“…The majority of research has required participants to estimate the strength of the relationship between two dichotomous variables. In contrast, the few studies that have examined numerical stimuli (i.e., perceived covariation between continuous or 591 ordinal variables) have reported relatively accurate perceptions of covariation (e.g., Alloy & Tabachnik, 1984;Baumgartner, 1995;Bettman, Creyer, John, & Scott, 1988;Lane, Anderson, & Kellam, 1985;Malmi, 1986;Wainer & Thissen, 1979;Well, Boyce, Morris, Shinjo, & Chumbley, 1988;Wright & Murphy, 1984). The experiments that follow were restricted to the case of numerical stimuli.…”