Here is an argument for the process of collapse of our knowledge of entangled quantum processes.
Certain quantum processes are entangled and yet are capable of showing the determinacy of entanglement. For example, one can use an electric circuit to separate ions and electrons [QM objects] on the plates of a capacitor. Then you separate the plates and put them in different places. As long as you do not know the direction of the potential difference of the electric circuit, your knowledge about the charge on the capacitor plates is zero.
After you measure the direction of the current of your experimental setup the quantum mechanical wave function of the un-knowledge collapses and you know, which plate is charged with more electrons and which with more ions.
Applying a voltmeter to the plates you discharge the plates and proof the prediction from the measurement of the current from the experimental setup. If the capacity of the plates is large and your measurement is performed quickly, you can repeat the measurement several times to verify your results.
However, there are processes where the generation of the entanglement and the subsequent measurement of the results is ambiguous. The point of this uncertainty is the claim that the wave functions of the entangled particles remain entangled until measurement.
Take the process of generation of entangled photon pairs. It is possible to design quantum dots, producing pair of photons, entangled in their polarization. The photons of each pair have the directions of polarization with an random angle but entangled by an angle difference. Say the angle for the first photon may be 33°, than the angel of the second photon is 33°+180°.
The point is, that the production of photons with always the same angle of polarization was not possible. And the proof, that an entanglement exist, was done by statistical methods. Not knowing the angles of polarizations of the photons, you should place the polarizers for the measurement in an arbitrary direction and not every measurement will be successful.
Instead of claiming the interference of wave functions until measurement one may claim that the entanglement was finished from the moment of pair production and exists until the disturbance from measurement processes. Both claims are not provable and meaningless for cryptographic and other technical uses.