Why is it faster (as in proportion to volume) to boil 4 cups of water than to boil 2 cups?

Double the ammount of water does not need doulbe the ammount of time to heat, since while the energy needed is doubled indeed, losses due to vaporization and radiation from the kettle should be approximately constant.

You can plot the time needed for a given ammount of water to boil and try to fit a function into that. With two data points you can manage to fit a straight line, corresponding to linear growth, although I do not expect that to be a good fit. Try doing measurements with 1, and 3 cups, too. Then you have more data and see what kind of function fits the data best. That way you can extrapolate to higher ammounts.