A relatively simple procedure for modeling periodic components in time series data is presented in this paper, along with an example of the procedure's use with communication data. Similar to multiple regression analysis, the described procedure has four steps that are based on information about periodic waves and their components, how to create models of wave components by using least squares estimations of amplitude and phase, the estimation of frequency, an algorithm for computing estimates in multiple component models, significance testing with these estimates, and harmonic analysis. The paper reports on the use of the procedure on data concerning news stories about the Concorde supersonic transport in the Washington, D. C., mass media, noting that the procedure's adjustments to hypotheses about intervals between Concorde stories helped predict cycles in the data that accounted for large amounts of variance. The paper concludes that as communication data becomes increasingly time based, procedures such as the one described should become more prominent. (RL)