An approximation method with a stepsize of h = 0.7 produces an error of 0.4. If the approximation method used has an error which is O(h^3), what would you expect the error to be if a stepsize of h = 0.2 had been used instead (give your answer rounded to 6 decimal places)?