On Measuring Bird Habitat: Influence of Observer Variability and Sample Size Academic Article uri icon

abstract

  • Studied the effects of observer variability when estimating vegetation characteristics at 75 0.04-ha bird plots. Observer estimates were significantly different for 31 of 49 variables. Multivariate analyses showed significant interobserver differences for 5 of the 7 classes of variables studied. Variable classes included the height, number, and diameter of trees, height to the 1st live tree branch, height and number of shrubs, and composite variables. Observer estimates were then compared with measurements of the same habitat variables. There was no clear pattern because estimates by each observer tended to deviate unpredictably from different measured values for 21 variables. Sample size requirements for selected variables ranged from 20-50 for measurements and from 20 to >75 for estimates. Studies that rely on ocular estimates might sacrifice accuracy in lieu of potential time and cost savings. -from Authors

published proceedings

  • The Condor

author list (cited authors)

  • Block, W. M., With, K. A., & Morrison, M. L.

citation count

  • 33

complete list of authors

  • Block, William M||With, Kimberly A||Morrison, Michael L

publication date

  • May 1987