Misled by unreliable research data, advertisers are underspending by as much as 30 percent and failing to realize the profits that more ads would generate, according to researchers at the University of California, Davis. Prasad Naik, an assistant professor of marketing, and Chih-Ling Tsai, a professor of management, both at the university's Graduate School of Management, say that with total spending on advertising at $201.6 billion last year, the lost opportunities are significant. In a case study of a leading cereal brand, they found measurements of advertising awareness were off by 40 percent and those for gross rating points, which represent how many people have seen an ad and how often, by 17 percent. "Errors inherent in research surveys lead advertisers to the mistaken conclusion that their ads are less effective than they really are," Naik says. "So advertisers mistakenly spend less and lose out on profits." Naik and Tsai are the first to apply "wavelets denoising" in the field of marketing to correct for errors introduced into research surveys by fallible memories and recording errors, for example. In the same way a blurred photograph of an automobile can be enhanced to provide a license plate number, the mathematical tool filters "noise" or errors from data and reconstructs a truer image of the data's pattern. Naik and Tsai's article "Controlling Measurement Errors in Models of Advertising Competition" will be published in a forthcoming issue of the Journal of Marketing Research.