Excellent.
Should have spotted the method where you write it as a Taylor series (necessarily with real coefficients, I guess, since the function's restriction to the real line must be real, as stated above) then apply complex conjugation to this series - then it just boils down to showing that the complex conjugate of a complex variable raised to a power is the power of the complex congugate of the variable. In other words
which isn't too hard, I hope!
The Identity Theorem method is a bit more sublte, but it makes sense.
Thanks Hurkyl, g_edgar.