The Cantabrian brown bear (Ursus arctos) population of northwest Spain has been monitored since 1982. Population trends have been estimated using counts of females with cubs-of-the-year (hereafter, F_CUB). A population viability analysis found a mean annual decrease of 4–5% for 1982–95, but with a stabilizing or even slightly increasing trend in the early 1990s. A recent paper in Ursus concluded that the population was “recovering,” with a 7.5% annual increase, based on the F_CUB index for 1994–2004. We show several factors limit the interpretation of an increasing trend based on the F_CUB data. First, data collection was not systematic, nor were spatial sampling and sampling effort sufficiently accounted for, leading to an arbitrary election of the period to estimate F_CUB trends. Second, data sets did not meet probabilistic analytical requirements. Third, the assumption that the F_CUB trend, albeit positive, directly reflects the population trend was not justified. In addition, we argue that alternative hypotheses explaining F_CUB trends should have been presented, particularly because of the absence of a correlation between population and range increases. Altogether, we call for caution when analyzing data about critically endangered populations like that of brown bears in the Cantabrian Mountains.