Premium
The selection of suitable indices for the measurement and analysis of fish condition
Author(s) -
Bolger T.,
Connolly P. L.
Publication year - 1989
Publication title -
journal of fish biology
Language(s) - English
Resource type - Journals
SCImago Journal Rank - 0.672
H-Index - 115
eISSN - 1095-8649
pISSN - 0022-1112
DOI - 10.1111/j.1095-8649.1989.tb03300.x
Subject(s) - statistics , index (typography) , set (abstract data type) , fish <actinopterygii> , measure (data warehouse) , selection (genetic algorithm) , data set , biology , condition index , correctness , regression analysis , mathematics , statistical analysis , data mining , fishery , computer science , algorithm , artificial intelligence , world wide web , programming language
A survey of the studies published in two leading fisheries journals revealed that the analysis and measurement of condition, based on length‐weight data, has been performed using a wide variety of indices and statistical procedures. Eight forms of index were identified which can be categorized into those which measure the condition of individual fish, i.e. condition factors‘, and those which measure the condition of subpopulations as a whole, i.e. regressions of log, 10 weight on log, 10 length and the parameters of such regressions. Analysis of a test data set indicated that both the form of index and properties of the data set size can dictate the patterns of condition observed. The various indices were reviewed in terms of appropriateness, simplicity and statistical correctness. It was concluded that an index should be selected only after a detailed examination of both the underlying assumptions of the index and the properties of the data set.