Both restriction to a specific population group or matching to a certain group will minimize the effect of potential confounding variable. But it will restrict the sample size to a specific group and so the sample size has to be enlarged enough to have the optimum power of the study which means more expenses and more time will be needed.