The analysis of the impact of agglomeration economies on firm location is one of the main issues on Regional Economics. However, despite of the broad and solid theoretical developments in this area, empirical findings show a worrying lack of robustness. This paper examines one possible cause: The Modifiable Areal Unit Problem (MAUP). To that end, it is estimated the influence of agglomeration economies on the firm location in Spain for five different divisions of the territory. The results indicate that the estimates are very sensitive to the definition of areas of analysis which cannot be neglected to contrast the hypotheses. Finally, it proposes a methodology to reduce the drawbacks of MAUP and exploit his potential benefits.