Journal article
Better inference from size-selectivity and catch-comparison studies using simultaneous bootstrap confidence bands
Fisheries research, Vol.300, pp.1-5
08/2026
Metrics
1 Record Views
Abstract
The use of 95% pointwise bootstrap confidence bands for assessing plausible shapes of size-selectivity curves and length-based catch-comparison curves has become ubiquitous within the published literature. It is now known that the widespread practice of using these pointwise confidence bands for formal inference about the curve suffers from the multiple comparisons problem, and that in practice they only provide approximately 75% confidence of including the true curve. Due to the popularity and familiarity with the pointwise confidence bands there is some ongoing reluctance to adopt more reliable methods of inference such as permutation tests. Here, a simultaneous-inclusion bootstrap confidence band is developed to resolve the multiple comparisons problem and permit continued use of bootstrap confidence bands. This corrected approach has been added to the Rpackage SELECTand is demonstrated here using a case study provided with that package. Nonetheless, this corrected approach does not overcome the potential for confidence bands to be innately misleading in terms of indicating plausible shapes of the true underlying curve. A statistically pragmatic and defensible workflow for future selectivity and catch-comparison studies is proffered.
Details
- Title
- Better inference from size-selectivity and catch-comparison studies using simultaneous bootstrap confidence bands
- Creators
- Russell B. Millar - University of AucklandMatt K. Broadhurst - Southern Cross University
- Publication Details
- Fisheries research, Vol.300, pp.1-5
- Publisher
- Elsevier B.V.
- Identifiers
- 991013390851802368
- Copyright
- © 2026 The Authors.
- Academic Unit
- Faculty of Science and Engineering; National Marine Science Centre
- Language
- English
- Resource Type
- Journal article