A theoretically driven meta-analysis of studies on symbolic understanding in digital context

• In Event: Poster Session 09
  In Poster Session: PS 09 Topic - Technology, Media & Child Development

Abstract

Young children demonstrate inconsistent use of symbols. For instance, toddlers are more likely to find a toy in a room using information presented in a photograph than in a miniature scale model of that room (DeLoach, 1991, 1995). This may be due to the relative salience of the scale model as an interesting object in its own right, obscuring the symbolic function of the model. In the new digital world, screen media can possess features of both photographs (e.g., two-dimensional) and scale models (e.g., potentially interactive) (Troseth et al. 2019). However, little is known about children’s symbolic understanding of screen media, and symbolic insight could be influenced by many features of both the symbol and referent (Kirkorian et al., 2019).

We aim to understand factors that moderate children’s symbolic representation, particularly using screen media. To this end, we will examine the magnitude of “symbol effects” (i.e., differences in performance when using symbols versus direct, unmediated demonstration) as a function of symbol type (screen, photograph, model) and other task features.

We predict that effect sizes for screen media will be more variable than other symbol types (H1) and have an average effect between photographs and models (H2). Additionally, we predict that screen media effect sizes will increase with symbol-to-referent scale differences (i.e., screen size to room size) (H3) and be smaller for interactive media than video (H4).

A multi-level meta-analysis will be conducted using existing studies on young children’s object retrieval. We will conduct a systematic literature search using databases in multiple disciplines (e.g., PsychINFO, Communication and Mass Media Complete) and relevant keywords (e.g., symbol* OR video* or touchscreen* OR object retrieval). We will reduce the results by scanning for relevance and identifying information needed to conduct the analysis, emailing authors when necessary to gather information. Additionally, we will review literature cited in or by relevant papers and solicit grey literature through listservs (e.g., Cognitive Development Society) and direct outreach to researchers with two or more relevant studies.

An omnibus analysis with a homogeneity test will compare the variance in effect sizes for each symbol type (H1). A moderator analysis will examine the extent to which the effect sizes differ by symbol type (H2), scale ratio (H3), and interactivity (H4).

This study will contribute to the literature on young children’s symbolic understanding. We will test specific predictions in theoretical models of children’s symbolic understanding in the digital world (Troseth et al., 2019), which could in turn inform effective media production to support early learning.

This meta-analysis is related to the first author’s dissertation, which proposed controlled experiments to test theoretical predictions about scaling and interactivity in children’s object retrieval using video. Data collection was disrupted by COVID-19. The authors will conduct this meta-analysis as a proof-of-concept to inform the experimental studies when data collection can resume. The meta-analysis will leverage existing publications and data sets, and the first author has experience conducting meta-analytic studies, lending sufficient feasibility for completing data analyses by March 2021.

Authors

• Mengguo Jing, Boston College
  Presenting Author

• Heather Kirkorian, University of Wisconsin - Madison
  Non-Presenting Author