Title
Recent Additions to the Data Analysis Tool Box
Author(s)
Meredith, Michael; Juat, Ngumbang
Published
2017
Abstract
New data analysis methods have been introduced into ecology in the last 20 years, in particular information criteria andBayesian analysis, while hypothesis testing has been de-emphasised. The introduction of Akaike‘s Information Criterion(AIC) shifted the focus from assessing evidence for individual effects to selection of the best models from among anumber of biologically-plausible and potentially complex models. This was applied for mark-recapture studies, includingspatially explicit capture recapture (SECR), and to occupancy estimation when these were introduced. Bayesian methodsusing BUGS software were developed for medical research and began to be used by ecologists around 2010. Fundamentaladvantages of the Bayesian approach are that output is suitable for decision making and prior information can beincorporated in a formal way. The BUGS language provides a simple, transparent way to specify models, givingecologists huge freedom to choose models appropriate to the specific situation they are investigating. This has led to thedevelopment of (1) integrated population models combining different types of information, e.g. fecundity, abundance andsurvival; and (2) hierarchical models allowing information to be shared across subsets of the data, eg. tiger surveys withvery few detections in some areas. Modelling of community occupancy or abundance shares information across species,providing estimates of species richness, biodiversity indices, and similarity measures which take account of probability ofdetection.
Keywords
Statistics;information theory;Bayesian analysis;hierarchical models
Access Full Text
A full-text copy of this article may be available. Please email the WCS Library to request.
![]()
Back
PUB21146