Area sampling is a special form of cluster sampling in which the sample items are clustered on a geographic area basis. For example, if one wanted to measure candy sales in retail stores, one might choose a sample of city blocks, and then audit sales of all retail outlets on those sample blocks.
The practical motivation underlying area sampling is that for many problems there is no current and accurate list of universe elements. For example, a complete list of the many different retail sources for candy would not exist. So it would be impossible to choose a probability sample of these outlets directly, using methods previously described. A solution is to use area sampling – first selecting a sample of geographic areas, then studying the particular universe units associated with the selected sample of geographic areas.
The process can be conceptualized this way. The original universe of interest for which there is no list is transformed into a universe for which there is a list. Such a list consists of city blocks, or ZIP code areas or counties or other geographically defined areas that can be identified on maps. A rule of association, uniquely linking each item in the universe of interest to a single physical area, is established. (In the candy outlet illustration, all outlets located on any block are ‘associated’ with that block) By drawing a probability sample of areas and using the rule of association, one obtains a probability sample from the universe of interest.
Application of Area sampling:
The basic idea of area sampling is both simple and powerful. It enjoys wide usage in situations where very high quality data are wanted but for which no list of universe items exists. For instance, many governmental agencies (e.g. Bureau of Labor Statistics) use area sampling.
However, the practical execution of a large scale area sample is highly complex. Typically an area sampling is conducted in multiple stages, with successively smaller area clusters being sub-sampled at each stage.
Example: A national sample of households is often constructed in a series of steps like this:
1. Create geographic strata, each consisting of a group of counties in more or less close proximity. Fifty or more such strata, containing all of the roughly 3,000 US counties, are commonly used.
2. Within each geographic stratum, choose a probability sample of one or more counties (or groups of counties such as metropolitan areas).
3. Within each sample county (or group of counties), choose a probability sample of places (cities, towns, etc).
4. Within each sample place, select a probability sample of area segments (blocks in cities, area with identifiable boundaries in other places, etc)
5. Finally, within sample segments choose a probability sample of households.
The details of constructing such an area sample will vary, depending upon the research situation.
Concluding Comments on Area sampling: Execution of area sampling designs is expensive and time consuming. They also require that researchers address a number of important questions concerning the sample design. For example, how strata should be built; how area sampling units should be defined at each stage (areas such as counties or places may be grouped together in many different ways); how many area sampling units should be selected at each stage and what selection process should be used; and so on. For best results, substantial information (detailed maps, statistical data by areas, cost data) is needed and expert statistical counsel is required.
Because of these considerations ‘pure’ area sampling is used only infrequently in marketing research practice. However, the reader should be aware of the concept for two reasons. First, many important secondary data (especially from the government) are derived from area sample. Second, the principle of area sampling is often used, in modified form, to provide a framework for the design of marketing research samples.