Presentation matters: Number of attributes presented impacts estimated preferences

Elizabeth S. Byrd, Nicole J.Olynk Widmar, Benjamin M. Gramig

Research output: Contribution to journalArticlepeer-review


Best–worst scaling is an increasingly employed methodology in which both the number of attributes shown in each choice task and the number of tasks can vary. Researchers face a tradeoff between the number of attributes shown per question and the total number of questions. U.S. residents (n = 818) were randomly assigned to see one of two best–worst presentations of the same six meat attributes (taste, convenience, safety, animal welfare, price, and nutrition). Significant differences were found in the estimated preference shares when respondents were shown two versus three attributes at a time. Both presentations ranked safety as the most important, taste as the second most important, and convenience as the least important meat purchasing attribute. However, the distributions of most of the preference share estimates were statistically different. Differences in preferences share estimates resulting from the presentation of questions has the potential to influence marketing, retailing, and other decisions. [EconLit citations: C83, M31, Q13].

Original languageEnglish (US)
Pages (from-to)377-389
Number of pages13
Issue number2
StatePublished - Mar 1 2018

ASJC Scopus subject areas

  • Food Science
  • Geography, Planning and Development
  • Animal Science and Zoology
  • Agronomy and Crop Science
  • Economics and Econometrics


Dive into the research topics of 'Presentation matters: Number of attributes presented impacts estimated preferences'. Together they form a unique fingerprint.

Cite this