Otherwise we're in a "no true Scotsman" situation because no one, not even you, can never truly know what your preferences are.
Otherwise we're in a "no true Scotsman" situation because no one, not even you, can never truly know what your preferences are.
You are asking people to reduce their highly complex internal evaluation process into a very small number of variables that you ask them to enter such as height preference.
The whole point of Gale Shapley (imo) is that it bypasses the need to reduce people's preferences down to single variables such as height. The idea is that you just give people a list and let them rank the list using whatever internal hidden black box process they want, the algorithm just operates on the output ranked list.
Trying to reduce people's internal evaluation process to a small number of variables (such as height) is where you run into problems, of the kind that you mentioned.