Lets say there are two entangled Qubits A, B, very far away from each other and described by the state
$|\Psi\rangle = |++\rangle + |--\rangle$
- In a reference system S, A measures + and one hour later B certainly also measures +. 
- But in another system S', moving relative to S, B could measure + first and one hour later A measures +. 
Therefore, it doesn't give sense to state that exactly one of the observers is "responsible" for collapsing the state. However, the scenario is often mentioned in such a way.
I often read that the moment A makes a measurement, the (later) measurement for B is already determined because A collapses the state. Isn't that misleading? Even worse, it is often claimed that B's state changes instantaneously when A makes a measurement.
What is the correct way of looking at this?
 
    