1. Comparison of Two Different Statistical Methods for Assessing Insecticide Efficacy in Small Plot Trials Targeting Wireworms, Limonius californicus (Mannerheim) (Coleoptera: Elateridae), in Potato
- Author
-
Erik J. Wenninger and Jeffrey Lojewski
- Subjects
0106 biological sciences ,Analysis of covariance ,fungi ,04 agricultural and veterinary sciences ,Plant Science ,Crop rotation ,Biology ,Pesticide ,01 natural sciences ,Treatment efficacy ,Plot (graphics) ,Toxicology ,040103 agronomy & agriculture ,0401 agriculture, forestry, and fisheries ,Analysis of variance ,PEST analysis ,Agronomy and Crop Science ,Limonius californicus ,010606 plant biology & botany - Abstract
The efficacy of insecticides is often assessed in small-plot field trials that compare insect damage and yield loss relative to an untreated check, typically in a randomized complete block (RCB) design. When insect damage is not uniformly distributed across a field, differences among treatments might reflect differences in local pest pressure rather than differences in treatment efficacy. One potential way to account for patchy distributions is to use a “running check” in which each plot contains an insecticide treatment and its own untreated check. Any benefits of this approach must be weighed against the additional labor involved. Wireworms (Coleoptera: Elateridae) are soil-dwelling insects that typically exhibit patchy distributions that can confound insecticide efficacy trials. Several species of wireworms feed on potato tubers, causing damage that makes the tubers unmarketable and, aside from crop rotation, insecticides remain a cornerstone of wireworm management in potato. The current investigation evaluated insecticide efficacy trials targeting wireworms in potato by comparing two different analyses for the same data sets spanning five years and seven different studies. Data collected using the running check approach were analyzed using analysis of covariance and compared with analysis of variance as if the experiment were designed as a RCB. In a majority of cases, both methods resulted in similar levels of statistical power and similar conclusions; however, including data from the running check as a covariate did sometimes eliminate or reveal differences between treatments compared to ANOVA. Overall, given the similarity of the results and the additional labor involved in using running checks, the standard RCB design would have been appropriate for the experiments assessed here. However, for wireworms and other pests that exhibit patchy distributions, care must be taken when designing experiments and interpreting results.
- Published
- 2019
- Full Text
- View/download PDF