Comparative studies in visual ecology of birds often rely on several assumptions on the evolution of avian vision. In this study, we show that when these assumptions are not upheld, conclusions may be strongly affected. To illustrate this purpose, we reanalysed the data of Avilés & Soler (J. Evol. Biol.22: 376-386, 2009) who demonstrated that nestling gape colouration in altricial birds is associated with visual system. We show that a slight change in analysis methodology leads to opposite conclusions. Such conflicting result raises the problem of applying powerful methods developed for continuous variables to a small sample and a small number of independent events of qualitative visual system shift in comparative analyses. Further, we show that the current trend to assume strong phylogenetic inertia of avian visual systems is contradicted by data and that the sequencing of the SWS1 opsin gene should be considered as an alternative approach.