Patuxent Wildlife Research Center

NAAMP III Archive - accepted papers
Home | Archive by Alphabetical Order | Archive by Category

Testing For No Trend

Philip M. Dixon and Joseph H.K. Pechmann

Summary

• Traditional tests of no trend are inappropriate to prove that population size is stationary.
• Equivalence tests define the null hypothesis to be that a trend is large, rather than that there is no trend.
• Two one-sided tests are recommended to test the equivalence null hypothesis.
• The approach is illustrated with data on trends in four amphibian populations.
• This approach requires a priori specification of an equivalence region, within which trends are considered to be small or close to zero. Specifying such a region may be difficult.
• A proposed equivalence region for a trend is a log-regression slope of (-0.0693, 0.0693). This corresponds to a population with a half-life/doubling time of more than 10 years.

[ Read this Paper ]

U.S. Department of the Interior
U.S. Geological Survey
Patuxent Wildlife Research Center
Laurel, MD, USA 20708-4038
http://www.pwrc.usgs.gov/naamp3/naamp3.html
Contact: Sam Droege, email: Sam_Droege@usgs.gov
Last Modified: June 2002