RESUMEN
PURPOSE: Reading and speech difficulties are common in childhood, yet it is not fully understood how much of their etiology is shared. This partly derives from methodological issues related to overlooking the potential co-occurrence between the two sets of difficulties. This study investigated the effects of five bioenvironmental predictors in a sample assessed for such co-occurrence. METHOD: A combination of exploratory and confirmatory analyses was performed on longitudinal data from the National Child Development Study. Exploratory latent class analysis was performed on children's reading, speech, and language outcomes at ages 7 and 11 years. Membership in the obtained classes was modeled using a regression with sex and four early-life predictors: gestation period, socioeconomic status, maternal education, and the home reading environment. RESULTS: The model yielded four latent classes that broadly reflected (1) average reading and speech, (2) excellent reading, (3) reading difficulties, and (4) speech difficulties. Early-life factors significantly predicted class membership. Male sex and preterm birth emerged as risk factors for both reading and speech difficulties. Protective effects against reading difficulties were identified for maternal education, and lower (but not higher) levels of socioeconomic status and the home reading environment. CONCLUSIONS: Co-occurrence of reading and speech difficulties in the sample was low, and differential patterns of effect of the social environment were supported. Reading outcomes were under stronger malleable influence than speech outcomes.