More-Or-Less Elicitation (MOLE) 1 Running head: MORE-OR-LESS ELICITATION (MOLE) More-Or-Less Elicitation (MOLE): Improving the Accuracy and Calibration of Estimates

Eliciting numerical estimates from people is a central challenge for psychology. Cognitive biases can affect estimates, making it challenging to measure what people know or believe. We present a novel elicitation method, More-Or-Less Elicitation (MOLE) that, rather than having people make a single, absolute judgment, asks for repeated, relative judgments across the range of possible estimates and constructs a probability density function (PDF) representing their beliefs. Study 1 uses a perceptual estimation task to compare these PDFs with traditionally elicited ranges, concluding that MOLE improves accuracy and precision. Study 2 shows that, compared to alternate, repeated-judgment methods, MOLE takes less time and does not require distractors. Study 3 shows MOLE’s efficacy in eliciting answers to general knowledge questions. We argue that MOLE benefits from using relative rather than absolute questions, avoids problems usually associated with repeated questions, and is a useful method for theoretical and applied researchers needing to elicit uncertain information.