An Original Solution for Completing Research through Snowball Sampling—Handicapping Method

Abstract

Snowball sampling, as a network research, has a number of advantages in registering “hidden populations”, such as: people with rare diseases and unrecognized by the community, illegal immigrants, drug users, people with deviant sexual behavior, undeclared workers, alcoholics, gambling or computer addicts, unemployed jobseekers, domestic and school violence, etc. The procedure is well known and widely applied, as are its drawbacks, the most important of which is the ambiguity regarding the size of the sample under investigation, as the number of interviewees cannot be predetermined, but it is up to the researcher to stop the survey when he/she considers the accumulation of information to be satisfactory or when the imposed limit of the research budget has been reached. The major disadvantage of this method of survey is the absence of objective, quantitative, scientifically based criteria, on the decision to stop the survey and this is precisely the aspect proposed to be resolved in the present paper. Our proposal, the handicapping method, adapts and develops the Wald test of the likelihood ratio for the case of investigating populations in terms of binary characteristics, a typical situation for sociological research. The procedure complements the snowball method, providing a completion under the conditions of assumed statistical risks. The results obtained in this original material can make a major contribution to the improvement of sociological research methods in a special field—sociological field survey/research in closed groups and non-visible populations.