Power Analysis


What Is The Power Analysis Package ?

The Power Analysis package ‘simr’, developed by Peter Green, in collaboration with Catriona MacLeod, at Landcare Research, provides tools that assess the power of different monitoring designs to detect trends or changes in monitoring data. The package is built for use in ’R’, an integrated suite of open-source software facilities that is widely used around the world. The package has been designed to provide a platform that allows for other packages to be added in the future. 
It addresses some recurrent challenges in reporting:

  • providing early warnings (e.g. overcoming difficulties with interpreting results or lack of confidence in results), 

  • learning for sustainability (e.g. power to detect the effect of a management intervention)

  • fine-tuning monitoring (e.g. identifying cost-effective designs to meet their specific monitoring goals and needs, by evaluating trade-offs in monitoring designs)

 

Link to the package:
https://cran.rstudio.com/web/packages/simr/index.html
Feedback and interest - The package has recently been publicised via online publications and conference presentations attracting great interest. It is already being applied in different areas, with over 1400 downloads (via RStudio alone)!


Use Of The Power Analysis Package

Power analysis package used within NZSD

285457_11.png

Utility of the package using NZSD industry stakeholders’ datasets to evaluate the power of existing sampling designs for detecting trends has been demonstrated.

Poster "How much data is needed for effective monitoring?" including a case study on Nitrogen use trends on kiwifruit orchards.

Prepared for the Stakeholder workshop - 6 August 2015

Power analysis package used by MfE

2612162_11_orig.png

This package was used to evaluate a citizen science dataset (the NZ Garden Bird Survey) for potential use in national environmental reporting. This work was carried out in collaboration with the MBIE-funded project, Building Trustworthy Biodiversity Indicators, and an external stakeholder of the NZSD project, the Ministry of Environment (MfE). 

September 2015: MfE just released the "Use of NZ Garden Bird Survey data in environmental reporting: Preliminary models to account for spatial variation in sampling effort" report. 

This report has been prepared by Catriona MacLeod and Peter Green (and others) from Landcare and is part of the Power Analysis Research topic. 


Other Information

December 2015
Following his presentation at SEEM 2015, Peter Green has been invited to run a workshop and give a talk at the Eco-Stats conference (Technological advances between Ecology and Statistics) in Sydney - 8-10 December 2015.

__3475564.gif

November 2015
Peter Green has presented the power analysis package (Power analysis for GLMMs by simulation) at the NZAS (NZ Statistical Association and Operations Research Society of NZ) in Christchurch - 24-26 November 2015. 


_7630067.jpg

November 2015
Catriona MacLeod has presented "Refining monitoring and reporting in production landscapes and beyond" introducing the power analysis package with an example from the kiwifruit sector at the NZES 2015 (New Zealand Ecological Society conference) in Christchurch - 16-19 November 2015.


June 2015
Peter Green has presented the new statistical package developed for power analyses in the statistical software, R, to a mixed audience of national and international academics and stakeholders involved in statistical analysis of ecological and environmental monitoring data - at SEEM 2015 (Statistics in Ecology and Environmental Monitoring conference) in Queenstown (NZ).

This resulted in an invitation to run a workshop at the Australian Eco‑Stats conference in December 2015 for international statisticians and ecologists.  It confirmed as well the opportunity for this package to be reused, developed further or adapted by statisticians and ecologists around the world for all manner of natural resource management and industrial production that step far beyond our agricultural sustainability agenda.