The study reexamines Phillips's findings regarding the linkage between the appearance of news stories on prominent suicides and the subsequent monthly rise of national suicides-the Werther effect. Extending Phillips's original data set to 1977, and employing the quasi-experimental method, it is found that stories on prominent suicides are likely to trigger a subsequent rise in national suicides. However, this rise may be related to the linkage of suicide with the business cycle, and the fact that more prominent suicides may occur in years when there is a downturn in the economy. This study analyzes suicide rate data with a multivariate time-series model and controls for seasonal effects, the average duration of unemployment and war. No significant linkage is found between the national suicide rate and stories on prominent suicides on the front page of the New York Times, Employing Boorstin's definition of celebrities, the prominent suicides on the front page of the New York Times between 1947 and 1977 are differentiated as celebrity and noncelebrity suicides. Examining only celebrity suicides, it is found that a significant rise in the national suicide rate occurs in the month after a celebrity commits suicide. Suicidal imitation by the public is more selective than originally hypothesized by Phillips. [ABSTRACT FROM AUTHOR]