How to use statistics to claim antagonism and synergism from binary mixture experiments
Research output: Contribution to journal › Review › Research › peer-review
Standard
How to use statistics to claim antagonism and synergism from binary mixture experiments. / Ritz, Christian; Streibig, Jens Carl; Kniss, Andrew.
In: Pest Management Science, Vol. 77, No. 9, 2021, p. 3890-3899.Research output: Contribution to journal › Review › Research › peer-review
Harvard
APA
Vancouver
Author
Bibtex
}
RIS
TY - JOUR
T1 - How to use statistics to claim antagonism and synergism from binary mixture experiments
AU - Ritz, Christian
AU - Streibig, Jens Carl
AU - Kniss, Andrew
N1 - CURIS 2021 NEXS 104
PY - 2021
Y1 - 2021
N2 - We review statistical approaches applicable for the analysis of data from binary mixture experiments, which are commonly used in pesticide science for evaluating antagonistic or synergistic effects. Specifically, two different situations are reviewed, one where every pesticide is only available at a single dose level and a mixture simply combines these doses, and one where the pesticides and their mixture are used at increasing doses. The former corresponds to using factorial designs whereas the latter corresponds to fixed-ratio designs. We consider dose addition and independent action as references for lack of antagonistic and synergistic effects. Data from factorial designs should be analyzed using two-way analysis of variance models whereas data from fixed-ratio designs should be analyzed using nonlinear dose-response analysis. In most cases, independent action seems the more natural choice for factorial designs. In contrast, dose addition is more appropriate for fixed-ratio designs although dose addition is not equally compatible with all types of dose-response data. Fixed-ratio designs should be preferred as they allow validation of the assumed dose-response relationship and, consequently, provide much stronger claims about antagonistic and synergistic effects than factorial designs. Finally, it should be noted that, in any case, simple ways of summarizing pesticide mixture effects may come at the price of more or less restrictive modelling assumptions. This article is protected by copyright. All rights reserved.
AB - We review statistical approaches applicable for the analysis of data from binary mixture experiments, which are commonly used in pesticide science for evaluating antagonistic or synergistic effects. Specifically, two different situations are reviewed, one where every pesticide is only available at a single dose level and a mixture simply combines these doses, and one where the pesticides and their mixture are used at increasing doses. The former corresponds to using factorial designs whereas the latter corresponds to fixed-ratio designs. We consider dose addition and independent action as references for lack of antagonistic and synergistic effects. Data from factorial designs should be analyzed using two-way analysis of variance models whereas data from fixed-ratio designs should be analyzed using nonlinear dose-response analysis. In most cases, independent action seems the more natural choice for factorial designs. In contrast, dose addition is more appropriate for fixed-ratio designs although dose addition is not equally compatible with all types of dose-response data. Fixed-ratio designs should be preferred as they allow validation of the assumed dose-response relationship and, consequently, provide much stronger claims about antagonistic and synergistic effects than factorial designs. Finally, it should be noted that, in any case, simple ways of summarizing pesticide mixture effects may come at the price of more or less restrictive modelling assumptions. This article is protected by copyright. All rights reserved.
KW - Faculty of Science
KW - Analysis of variance
KW - Antagonism
KW - Dose addition
KW - Dose-response analysis
KW - Factorial design
KW - Fixed-ratio design
KW - Independent action
KW - Synergism
U2 - 10.1002/ps.6348
DO - 10.1002/ps.6348
M3 - Review
C2 - 33644956
VL - 77
SP - 3890
EP - 3899
JO - Pest Management Science
JF - Pest Management Science
SN - 1526-498X
IS - 9
ER -
ID: 257602028