Many aptitude scales measure general or academic knowledge with a format in which answers are scored as either correct or incorrect. In contrast to traditional scoring procedures, quantifying performance on scales that measure interpersonal knowledge, as well as knowledge of other ambiguous relationships, requires the opinions of multiple experts, and individual responses cannot be easily or unambiguously evaluated. Given this type of uncertain knowledge domain, an obscure and little-appreciated procedure to measure expertise is based on the distance between expert and participant estimates of the relative strengths of a set of probabilistic relationships. Using this format, three scales were developed to measure knowledge associated with leadership, subtle indicators of alcohol abuse, and appropriate dinner behaviors. Data were collected with these scales using U.S. Air Force recruits for whom Armed Services Vocational Aptitude Battery scores were available. Results demonstrate (a) the applicability of the probabilistic response format to efficiently measure individual differences in nontraditional knowledge domains, such as interpersonal skills, and (b) the existence of a seperate first-order factor that may be interpreted as social insight and has a substantial second-order loading on psychometric g.