Premium
Sample‐Size Requirements for Evaluating Population Size Structure
Author(s) -
Vokoun J. C.,
Rabeni C. F.,
Stanovick J. S.
Publication year - 2001
Publication title -
north american journal of fisheries management
Language(s) - English
Resource type - Journals
SCImago Journal Rank - 0.587
H-Index - 72
eISSN - 1548-8675
pISSN - 0275-5947
DOI - 10.1577/1548-8675(2001)021<0660:ssrfep>2.0.co;2
Subject(s) - sample size determination , statistics , frequency distribution , sample (material) , mathematics , sampling (signal processing) , similarity (geometry) , exponential function , population , computer science , artificial intelligence , mathematical analysis , physics , filter (signal processing) , image (mathematics) , computer vision , thermodynamics , demography , sociology
A method with an accompanying computer program is described to estimate the number of individuals needed to construct a sample length‐frequency with a given accuracy and precision. First, a reference length‐frequency assumed to be accurate for a particular sampling gear and collection strategy was constructed. Bootstrap procedures created length‐frequencies with increasing sample size that were randomly chosen from the reference data and then were compared with the reference length‐frequency by calculating the mean squared difference. Outputs from two species collected with different gears and an artificial even length‐frequency are used to describe the characteristics of the method. The relations between the number of individuals used to construct a length‐frequency and the similarity to the reference length‐frequency followed a negative exponential distribution and showed the importance of using 300–400 individuals whenever possible.