Abstract
Most adults, especially women, hold infants and objects representing infants, such as dolls, preferentially on the left side. The attention hypothesis credits the effect to left-directed attention for perception of emotionally salient targets, faces being prime examples. Support comes from studies showing stronger left visual hemispace (LVH) biases in left-holders than right-holders on the Chimeric Faces Test (CFT), but control tests with non-social/emotional objects are needed. We therefore observed young women holding a doll, a book, and a bag, and compared their scores with their performance on the CFT. We also assessed their handedness to check on its possible role. Overall, only the doll elicited a significant side bias, with 57% of all holds on the left, 2% in the middle, and 41% on the right. On the CFT, only left-holders had an LVH bias, whereas right-holders had no bias in either direction. Only the doll-hold scores were consistently related to CFT scores, and for none of the objects was handedness related to side-of-hold.
Acknowledgements
Some of this work was presented at the Eighteenth Annual Conference of the Human Behavior & Evolution Society, Philadelphia, Pennsylvania, 7–11 June 2006. We thank Michael Corballis and the two anonymous reviewers for their helpful comments on an earlier version of this report.
Notes
1Although all participants completed both sessions, on some of the tests of holding, after scores were entered into a database, computer error led to some records being corrupted, so that the numbers of participants represented in this and other analyses varied slightly, from 53 to 66, depending on the task and the session. So far as we can tell, the missing scores represented a random component of the sample and had nothing to do with what participants did or failed to do (e.g., failing to perform certain tasks or to answer certain questions). We therefore do not believe that they are likely to be important for data interpretation. To be certain, we compared the scores from the missing-data and complete-data participants on all of the relevant variables, such as CFT and handedness. There were no significant differences between them.