The primary roles for population viability analysis (PVA) in bird conservat
ion are to predict the consequences of human actions for populations, and t
o assist in the design of research and monitoring programs. PVAs have limit
ations that compromise their ability to predict the future accurately. But
all predictions are error prone. The lack of completely reliable and verifi
able predictions does not detract from the utility of PVAs because their pr
imary output is to reinforce the rigour and transparency of environmental d
ecision making. The imperative for PVAs is driven by the weaknesses of alte
rnative approaches that rest on undefined and unrepeatable subjective proce
sses. Subjective estimation is known to be biased and inaccurate and this r
eview contends that many alternatives to PVA will produce worse results.