The assumption that an observed difference is due to chance rather than some other, hypothesized causal factor.