Can you force an electron spin to be up or down so that when you measure it, you know what you will expect? I imagine that with an entangled electron pair, forcing the spin on one would enable faster than light communication because you would know the other's state. I imagine if you have a number of entangled electrons, if it was possible to filter the pool of electrons such that only known states will pass the filter when you measure them, then it would be possible to send a complex message. Even if you can force the spin with greater than random chance (noise), it might still be useful to encode a message and decipher it on the other end.
It is possible to know the spin state of an electron leaving a sample. In this way one can "prepare" electrons in a given state. That is how polarization experiments are done.
If a message is encoded in an electron beam, it will not arrive faster than light. The light velocity is the limit at which electrons might travel.
It is not possible to transmit a message, faster than light, even using entanglement,even if one measures the spin polarization of each individual electron. Suppose for simplicity that each electron is entangled with the parent atom, the entanglement is fixed at the time of departure. One cannot change the parent atom's spin state and transmit the effect to the electron that has already left. One will just destroy the entanglement. One cannot "write a message" after departure of the beam.
The assumption, that the states of two entangled by their spin electrons are indifferent - hence a superposition - is right for our knowledge about this states and mathematically correct. I'm not so familiar with pair production of electrons, but let's take the pair production of photons in a spontaneous parametric down-conversion. The two photons are entangled by their spin always in the same manner. Until nobody measure one of the photons state (the spin) our knowledge about the concrete distribution of two states to the two photons is zero. What we know is, that both states are exist and we get a confirmation about this at any time, after we measured both particles. Measuring one particle, we predict the state of the second and get the right result for the second sate by measurement in a statistical satisfying result. Wy statistical satisfying? Because the environment very fast destroys the state. The real art of the quantum cryptography is it, to prevent the state of entangled particles from destruction.
Long story, short conclusion: After we measure the state of one of the two entangled particles the experience value about this process from previous measurements of both particles let us make the prediction about the state of the second particle. But only the measurement of the second state gives us certainty. The tricky thing is, that statistically the experience value is smaller than 1, means, that not all measurements give the right correlations (see the environment influence) but we are legitimate due to understanding the process with its conservation of state variables or due to our experience from past experiments to expect the entanglement for artfully prepared spontaneous parametric down-conversion setups. And now, please, following the causality chain, at which point this this interpretation of entanglement is less truth or more wrong than the collapse of a probability function of two entangled particles? In both cases we have ingoing parameters and resulting parameters of the involved particles. The time between the start of the process and the measurement of the states is hidden and any interpretation has to be wellcome as long as it predicts the right results.
But - the faster than light communication will not appear at all as a question in the explanation I give. And this is the important point, that after an interpretation of a process we have or we have not new questions.