This completely depends on the specific error amplifier circuit. In general there is no buffer amp between the output divider and the error amplifier network, in which case the impedance of the output divider will need to be taken into account in the error amp transfer function, so then the above statement is correct. Personally I like to put a buffer amp between the divider and the error amp network, since it allows you to easily vary the K of your loop gain without messing up your error amp. But that's just me.
This doesn't apply to a transconductance amp, since the input of the amp is always high impedance, and there is no external feedback in the error amplifier.