Replication defines reliability – A meta-analysis of aerial insecticide effects on forest arthropods

Aerial insecticide application is considered a last-resort measure to prevent tree mortality and stand dieback from insects and is thus an important tool for maintaining forest ecosystem services, particularly in the face of global warming. However, concerns have been raised about the potential adverse side effects of insecticide applications in forests on arthropods. Further contributing to the ongoing debate are several studies that have reported effects that diverge from predictions based on known insecticide properties and application techniques. Yet the empirical evidence for such unpredictable toxic effects remains inconsistent, presumably due to an inadequate study design.

Here, we provide a systematic meta-analysis of 22 studies on the side effects of aerial applied insecticide, that are currently authorized by German plant protection agencies. In only five studies, we confirmed the employment of true replication; in the remaining 17 studies, we could not exclude the possibility of pseudoreplication. In studies with true replication, we found that, on average, 94 % of the statistically analyzed non-target arthropod study groups responded as predicted. By contrast, only 64 % of responses aligned with these predictable toxic effects in studies lacking true replication. Our findings suggest that the empirical basis of the reported unpredictable toxic effects may not be robust due to inappropriate study design, particularly with respect to replication.

We provide guidance on incorporating true replication into field design, and recommend appropriate data pooling strategies in statistical analysis to avoid temporal and sacrificial pseudoreplication—both essential for generating reliable and interpretable results.