Parent ability to diagnose the cause of non-contingent crying in an infant was investigated through use of a new methodological instrument. Problems programmed on a microcomputer presented 25 information units leading to only one correct causal hypothesis about infant crying and 25 information units similarly structured about an adult woman's insomnia. Subjects were 30 college-educated, upper-middle class women divided into four groups: inexperienced in infant caregiving, primiparae, multiparae, and (nulliparous) pediatric nurses. All subjects were instructed to select the fewest and only the most important information units in order to determine which of nine causal hypotheses was correct. Results suggest that women with some infant caregiving experience are, on average, more efficient and accurate in diagnosing the cause of crying in babies than women who have not had that experience. The three experienced groups (primiparous, multiparous, and nurses) only differed reliably from the nulliparous group, suggesting a novice-expert dichotomy. Nevertheless, the nulliparae performed surprisingly well, suggesting a cultural effect involving the provision of experiences to women to aid them in their future task of mothering. It is concluded that the use of computer-presented problems offers a promising new approach for addressing cognitive and cognitive-behavioral questions. (RH)