That said, at Nature News there's a report about a new experiment that backs up entanglement:
The latest effort to explore the phenomenon, to be published1 in Physical Review Letters on 7 February, uses light emitted by stars around 600 years ago to select which measurements to make in a quantum experiment known as a Bell test. In doing so, they narrow down the point in history when, if they exist, hidden variables could have influenced the experiment.Here' more, with a particularly important line highlighted by me:
“It’s a beautiful experiment,” says Krister Shalm, a quantum physicist at the US National Institute of Standards and Technology (NIST) in Gaithersburg, Maryland. Although few expected it to disprove quantum mechanics, such experiments “keep pushing alternative theories to be more and more contrived and ridiculous”, he says. Similar techniques could, in the future, help to protect against hackers who try to crack quantum-cryptography systems, he adds.
But they left open another loophole — one that is more subtle, and impossible to fully close, says Andrew Friedman, an astronomer at the Massachusetts Institute of Technology in Cambridge, and a co-author on the latest paper. Bell tests also assume that experimenters have free choice over which measurements they perform on each of the pair of photons. But some unknown effect could be influencing both the particles and what tests are performed (either by affecting choice of measurement directly, or more plausibly, by restricting the options that are available), to produce correlations that give the illusion of entanglement.
To narrow this freedom-of-choice loophole, researchers have previously put 144 kilometres between the source of entangled particles and the random-number generator that they use to pick experimental settings5. The distance between them means that if any unknown process influenced both set-ups, it would have to have done so at a point in time before the experiment. But this only rules out any influences in the microseconds before: the latest paper sought to push this time back dramatically, by using light from two distant stars to determine the experimental settings for each photon. “We outsource the choice to the Universe itself,” says Friedman.
The team, led by physicist Anton Zeilinger at the University of Vienna, picked which properties of the entangled photons to observe depending on whether its two telescopes detected incoming light as blue or red. The colour is decided when the light is emitted, and does not change during travel. This means that if some unknown effect, rather than quantum entanglement, explains the correlation, it would have to have been set in motion at least around 600 years ago, because the closest star is 575 light-years (176 parsecs) away, says Friedman, who hopes to eventually push back this limit to billions of years ago by doing the experiment with light from more distant quasars. Their results found a level of correlation that supports ‘action at a distance’1....
Others argue that although, fundamentally, the loophole is never closable, such experiments are valuable because new theories necessarily become more improbable and contrived, or eventually, end up assuming that everything in the Universe was determined at the time of the Big Bang — a philosophical view that most physicists reject. Reworking experiments to reduce and make better assumptions is therefore worthwhile, says Shalm.