We show how to achieve lattice-spacing-independent results in numerical sim
ulations of finite-temperature stochastic scalar field theories. We general
ize a previous approach by obtaining results which are independent of the r
enormalization scale. As an application of our method, we examine thermal p
hase mixing in the context of Ginzburg-Landau models with short-range inter
actions. In particular, we obtain the lattice-spacing and renormalization-s
cale-independent critical value of the control parameter which determines t
he free-energy barrier between the two low-temperature phases. We also prop
ose a simple procedure to extract the critical value of control parameters
for different choices of lattice spacing.