CUMULATIVE ADVERTISING EFFECTS: THE ROLE OF SERIAL CORRELATION; A REPLY

In a recent paper, Houston and Weiss applied a non-linear estimation procedure to a serial correlation model of advertising effect. Although the methodology is interesting, their conclusions, based on the annual Lydia Pinkham data, are in error. This paper locates the source of the error and corrects the conclusions. It also applies the Houston and Weiss methodology to the monthly Lydia Pinkham data with interesting results.