Barbie-Cueing Weight Perception

It was previously reported that Barbie feels heavier than Ken when both dolls are matched for mass. However, we felt it was unclear from this earlier report if the effects went beyond a typical size-weight illusion. By providing better controls, we conclude more confidently that doll features other than size influence weight perception. Specifically, conceptual knowledge, in the form of culturally reinforced biases, seems to affect how we perceive their weight.

smaller object to be lighter and therefore apply less force than is required to lift it. Additional force is needed, causing the object to feel heavier. Conversely, the larger object is lifted with excessive force, and applying corrective forces to stabilise it in the air causes it to feel lighter.
Because Dijker (2008) was interested in the influence of culturally reinforced biases, he created a cohort of uncanny dolls-each with the same mass but with an exaggerated quality. Specifically, the dolls emphasised cuteness, youth, old age, masculinity, femininity, different races, or some combination of these. Participants rated the dolls' expected (pre-lift) weight and perceived (post-lift) weight. As hypothesised, culturally reinforced biases seemed to affect both measurements. Participants expected dolls emulating strength and masculinity to weigh more, yet perceived them to weigh less. Participants in a control experiment lifted a series of cans that varied in volume but weighed the same as the dolls. Participants reported an SWI; smaller cans felt heavier. Still, this illusion seemed weaker than that found with the dolls. Taken together, Dijker concluded that culturally reinforced biases influenced the dolls' perceived weight beyond a simple SWI.
However, we felt that there were some shortcomings with the design. Dijker (2008) did not report the volume of the dolls nor did he match the cans' volumes to the dolls. In our view, it is imperative that control objects match the dolls in volume as well as weight to confidently conclude that the dolls elicit a stronger weight illusion than a typical SWI. Thus, Experiment 1 aimed to do precisely that, using 31 right-handed adults from the La Trobe University community (16 females, 15 males; M age ¼ 23.84 years, SD age ¼ 4.68). All participants in this study gave written informed consent to the procedures, which were approved by the University's ethics committee.
Stimuli were two sets of objects (i.e., dolls and control objects), matched in weight, volume, height, and colour (Figure 1 Following these treatments, we gave the dolls a bath to determine the amount of water displaced after submersion (Figure 1(e)). These values were taken as their volumes. Using this information, we created the control objects-3D-printed cylinders-that matched the dolls in both volume (Ken: 328.25 cm 3 , Barbie: 246.29 cm 3 ) and height (Ken: 31.2 cm, Barbie: 26.5 cm). The cylinders were painted a similar colour to the dolls' skin, because colour can influence perceived weight (Walker, Francis, & Walker, 2010). Finally, we inserted lead pellets in the centre of the cylinders, held in place by foam, so that they also weighed 122 g.
In Experiment 1, participants closed their eyes and one stimulus pair (dolls or cylinders) was placed on the table in front of them. Participants then opened their eyes, hefted one stimulus at a time, and provided heaviness ratings using an absolute magnitude estimation procedure described elsewhere (Buckingham & Goodale, 2013). They used a different hand for each stimulus and then lifted each of them with the opposite hands. This procedure was repeated for the second stimulus pair. The order in which the pairs were presented, and the starting hand used to heft the stimuli, was counterbalanced across participants.
Although the pairs were matched in physical volume, the greater perceived weight difference for Barbie and Ken could potentially be explained by a greater difference in perceived volume for the dolls, given that object shape and structure can influence perceived volume (Raghubir & Krishna, 1999). Thus, in Experiment 2, 16 new righthanders (8 females, 8 males, M age ¼ 21.56 years, SD age ¼ 1.27) indicated the perceived volume of each doll/cylinder by pouring a representative amount of water from a large, transparent container (capacity: 1.8 L; approximately 8.5 cm Â 8.5 cm Â 26 cm) into a smaller, transparent container (capacity: 0.9 L; approximately 9 cm Â 9 cm Â 12 cm). They were asked to pour the amount of liquid they felt would fill each stimulus if it were hollow. Stimuli were presented as pairs (with appropriate counterbalancing as in Experiment 1) and participants performed two trials for each pair. They did not touch the objects. We drilled holes in Ken to reduce his weight (c) and added lead to Barbie to increase hers (d). Ken and Barbie's volumes were determined by method of water displacement (e). The dolls did not experience pain or drowning and tolerated their treatments well given their inanimate disposition.
These results are informative. We can now say with more confidence that dolls can influence weight perception beyond a simple SWI, and that this difference is not due to either physical or perceived differences in volume. Conceptual knowledge, in the form of culturally reinforced biases, seems to affect how we perceive their weight. This has important theoretical implications as it suggests that weight perception can be influenced by a top-down mechanism. Whether or not the features represented by dolls influence perception via variations in lifting behaviour remains under debate (see Dijker, 2008, for his original proposals). The findings also demonstrate how cultural biases can permeate even basic perceptual processing, including our conscious experience of the weight of objects around us.

Declaration of Conflicting Interests
The author(s) declared no potential conflicts of interest with respect to the research, authorship, and/or publication of this article. * * Figure 2. Mean standardised heaviness ratings (a; Experiment 1) and perceived volume as measured with the water task (b; Experiment 2) with standard errors for the small (blue) and large (yellow) dolls and cylinders. Asterisks (*) denote significant differences after correcting for multiple comparisons (p < .05).