A new era for ERA? Using Bayesian Network Models to improve ERA, with Jannicke Moe

Credit: antishock – stock.adobe.com

“Not difficult, just different,” is how one author in the latest IEAM podcast describes Bayesian Network models (BNs) to researchers that are unfamiliar with—and often intimidated by—them. A recent special series aims to dispel the esoteric aura that surrounds this approach by showing how BNs have improved ecological risk assessments in the past 20 years, with the goal of encouraging more practitioners to employ BNs and continue evolving the practices of ERA and environmental management. The Guest Editors of the series are Jannicke Moe, John Carriger, and Miriam Glendell.

Guest Editor Jannicke Moe talks to us about advantages of BNs, recent developments, and highlights the research presented in the series—10 articles demonstrating the application of BNs to various environmental assessment and management scenarios involving climate change, ecological and socioeconomic endpoints, machine learning, diagnostic inference, and model evaluation.

Listen on iTunes or YouTube.

Read More »

Do no harm: Evaluating non-lethal fish sampling, with Alyse Kambeitz

31536015808_6132e55663_o (1)
Fish sampling with a seine. Credit: USFWS (Public Domain Mark 1.0).

The goal of any environmental monitoring program is to assess and protect the health of the organisms being monitored. Yet the most common methods require the sacrifice of a large number of individuals to collect enough data to ensure the well-being of the entire population. A new study published in IEAM set out to find a better way to monitor fish populations in Canadian waters affected by mining activity. We spoke with lead author Alyse Kambeitz to hear more. Access the article in the November 2019 issue of IEAM.

Listen on iTunes or YouTube.

Read More »

Get Real! Stuart Hurlbert on Pseudoreplication and Other Sins of Statistical Analysis

zebrafish-replicates
Experimental zebrafish (Danio rerio). Credit: Novartis AG, CC BY-NC-ND 2.0.

Think you know stats? Stuart Hurlbert first described pseudoreplication—a common but serious statistical error—in 1984. Despite widespread knowledge of the error, pseudoreplication is often misinterpreted, and literature surveys show that the error is on the rise in certain fields. Listen to Hurlbert define pseudoreplication and other related errors, plus hear why we shouldn’t dichotomize results as “significant” and “non-significant,” what’s missing from basics stats courses, and what’s next on his list.

Access the Learned Discourse by Hurlbert and Lombardi in the January 2016 issue of IEAM.

Podcast available on iTunes and YouTube

Read More »

Bayesian Networks for the Uninitiated, with David Barton

bayes-theorem2
Bayes’ Theorem. Credit: Daniel Hjort, CC BY-NC-ND 2.0.

Baffled by Bayesian statistics? You’re not alone.

Join us as we speak with Dr. David Barton, Guest Editor of the special series “Bayesian Networks in Environmental and Resource Management,” to discuss the basics of Bayesian approaches in environmental management. The series is composed of seven case study articles, each of which applies the Bayesian network approach to environmental and resource management problems around the world. Access the series in the July 2012 issue of IEAM.

Podcast available on iTunes and YouTube

Read More »

Well Past Time to Stop Using NOELs and LOELs, with Wayne Landis and Peter Chapman

Drs. Landis and Chapman are authors of an editorial in the October 2011 issue of IEAM entitled, “Well Past Time to Stop Using NOELs and LOELs.” The editorial was essentially a call to end the use of these two measures in favor of more statistically robust approaches. Join us as we hear more from Wayne and Peter on their call to move away from relying solely on hypothesis testing.

Podcast available on iTunes and YouTube

Read More »