-
Views
-
Cite
Cite
Raphael Gillett, Sampling with Fallible Memory: An Occupancy Problem, Journal of the Royal Statistical Society Series D: The Statistician, Volume 45, Issue 3, September 1996, Pages 299–305, https://doi.org/10.2307/2988468
- Share Icon Share
Summary
The act of sampling from a small finite population is an everyday human activity, the success of which depends on the efficacy of people’s working memory for previously sampled objects. Accordingly, a means of assessing working memory capacity is often desired. An extension of the classical occupancy distribution supplies a framework for testing hypotheses about memory capacity, using a model of working memory in which an m-slot buffer holds a record of m previous selections from a population of n objects. A person does not commit the error of resampling objects within the buffer, and samples with replacement objects outside the current memory set.