Nancy Estévez-Pérez

Background. The standardized identification, psychoeducational assessment, and diagnosis of children at risk of reading comprehension (RC) difficulties is a highly specialized, time-consuming, and cumbersome process for teachers, psychologists, and researchers. Following the graph theory framework, text summaries, a ubiquitous RC measure used in schools, can be represented as networks of words (nodes) connected by arcs (TSGraphs). Do their resulting topological properties highlight individual variability in traditional reading/language comprehension measures?

Objective. The objective of this study was to determine whether there is a significant association between individual variability in the connectivity measures of the TSGraphs of selected texts produced using graph theory and individual variability in traditional standardized measures of reading and language comprehension in Cuban school-age children.

Design. Two correlational studies were conducted. Study 1 evaluated the association between the TSGraph properties and the reading comprehension of good and poor fifth-grade readers (N=21). Study 2 evaluated the association between the TSGraph properties and language comprehension in sixth-grade children at risk of intellectual disability (IDr) and typically developing (TD) controls matched in age and gender (N=42). Reading fluency, intellectual capacity, and vocabulary were controlled for in both studies.

Results. Study 1 showed a significant association between TSGraph density and reading comprehension in fifth graders after controlling for reading fluency. Study 2 found that density significantly covaried with language comprehension in sixth graders after controlling for intellectual capacity.

Conclusion. Topological measures of text summaries show promise for the assessment and characterization of reading and language comprehension. This is the first such study conducted on native Spanish speakers. Additional experimental studies in larger samples are required.