How does the acquisition of information about a person affect the liking of that person? A recent set of studies suggests that liking decreases as people acquire more information (Norton, Frost, & Ariely, 2007). We test this “less-is-more” hypothesis along with an alternative hypothesis based on information integration theory. According to this alternative, people average available person information in an unbiased manner so that the liking of a person described by a random sample of any number of traits from a trait universe approximates the degree of liking that would be obtained if all trait information were known. The correlation between liking and the number of traits should be zero. We present the results of computer simulation and 2 empirical person-judgment studies. Using Bayesian analyses, we find that the evidence is more consistent with the information-integration hypothesis than with the “less-is-more” hypothesis.