Abstract
A comparison is made between the use of a stratified universe and an unstratified universe of cultures for sampling purposes. Samples are taken from the Ethnographic Atlas (1) by selecting from culture clusters (a first level of stratification of cultures) and (2) by a random selection from the total list of cultures in the Atlas. The analysis is done regionally, for Africa and North America. A chi-square, two-sample test indicates the essential similarity of the stratified and the unstratified universes; no significant statistical difference is found among the samples.
Matched-pair, signed-ranks tests are used to determine (1) the comparative stability of successive samples taken from the stratified universe and those taken from the unstratified universe and (2) the interchangeability of cul tures within culture clusters. Results for Africa indicate no significantly greater similarity of successive samples based on culture clusters compared with successive random samples based on the totality of cultures. For North America, successive samples from culture clusters seem to be significantly more stable than random samples from the totality of cultures only in regard to Ethnographic Atlas data. General interchangeability of cultures within culture clusters is not borne out by the data examined.
Get full access to this article
View all access options for this article.