We examine the computation of the nucleation barrier used in the expre
ssion for false vacuum decay rates in finite-temperature field theory.
By a detailed analysis of the determinantal prefactor, we show that t
he correct bounce solution used in the computation of the nucleation b
arrier should not include loop corrections coming from the scalar fiel
d undergoing decay. Temperature corrections to the bounce appear from
loop contributions from other fields coupled to the scalar field. We c
ompute the nucleation barrier for a model of scalar fields coupled to
fermions, and compare our results to the expression commonly used in t
he literature. We find that, for large enough self-couplings, the incl
usion of scalar loops in the expression of the nucleation barrier lead
s to an underestimate of the decay rate in the neighborhood of the cri
tical temperature.