Ah, but one could easily design an experiment to eliminate this bias. The truly skeptical should do just that!
Natural selection already has: had purple been of such great benefit to plants, it would be (at least, more) ubiquitous in the plant kingdom.