Adults' performance on horizontality tasks: Conflicting frames of reference.

Many adults fail the Piagetian water-level task designed to assess children's abilities to use coordinate axes for representing physical phenomena. Although McGillicuddy-De Lisi, De Lisi, and Youniss (1978) reported that adults' difficulties virtually disappear with a crossbar horizontality task, this advantage might reflect the absence of an embedding nonoblique context. Consistent with this hypothesis, college students in Study 1 performed no better on an embedded crossbar task than on the water-level task. Performance was significantly better on a disembedded crossbar task. The predicted superiority of field independence was not found. In Study 2, college students' performance on the embedded crossbar and water-level tasks again did not differ significantly. Performance on disembedded crossbar tasks was better than on the water-level task, regardless of whether or not the symmetrical, pivoted nature of the crossbar was emphasized