Slovin's formula, somtimes also spelled "Sloven's forumula (e.g., Altares et al. 2003, p. 13), is an ad hoc formula lacking mathematical rigor (Ryan
2013) that gives an estimate of the sample size needed
to obtain statistically meaningful results when sampling from a population of unknown
characteristics. If 
is the population size and 
is the allowed probability of an error in a small representative
sample, then the number of samples 
needed is approximately
 |
The formula was supposedly published by Slovin in 1960 (with reference reportedly given in Guilford and Frucher 1973) and independently by Yamane (1967).
It is unclear exactly how useful the formula is in practice due not only to the imprecise definition of 
,
but also because although the size of a sample must depend on population variability
in the measured property, Slovin's formula lacks a parameter representing such variability
(Ryan 2013.)
