Early Writing Development Across Preschool: Examining a Two-Factor Model of Transcription and Composing

• In Event: Advances in Writing Development and Assessment from Preschool to Elementary Grades

Abstract

Purpose: Theoretical and conceptual notions of writing in childhood emphasize that children must successfully leverage multiple, distinct (yet interrelated) writing processes in order to compose text (see Berninger & Swanson, 1994; Juel, Griffith, & Gough, 1986; Kaderavek, Cabell, & Justice, 2009; Kim, 2020; Puranik & Lonigan, 2014). A major limitation in preschool writing research, however, is that studies employ correlational designs that limit our ability to examine growth. Longitudinal research that explores the nature of writing processes across this time period are severely needed. To this end, the primary goal of this study was to propose a two-factor framework (transcription and composing skills) of young children’s early writing based on previous theoretical models and evaluate the reliability of the model with longitudinal data.

Methods: A total of 496 children (approximately 9 per classroom) were recruited from 54 preschool classrooms located in two US states. Children were assessed three times throughout the preschool year on four writing measures: name writing, letter writing, spelling, and story writing. Children were assessed individually in a quiet room at their preschool centers by trained research assistants, with assessments lasting about 10 minutes. Writing samples were scored on a variety of print, orthographic, and composition features.

Results: Confirmatory factor analysis (CFA) was used to compare the model fit indices of a two-factor model (transcription and composing skills) to the one-factor model at three time points (Fall, Winter, Spring). All of children’s early writing skills at different time points are moderately or highly correlated (rs > 0.30-0.77). Children’s early writing skills show an ascending trend across four writing tasks. CFA demonstrates that at all three time points the two-factor model fits statistically better than the one-factor model. The model indices for the one-factor model at fall time is: χ^2(30) = 1613.30, p<.001, CFI=0.554, RMSEA=0.206. The model fit indices improved significantly with the two-factor model, which is χ^2(31) = 1613.30, p<.001, CFI=0.911, RMSEA=0.093. Findings indicated that children’s writing skills in preschool appear adequately captured by attending to transcription and composing processes.

Authors

• Xiao Zhang, Georgia State University
  Presenting Author

• Gary E Bingham, Georgia State University
  Non-Presenting Author

• Ryan Bowles, Michigan State University
  Non-Presenting Author

• Hope K Gerde, Michigan State University
  Non-Presenting Author