The adding-and-averaging effect in bundles of information: Preference reversals across joint and separate evaluation.

When does adding mildly favorable information (e.g., job experience at the Railway Credit Union) alongside highly favorable (e.g., job experience at Goldman Sachs) increase versus decrease evaluations of a bundle of information like a resume or product bundle? We posit that whether that package of information is evaluated by itself-in separate evaluation (SE)-or side by side with another package-in joint evaluation (JE)-matters. Across a variety of contexts, four studies show that people "average" in SE and "add" in JE. Consequently, mildly favorable information hurts evaluations in SE but helps in JE. Study 1 demonstrated this "adding-and-averaging effect" among persons with expertise: law professors judging law faculty candidates. Adding middle tier academic publications to a higher tier publication on a CV decreased evaluations of a candidate judged in SE but increased evaluations of the same candidate in JE. Study 3 examined a linear pattern prediction, showing that each piece of mildly favorable information linearly added to the overall impression of a package of information in JE but linearly detracted from evaluations of the identical target in SE. Finally, Study 4 traced these differences in evaluative judgments to a shift in reference points brought about by evaluation mode. Implications for the organization specifically and our understanding of judgment and decision making processes more generally are considered. (PsycINFO Database Record