TY - JOUR
AU1 - Hayes, John P.
AU2 - Steidl, Robert J.
AB - Reed and Blaustein (1995) add an important perspective to the debate over amphibian population trends, asserting that if a null hypothesis of no trend is not rejected and the power of the statistical test to detect the trend was low, then the studyâs results can only be considered inconclusive. Too frequently researchers infer that failure to reject a null hypothesis indicates the null is true, without considering statistical power of these tests and the consequences of committing Type II errors (Peterman 1990). Reed and Blaustein correctly assert that estimating the Type II error rate is essential when evaluating evidence suggesting no decline. Reed and Blaustein attempted to estimate the statistical power of tests using data from studies of eight amphibian populations, none of which demonstrated statistically significant trends with time. Although the authors focused on amphibian populations, they implicitly addressed the more general question of whether it is reasonable to conclude a null hypothesis to be true when it is not rejected (P ); in their analysis, the null hypotheses were that no significant correlations existed between amphibian population sizes and time. Reed and Blaustein suggest that it is only reasonable to conclude the null to be true
TI - Statistical Power Analysis and Amphibian Population Trends
JF - Conservation Biology
DO - 10.1046/j.1523-1739.1997.96034.x
DA - 1997-02-17
UR - https://www.deepdyve.com/lp/wiley/statistical-power-analysis-and-amphibian-population-trends-0Aow97RfVS
SP - 273
VL - 11
IS - 1
DP - DeepDyve