Wouldn't it be easy to find location correlates if this was the case? Strong EMF/RF radiation is rather spatially confined. If it was affecting insects, we should expect negative results in sparsely populated areas.