Poster #4 - Infants Use Gender-Typed Preferences to Chunk Items in Working Memory

• In Event: 2-055 - Poster Session 06
  In Poster Session: PS 06 Section - Attention, Learning, Memory

Integrative Statement

In the absence of perceptual or conceptual chunking cues, infants are unable to keep track of more than three identical objects in working memory (WM) at a given time (Feigenson & Halberda, 2004; 2008). For example, when three identical objects are hidden inside a box, infants continue searching the box when only a subset is retrieved, but do not continue searching when all of the objects have been retrieved. In contrast, infants fail to continue searching the box if four identical objects are hidden and only a subset is retrieved. Previous research has shown that infants can harness their knowledge of the social world to “chunk” items into sets, thereby overcoming typical capacity limits to remember four objects in WM – infants chunk dolls into sets who socially interact in pairs (Stahl & Feigenson, 2014), and dolls who belong to the same linguistic group (Stahl & Feigenson, 2018). The current study examined whether infants can use other social information, namely gender-typed preferences, to group agents into gender-based sets to remember four items in WM.

In Experiment 1, 20- to 26-month-olds (n=15) viewed sets of identical animal dolls that had no identifiable gender. Infants watched as four animals were hidden inside of a box. Prior to hiding, each animal stated a gender-typed toy preference – two animals stated, “I like dolls!” (a stereotypically “girl” preference) and the other two stated, “I like trucks!” (a stereotypically “boy” preference). By 18 months, infants associate such toy preferences with particular genders (Serbin et al., 2001) - here, the question was whether infants would group the animals into sets based on these gender-typed preferences, even though they were all perceptually identical and no other grouping cues were available for chunking.

We compared how long infants searched inside the box when only a subset of the four animals had been retrieved, relative to how long they searched when all animals had been retrieved. Infants searched significantly longer when more objects remained in the box than when it was empty (p<.05). Because infants cannot typically keep track of four identical objects in WM without chunking cues, these results show that infants chunked the animals into sets based on their stated gender-typed preferences.

However, an alternative explanation is that infants can use any distinct preferences, irrespective of whether those preferences are gendered, to chunk agents into sets. In Experiment 2, another group of infants will participate in a nearly identical procedure to that of Experiment 1, except the stated toy preferences will be gender neutral. One pair of animals will state that they like balls, and another pair will state that they like books. If infants again successfully remember all four objects, then they may be able to use social agents’ preferences more generally to form meaningful chunks. However, if infants fail to remember all four objects, then agents’ preferences might need to be linked to a more stable trait (i.e., the agents’ inferred gender based on their preferences) in order to be utilized as a meaningful chunking cue.

Authors

• Aimee E Stahl, The College of New Jersey
  Presenting Author

• Elizabeth Staroselsky, Columbia University
  Non-Presenting Author