Poster #38 - How Picture Books Support Word Learning: A Human Simulation Paradigm Study

• In Event: Poster Session 12

Abstract

Picture books have been shown repeatedly to be an important tool to support early language learning (Dickinson et al., 2012; Mol et al., 2008). Studies investigating the mechanisms by which they support learning have largely focused on distinct features of picture books, with some research focusing on linguistic information (Montag et al., 2015), others on social dynamics (Ninio & Bruner, 1978), and yet others on visual properties (Flack & Horst, 2018). How the various properties independently and jointly support learning however remains unknown and is a topic of great interest. The Human Simulation Paradigm (HSP) is one well-established method for characterizing and quantifying how different sources of information support one facet of language learning: identifying the meanings of words (Gillette et al., 1999). We adapted the HSP to examine and compare the informativity of the linguistic and visual worlds of picture books for word meaning identification. Of particular interest was the relative informativity of the linguistic and visual signal in children’s picture books for the identification of early-learned nouns (e.g., bird, chair, duck), whether this informativity differs as a function of noun type (i.e., concrete vs. abstract), and how the informativity changes as a function of the age of the picture book’s target audience.

Methods:
Adult participants (N = 64) were randomly assigned to the Illustration or Text conditions. In both, participants completed 40 trials in which they guessed the identity of a “mystery noun” present in the original text. Participants either saw the illustration in which the mystery noun appeared (Illustration condition; Fig 1B) or the text surrounding that mystery noun (Text condition; Fig 1C). Half of the nouns were basic-level object categories (BLOCs – see Kako, 2005); the other half were non-basic-level object categories (non-BLOCs). Participants either saw stimuli from picture books for 0-to-5-year-olds; the other half saw stimuli from picture books for 3-to-8-year-olds.

Results and Discussion:
Our study revealed three findings on the sources of word meaning in picture books. First, participants in the Text condition (M = 0.35, SD = 0.23) outperformed those in the Illustration condition (M = 0.15, SD = 0.22), t(39) = 4.58, p < 0.001, highlighting the value of linguistic information in picture books (Fig 2A; see also Cameron-Faulkner & Noble, 2013; Montag et al., 2015). Second, consistent with studies on the sources of information about word meaning in child-directed speech (Kako, 2005), the text advantage was unique to the less imageable, more abstract non-BLOC nouns (Mtext = 0.38, SDtext = 0.23; Millustration = 0.04, SDillustration = 0.08; Fig 2B). Finally, these trends on the role of illustrations and text in highlighting noun meaning across word types appeared the same regardless of the target audience (see Fig 2C-D).

Picture books are widely regarded as an invaluable tool for learning (Hudson et al., 2017; Wagner, 2017). Understanding how different features of picture books support word learning advances our knowledge of how picture books support learning. These data may ultimately help educators and practitioners in deploying picture books for intervention purposes.

Authors

• Amanda Mankovich, University of Connecticut
  Presenting Author

• Sumarga H. Suanda, University of Connecticut