Simple View of Reading in Chinese: A One-stage Meta-analytic Structural Equation Modeling

• In Event: Leveraging Innovative Methods in Meta-Analysis to Advance Developmental Theory

Abstract

Based on the framework of Simple View of Reading (SVR), the current meta-analysis aimed to investigate the relations among important reading skills in Chinese, the most spoken language in the world. We also aimed to explore moderators with theoretical and practical significance (i.e., grade level; location [Mainland vs. Hong Kong]) that may influence the relations among these Chinese reading skills. We hope our meta-analytic attempt can examine the generalizability of SVR to non-alphabetic languages, and more importantly, provide implications for reading instructions and identification of reading problems among Chinese children.

With the one-stage meta-analytic structural equation modeling (MASEM) analysis based on 49,416 individuals from 267 independent samples and 210 studies, we systematically investigated models including meta-linguistic skills, decoding, language comprehension, and reading comprehension for Chinese population. Unlike the traditional correlation meta-analysis that can only study correlations between two variables, with MASEM and extracted correlation matrices from reviewed studies, researchers can create a pooled correlation matrix, based on which inter-relations among more than two variables can be simultaneously modeled and tested. That is, MASEM enables researchers to examine more complex relations among variables (including latent variables), to determine their hypothesized model using model fit indices like SEM, and to directly examine whether a variable moderates a specific path (i.e., explaining between-study heterogeneity with one-stage MASEM). Thus, MASEM is often a preferred meta-analytic method for model testing (involving more than two variables).

Findings showed (Figure 1): 1) decoding and language comprehension were moderately related and together explained 52.7% variance of reading comprehension; 2) meta-linguistic skills made significant direct and unique contributions to decoding and showed a strong relation with language comprehension; however, meta-linguistic skills did not make direct contributions to reading comprehension beyond decoding and language comprehension; 3) location (Mainland vs. Hong Kong) did not emerge as a significant moderator in the model; 4) grade level significantly explained the between-study heterogeneity on the relation between decoding and reading comprehension, such that decoding made more contributions to reading comprehension before grade 2 than after; and 5) the effects of language comprehension on reading comprehension stayed stable with grade, and so did meta-linguistic skills on decoding. These findings, taken together, suggest that the Simple View of Reading (SVR) can be applied to reading in non-alphabetic languages such as Chinese. For Chinese reading development, grade 2 may be the transitional grade where the effects of decoding on reading comprehension started to decrease significantly. The null direct effects of meta-linguistics skills on reading comprehension further support the parsimonious structure of SVR (decoding and language comprehension) in explaining reading comprehension in Chinese.

Authors

• Peng Peng, University of Texas at Austin
  Presenting Author

• Kejin Lee, University of Nebraska - Lincoln
  Non-Presenting Author

• Jie Luo, Beijing Normal University
  Non-Presenting Author

• Shuting Li
  Non-Presenting Author

• R. Malatesha Joshi, Texas A&M University
  Non-Presenting Author

• Sha Tao, Beijing Normal University
  Non-Presenting Author