&NA;Recent family, twin and adoption studies of psychiatric illness have often used “super‐normal” control groups, from which individuals are excluded if they demonstrate any significant psychopathology. Relatives of such “super‐normal” control groups will have lower rates of psychiatric illness than the general population. Since such screening is not usually applied to index probands, rates of “screened‐out” disorders may be increased in relatives of index versus “super‐normal” control probands, producing spurious evidence for coaggregation. An algebraic model demonstrates that this artifactual coaggregation will be strongest for “screened‐out” disorders that are both highly familial and common in the population.