Poster #32 - Attention to Number in Adults and Young Children: A State-Trait Analysis

• In Event: 1-005 - Poster Session 01
  In Poster Session: PS 01 Section - Cognitive Processes

Integrative Statement

The question of how readily people attend to the numerosity of objects in the environment—as opposed to other properties, such as color or size—has recently emerged as an intriguing topic of research in numerical cognition. A number of studies have aimed to identify stable individual differences and their relations to numerical abilities (e.g., Spontaneous Focus on Numerosity, or SFON; Hannula, Lepola, & Lehtinen, 2010). Others, meanwhile, have investigated the degree to which both adults’ and children’s attention to number is sensitive to context. If attention to number is substantially context-sensitive, such attention can potentially be manipulated to yield instructional benefits. The present study seeks to improve our understanding of the relative contributions of contextual factors versus a stable trait in situations where people’s attention may (or may not) be drawn to numerosity.

Using data collected from 146 adults and 191 children (Pre-K–2nd grade), we constructed latent state-trait models (Steyer, 1987) of responses to developmentally appropriate versions of a published experimental measure of "Attention to Number." Participants were presented with a stimulus figure (e.g., three orange triangles) and chose a match from a set of four figures, three of which aligned with the stimulus on one of several featural dimensions (numerosity, shape, pattern, location, etc.; see Figure 1). The full set of 16 items comprised a 2✕2 matrix of conditions (high/low salience of competing features ✕ natural/geometric object shapes). After initial responses to all items were recorded, participants revisited each item and indicated second, third, and/or fourth matches among the remaining choices. For each item, participants received a score from 0–4 that gauged the priority of their attention to number. In addition to assessing the primacy of numerosity versus other features as a basis for matches, we also sought to learn whether “trait-like” attention to number was similarly evident in adults and young children.

Results showed that a single-trait multi-state (STMS) model was a good fit to adults’ data, but a very poor fit for children’s data (see Table 1). Among adults, a measure of occasion-specificity (OSpec) indicated that for items with natural shapes (e.g., apples), differences across task conditions (i.e., “state”) accounted for approximately 40% of the variance explained by the model, with 60% accounted for by a stable trait. For items with geometric shapes, state accounted for 56% of the explained variance. Thus, in our results, contextual differences and a stable trait contributed relatively equally to adults’ attention to number. Meanwhile, although the STMS model was a poor fit to children’s data, a simple state-only model differentiating high- and low-salience conditions exhibited excellent fit, χ2(1) = 0.032, p = .859, RMSEA < .001, CFI = 1.00. These results show that trait-like aspects of attention to number likely do not emerge until sometime after early childhood. Future research should aim to discover when a stable trait contributing to attention to number begins to emerge, as well as how this emergence relates to development, experience with mathematics, and subsequent mathematics performance.

Authors

• Luke Rinne, University of Minnesota
  Presenting Author

• Michele Mazzocco, University of Minnesota
  Non-Presenting Author