Using a simplified psychology of perception and some additional assumptions, a system of twelve factors describing events is presented that together are used as a definition of 'newsworthiness'. Three basic hypotheses are presented: the additivity hypothesis that the more factors an event satisfies, the higher the probability that it becomes news; the complementarity hypothesis that the factors will tend to exclude each other since if one factor is present it is less necessary for the other factors to be present for the event to become news; and the exclusion hypothesis that events that satisfy none or very few factors will not become news. This theory is then tested on the news presented in four different Norwegian newspapers from the Congo and Cuba crises of July 1960 and the Cyprus crisis of March-April 1964, and the data are in the majority of cases found to be consistent with the theory. A dozen additional hypotheses are then deduced from the theory and their social implications are discussed. Finally, some tentative policy impli cations are formulated.