A recurring error in evaluating the effects of different pesticides, pollutants and fertilizers with a zero level

Abstract

Abstract Background The Quenouille-Addelman solution has been proposed to properly analyze linear models with a crossed or factorial arrangement of treatments that includes a qualitative/categorical and a quantitative factor with a zero level, a situation particularly prevalent in ecotoxicological studies. However, a review of the recent literature reveals that this solution isn’t used, perhaps due to a lack of recognition that zero-level factors can produce incomplete factorial arrangements. Results Using practical examples, I demonstrate that the conclusions of a study can be substantially altered if the Quenouille-Addelman solution is not used when warranted. Conclusions Suspecting that the lack of a detailed method may have contributed to the underutilization of the solution, I describe how to apply the solution using current statistical software packages and discuss how the solution can be adapted to address some experimental situations not previously considered.