Training Reduces Observer Variability in Visual-Based Assessments of Stream Habitat

Abstract

Habitat assessments are an important component of most water quality evaluations in the USA. Habitat assessments are often based on an observer rating the habitat quality of a site by matching specific habitat features to written descriptions. We examined the effect that training has on reducing observer variability by using 2 separate groups of observers to assess the habitat quality at the same sites. One group received training on habitat assessments; the 2nd group conducted the assessment using the written descriptions but without training. The trained group's assessments were less variable than those of the untrained group of observers. We attribute this difference to a greater disparity among untrained observers in the interpretation of individual habitat features. However, after equal training and experience, variability of both groups' habitat assessments was not significantly different at a 2nd stream site. Finally, the variability of the habitat assessments at the 2 sites was not equal, indicating that training at 1 site does not necessarily prepare the observers to evaluate all sites. Thus, training in a range of habitat types is a crucial part of implementing visual-based habitat assessments.