Yes, you are overlooking something very simple: the circle we're asked to find doesn't (necessarily) have a radius of sqrt(13/4).
The circle with "diametrically opposite end pts. (3,0) and (1,-3)" has a radius of sqrt(13/4). Its equation is easy to find, since you know two points on it. AKG did it in his first post. This circle is not a tangent to 2x + 3y + 1 = 0.
What we're looking for is a totally different circle (that IS a tangent to 2x + 3y + 1 = 0). The only connection this new circle has with the old circle, is that they are
orthogonal to each other.
*edit* That's not to say that you can't find a circle with a radius of sqrt(13/4) that is tangent to 2x + 3y + 1 = 0 at (1, -1). Let the circle's equation be (x - a)^2 + (y - b)^2 = 13/4. Work out its intersection with the tangent line. The discriminant must be zero. You now have an equation relating a and b. But since you also know that (1, -1) must lie on the circle, you have another equation with a and b. Two equations, two variables, should be solvable. In fact, if you slog through the calculations (or enlist a computer to do it for you ;)), you'll find that the equation is x^2 + (y + 5/2)^2 = 13/4 or (x - 2)^2 + (y - 1/2)^2 = 13/4.
You can apply the same technique (letting the circle's equation be (x - a)^2 + (y - b)^2 = r^2, though) to solve the original problem, only you have to apply the condition given on Mathworld... Very un-pretty though.
And now I see the mistake you did. If you have this equation:
-2/3 = -\frac{(1 - a)}{(1 - b)}
1 - a is not necessarily equal to 2, and 1 - b is not necessarily equal to 3. Since \frac{2}{3} = \frac{4}{6}, is 2 = 4 and 3 = 6?