It shouldn't matter. In either case, the true root is the same so the approximation yields the same answer.
Consider f(x) = x - ex2
f(0.6) ~ - 0.098
f(0.7) ~ 0.087
When you check 0.65 you get
f(0.65) ~ - 0.005406
So the true root is between 0.65 and 0.7, hence the true root rounds up 0.7 to one decimal place
Consider f(x) = ex2 - x
f(0.6) ~ 0.098
f(0.7) ~ - 0.087
When you check 0.65 you get
f(0.65) ~ 0.005406
So the true root is between 0.65 and 0.7, hence the true root rounds up 0.7 to one decimal place