“…sales nor the advertising history are very successful in reducing the variance in the series. In the literature on sales and advertising, several sources report R-squares of 0.70 and more, which leads one to observe that (1) R-square is not a good measure of association on longitudinal data (see Pierce, 1977) and (2) that one should not expect very good fit with a limited information set such as the one used in this study. Secondly, one can conclude that the exogenous influences and the added noise are about equally important in explaining deviations from the seasonal pattern in sales (V12S,).'…”