Premium
A comparison of randomized intervention analysis and two‐factor analysis of variance for analysis of angler effort in a floodplain river fishery
Author(s) -
Cloutman D. G.,
Jackson D. C.
Publication year - 2003
Publication title -
fisheries management and ecology
Language(s) - English
Resource type - Journals
SCImago Journal Rank - 0.693
H-Index - 55
eISSN - 1365-2400
pISSN - 0969-997X
DOI - 10.1046/j.1365-2400.2003.00330.x
Subject(s) - statistics , analysis of variance , mathematics , repeated measures design , one way analysis of variance , statistical analysis , stocking , pooled variance , fishery , biology , confidence interval
Probability values were compared among randomized intervention analysis (RIA) and untransformed and log e ‐transformed two‐factor analyses of variance ( anova ) to test effects of fish stocking and a public relations programme on angler effort. The data set consisted of a time series of angler vehicle counts at single control and impact locations in a before–after‐control‐impact (BACI) experimental design. A randomized complete block anova indicated that probability values from RIA and the untransformed and log e ‐transformed two‐factor anova did not differ significantly ( P = 0.345), and that they did not result in different conclusions concerning angler effort. Randomized intervention analysis is superior to two‐factor anova in not having to conform with assumptions of parametric statistics, and it tests for temporal autocorrelation. Two‐factor anova is superior to RIA in providing statistical inferences about possible differences in main effects, i.e. control and impact locations and pre‐ and post‐treatment periods, if the interaction term is non‐significant. To achieve the most comprehensive analysis, it is recommended that both RIA and two‐factor anova (whether untransformed or transformed) be performed during field experiments of angler effort when using a BACI design at single control and treatment locations.