TitleSample size weighting in probabilistic inference
NameLindemann, Natalie Ann (author), Gelman, Rochel (chair), Chapman, Gretchen B. (co-chair), Singh, Manish (internal member), Rips, Lance (outside member), Rutgers University, Graduate School - New Brunswick,
DescriptionHow do people evaluate data on the basis of sample size? Normatively, sample size is an important factor that one should consider when making judgments and inferences from sample data. Previous research is mixed regarding whether or not laypeople are sensitive to sample size. However, in this paper I show that laypeople attend to sample size, but that their sensitivity decreases as sample sizes become larger. This curvilinear functional form is found for both high and low numerate subjects across two different judgment tasks. However, high numerate subjects consistently show greater sensitivity to sample sizes than lower numerates, although they still underweight sample size relative to normative standards. Low numerate subjects’ sensitivity to sample size may be increased by providing raw data and instructions that sample size matters.
NoteIncludes bibliographical references
Noteby Natalie Ann Lindemann
CollectionGraduate School - New Brunswick Electronic Theses and Dissertations
Organization NameRutgers, The State University of New Jersey
RightsThe author owns the copyright to this work.