Selective attention and secondary message analysis: A reconsideration of Broadbent's filter model of selective attention

If the type of channel sharing which Broadbent (1958) hypothesized for low information messages also occurs with high information verbal messages, then his filter model is an adequate model of the selective attention process. Analysis which proceeds by a hierarchy of tests and the fluctuating redundancy of primary prose messages can together account for the characteristics of secondary message detection. This hypothesis was tested using four levels of redundancy in the primary message, and two types of secondary message targets, requiring analysis to two different levels in the hierarchy of tests for detection. It was predicted that there should be more target detection as redundancy increased, and more detection of the targets requiring the lower level of analysis than of those requiring the higher. The predictions were confirmed with the exception that the higher level targets were detected with equal frequency at all levels of redundancy. A possible reason for this is suggested.