UC San Diego

Framing effects are said to occur when equivalent descriptions of objects or events lead to different choices. Attribute frames refer to logically equivalent descriptions along a single dimension. For example, ground beef might be described as "85% lean" or, equivalently, "15% fat". Typically, one frame is positive and one is negative and people evaluate the object more favorably when presented with the positive frame (e.g., "85% lean" beef is viewed more positively than "15% fat" beef). This phenomenon is known as a valence-consistent shift in evaluations. However, missing from the literature on attribute framing is a double frame condition, in which positive and negative frames are presented together. Because some industries (e.g. the FDA) use the double frame as the standard for communicating, understanding how these frames are evaluated has practical implications. Additionally, how double attribute frames are evaluated has theoretical implications: The most common explanation of attribute framing, the Associative Account, predicts how double attribute frames will be evaluated. Such a prediction however, has not been well tested. This dissertation systematically explores the evaluation of double attribute frames, the implications for the Associative Account, and potential alternative explanations for such evaluations. Experiments 1a - 1d examine how double attribute frames are evaluated across stimulus and procedure differences including context, quantitative information, frame descriptors, response question, and response measure. These experiments reveal a stable, consistent, and surprising pattern of results in the evaluation of double attribute frames. Experiments 2a and 2b examine methodological factors common to most attribute frame experiments that may account for the observed pattern of double attribute frame evaluation, possibly by affecting frame saliency. Finally, Experiments 3a and 3b examine whether an alternative explanation for attribute framing, the Information Leakage Account, as well as linguistic cues, can offer insight into the observed pattern of results in the evaluation of double attribute frames. Taken together, this body of work demonstrates a novel, unpredicted pattern of evaluations of attribute frames and raises questions for the standard theoretical explanation of such effects