U.S. flag

An official website of the United States government, Department of Justice.

Methodological and Design Issues: Techniques for Assessing the Representatives of Snowball Samples (From Collection and Interpretation of Data From Hidden Populations, P 31-43, 1990, Elizabeth Y Lambert, ed. -- See NCJ-128609)

NCJ Number
K M van Meter
Date Published
13 pages
Recent research efforts have provided a formal and operational tie between ascending and descending methodologies used to study hidden populations. Snowball sampling and hierarchically ascending (HA) classification analysis are useful once problems with the calculation of general population estimates and the explanation of variance are resolved.
There is an opposition between studying social groups not accessible through established means of sociological research and using research methodologies based on intensive, detailed, data collection strategies. There is also a clear distinction between ascending and descending methods in the case of classification analysis; the classes obtained by HA classification are called polythetic classes. Cross-classification analysis is the Cartesian crossing of two automatic classification analysis and snowball sampling are compatible in ascending methodology and have been used to study drug abusing behavior in several European cities. When conducting this sort of analysis, researchers must be aware of possible random error or sampling error, sample bias, and response bias which are the major sources of inaccuracy in data collection. 31 references