Neonicotinoids and bees: no-use, misuse and abuse of statistics

Speaker: Prof. Jeremy Greenwood (CREEM)

Abstract

The battle over organochlorine pesticides half a century ago taught wildlife conservationists that the agrochemical industry may use strange standards of scientific integrity in arguing the case that its products do not have undesirable side-effects.

Both honey bees and bumble bees may be suffering from the effects of neonicotinoids, which dominate the current market for agricultural pesticides.  A study by a major manufacturer claimed to show that they posed little risk to honey bees but this was based merely on inspection of the experimental results, with no formal scientific analysis.  With Rob Schick, Steve Buckland and I analysed the data from this study, estimating the effect sizes on bees of being exposed to neonic-treated crops.  The confidence limits of the estimates were so wide that it was generally impossible to rule out the possibility that the effects were large and important.

A few other similar experiments have been carried out.  Most have claimed or implied no serious effects of neonics but these claims have all been based on statistical analyses that range from inadequate to completely flawed.