in the sample.

Systematic sampling

Selecting every nth unit of the target population from a list that is randomly ordered.

Stratified sampling

Dividing a population into groups or strata (e.g., age group, race/ethnicity) and then randomly selecting from that group.

Cluster sampling

Generally a two-staged process. Initially, the total population is divided into clusters or groups, and then a random sample of clusters is selected. In the second stage, a random sample is selected from within each of these clusters.

Nonprobability sampling

Sample members are selected on the basis of availability. In others words, everyone in the target population does not have a chance of being included in the sample; the selection of members is nonrandom.

Convenience sampling

Sample members are selected on the basis of conve- nience—they are available and convenient (e.g., caregivers who attend community support groups).

Quota sampling

Similar to convenience sampling, but the goal is to select a certain quota or number of members of a sample that have a certain characteristic (e.g., socioeconomic status).

Purposive sampling

Sampling a specific group of individuals to address a very specific need or purpose (e.g., those who did not do well in a particular training program).

Snowball sampling

Gathering data from a few members of the target population and then asking those members for information regarding the location of other potential members of the population.

Adaptive allocation sampling

Adaptive allocation sampling, similar to cluster sampling, is a staged sampling approach. The initial sample is obtained using a conventional approach such as random sampling and then that sample is examined to determine if there are some geographic areas that exhibit more of the behavior/ phenomena of interest on the basis of observations from a few select variables.

 
Source
Found a mistake? Please highlight the word and press Shift + Enter  
< Prev   CONTENTS   Next >

Related topics