This paper examines the reactions of investors to the arrival of unexpected information in five major US equity markets from 1990 to 2001, a period characterized by high daily trading volume and the increasing presence of noise-traders. Market surprises are identified using a strictly quantitative approach, cumulative abnormal returns are calculated and tracked for a period of 30 days after each favourable or unfavourable event. The empirical results provide evidence that investors' reactions during the sample period are consistent with the prediction of the Uncertain Information Hypothesis in all markets except NASDAQ. One major implication of these results for investors is that implementing a contrarian strategy of buying current losers and selling current winners will not generate superior returns.