Some statistical modeling problems in the advertising industry: A look at media exposure distributions

Research output: Contribution to journalArticleResearchpeer-review

14 Citations (Scopus)


Advertisers have been estimating the potential audience for their advertising campaigns for over 30 years, with varying degrees of success. When abstracted into random variables, their modeling problems become rather tantalizing for statisticians due to awkward inter- and intramedia correlations. The purpose of this article is to introduce more statisticians to fertile areas of research in the advertising industry.

Original languageEnglish
Pages (from-to)254-260
Number of pages7
JournalThe American Statistician
Issue number4
Publication statusPublished - 1 Jan 1992
Externally publishedYes


  • Beta-binomial
  • Canonical expansion
  • Dirichlet-multinomial
  • Duplication of viewing law
  • Log-linear model

Cite this