The six articles in this special section are based on a NEC-63 (Research Committee on Commodity Promotion) sponsored symposium held in Washington, DC March 29 -30, 1999 entitled "Checkoff Promotion Response Measurement: Improving the Science and Setting Priorities." Over the past decade we have seen substantial advancement in efforts to measure the impacts of generic promotion programs. The symposium was designed to assess where we are in this research, but also motivated, in part, by increasing concerns that published estimates of advertising response may be fragile. If this is true, then policy decisions based on the parameters estimates may not be sound. A related concern is that benefit-cost estimates are often based on models that are incomplete. Omitted elements may include supply response, demand interrelationships, trade, policy setting, marketing channel, opportunity cost, and tax shifting, any one of which could cause the reported benefit-cost ratios to be misleading.
With these concerns in mind, and given the 1996 Congressional mandate that all mandatory commodity promotion programs be evaluated at least once every five years, this seemed like a good time to take stock of progress in the area. Accordingly, we asked George Davis and Bill Tomek to focus on the "big picture" issues of theory and measurement. The program also included several contributed papers that addresses the robustness issue. All the articles have been peer reviewed by two anonymous referees in conformance with the journal's standard procedures. We believe they make a useful contribution to the literature and hope that they will stimulate further improvements in research quality.
Starting from the premise that "there is no such thing as . . . 'the scientific method,'" Davis gives a wide-ranging epistemological discussion of the subject matter before concluding that researchers would do well to focus on "modeling the market, not marketing the model" (to paraphrase). Tomek, in collaboration with Harry Kaiser, urges the adoption of a general-to-specific modeling philosophy as a partial solution to the robustness issue. The modeling philosophy is demonstrated using time series data on milk advertising. In a replication attempt, Coulibaly and Brorsen conclude that the effects of beef advertising are probably too small to measure reliably. Ron Ward, in an invited reply to Coulibaly and Brorsen's work, questions the replication effort, and presents new evidence to suggest that beef advertising impacts are indeed measurable. Chanjin Chung and Kaiser revisit the issue of whether Gross Rating Points may be a better measure of advertising exposure than dollar expenditures and conclude "probably not." Gary Williams focuses on opportunity cost,