Bell's theorem supposedly demonstrates an irreconcilable conflict between quantum mechanics and local, realistic hidden variable theories. In this paper we show that all experiments that aim to prove Bell's theorem do not actually achieve this goal. Our conclusions are based on a straightforward statistical analysis of the outcomes of these experiments. The key tool in our study is probability theory and, in particular, the concept of sample space for the dichotomic random variables that quantifies the outcomes of such experiments. We also show that an experimental proof of Bell's theorem is not, in principle, impossible, but it would require a completely different experimental apparatus than those commonly used to allegedly achieve this objective. The main consequence of our work is that we cannot dismiss local realistic hidden variable theories on the basis of currently available experimental data.