ABSTRACT
After early research into the cocktail party phenomenon a number of theories of focused attention were proposed. Broadbent proposed an early filter model. This suggests that there is a selective filter that blocks out some information but allows the required information through to a limited capacity channel. However, this does not explain the cocktail party phenomenon. Treisman believed that the filter was more flexible and that rejected information was not blocked but was attenuated. Since rejected information was available for semantic analysis (in a weakened form) important information could be detected. Deutsch and Deutsch proposed a late selection model in which all information is initially processed and then selection takes place based on the pertinence of information. This explains much of the research evidence but it is an uneconomical model (why analyse all information if only a small proportion is to be used?).
