Extensive experimental tests of the Bell inequality have been conducted over time and the test results are viewed as a testimony to quantum mechanics. In considering the close tie between quantum mechanics and statistical theory, this paper identifies the mistake in previous statistical explanation and uses an elegant statistical approach to derive general formulas for two-particle Bell tests, without invoking any wavefunctions. The results show that, for the special case where the spins/polarizations are in the same, opposite, or perpendicular directions, the general formulas derived in this paper convert to quantum predictions, which are confirmed by numerous experiments. The paper also investigates the linkages between the statistical and quantum predictions and finds that vector decomposition and probability law are at the heart of both approaches. Based on this finding, the paper explains statistically why the local hidden variable theory fails the Bell tests. The paper has important implications for quantum computing, quantum theory in general, and the role of randomism and realism in physics.