An experimental design methodology was applied to evaluate the decolourization of crystal violet (CV) dye by electrocoagulation using iron or aluminium electrodes. The effects and interactions of four parameters, initial pH (3-9), current density (6-28 A m(-2)), substrate concentration (50-200 mg L-1) and supporting electrolyte concentration (284-1420 mg L-1 of Na2SO4), were optimized and evaluated. Although the results using iron anodes were better than for aluminium, the effects and interactions of the studied parameters were quite similar. With a confidence level of 95%, initial pH and supporting electrolyte concentration showed limited effects on the removal rate of CV, whereas current density, pollutant concentration and the interaction of both were significant. Reduced models taking into account significant variables and interactions between variables have shown good correlations with the experimental results. Under optimal conditions, almost complete removal of CV and chemical oxygen demand were obtained after electrocoagulation for 5 and 30 min, using iron and aluminium electrodes, respectively. These results indicate that electrocoagulation with iron anodes is a rapid, economical and effective alternative to the complete removal of CV in waters. Evolutions of pH and residual iron or aluminium concentrations in solution are also discussed. (C) 2010 Elsevier B.V. All rights reserved.