Environmental Economics

Optimal Experimental Design for Conjoint Analysis

  • Abstract
    This project will determine optimal attribute levels and choice sets for conjoint analysis questions that, given a fixed number of observations, will provide the most information possible about parameter estimators of interest such as mean or median willingness to pay. This research will extend the existing literature on optimal design of conjoint analysis surveys. It will derive optimal designs as opposed to efficient designs.
  • Metadata
    EPA/NSF ID:
    9613045
    Principal Investigators:
    Kanninen, Barbara
    Technical Liaison:
    Research Organization:
    Minnesota, University of
    Funding Agency/Program:
    NSF/Valuation
    Grant Year:
    1996
    Project Period:
    September 1, 1996 to August 31, 1999
    Cost to Funding Agency:
    $82,563
  • Project Reports
    "Optimal Experimental Design for Binary Choice Experiments," June, 1998, Word for Windows, 264KB:


    DRAFT1~1.DOC - DRAFT1~1.DOC

  • Project Status Reports

    To assess the total value, including use and nonuse values, of nonmarket goods such as environmental amenities, researchers often apply survey techniques that allow them to explore public preferences for hypothetical goods or services. The standard survey technique for this purpose has been the contingent valuation (CV) method. Recently, conjoint analysis has been used in several environmental contexts. Conjoint analysis is a marketing technique that can be used to assess values for attributes of market or nonmarket goods based on survey respondents’ willingness to trade-off different bundles of these attributes.

    In a conjoint analysis survey, respondents are presented with a set of scenarios that differ in terms of a series of attributes and are asked to rank the alternative scenarios, or choose their most preferred. The scenarios in the choice set differ by the levels of the different attributes. A major cost consideration in conducting surveys for environmental valuation is the per unit cost of survey administration. At current costs, sample sizes are often limited to the smallest that researchers feel is necessary for a particular problem. By employing optimal survey design techniques, prac- titioners can increase the informational content of each observation, producing the equivalent effect of a larger sample size.


    This project’s goal is to determine optimal attribute levels and choice sets for conjoint analysis questions that, given a fixed number of observations, will provide the most information possible about para- meter estimators of interest such as mean or median willingness to pay.

    This project will extend the existing literature on the optimal design of conjoint analysis surveys in two ways: (1) it will consider attribute levels as well as choice sets as variables in the optimization problem, and (2) it will derive “optimal” designs as opposed to “efficient.” The focus will be on deriving “D-optimal” designs, that is, designs that maximize the determinant of the information matrix.