Abstract
Actual realisations of EPR experiments do {\em not} demonstrate non-locality. A model is presented that should enable non-specialists as well as specialists to understand how easy it is to find realistic explanations for the observations. The model also suggests new areas where realistic (``hidden-variable'') models can give valid predictions whilst quantum mechanics fails. It offers straightforward explanations for some anomalies that Aspect was unable to account for, providing perhaps the first experimental evidence that a hidden-variable theory can be {\em superior} to quantum mechanics. The apparent success of quantum mechanics in predicting results is shown to be largely due to the use of unjustifiable and biased analysis of the data. Data that has been discarded because it did not lead to a valid Bell's test may give further evidence that hidden variables exist.