[xnull6guest]

> Are there any experts here who can give a reason why error correction in quantum computing will scale polynomially or better?

Because it takes a finite number of Toffoli/Hadamard/whatever gates to error correct local errors.

That is to say it's like check (or ECC) bits in traditional RAM. It's a couple extra transistors per memory location. Error correcting codes in general tend to be very lightweight and can be made to work locally - the the quantum world is the same.