This study adds to the existing body of research by (a) including a comprehensive analysis of published reviews and primary studies investigating the effects of early identification of hearing loss and intervention on language outcomes; and (b) using advanced statistical techniques to further examine existing data on nearly 5,200 children with hearing loss. Analysis of reviews of primary studies showed these reviews exhibited severe sampling bias, lacked systematic methods for analyzing

more »

... and did not include a common metric for comparing results across studies nor a mechanism for analyzing how findings from primary studies covaried with other important factors such as parental involvement, fidelity of intervention, or study quality. Without a more rigorous analysis of primary studies, the conclusions drawn from these reviews are questionable. The review of primary studies revealed many methodological problems including weak experimental designs, small sample sizes, attrition or questionable sample selection methods, differences in length of treatment and characteristics of the participants, and inadequate reporting. Many researchers unjustifiably concluded that earlier intervention produced better developmental outcomes. However, almost half of the studies in which children were assessed at older ages showed no relationship between age at identification or intervention and language outcomes. Use of structural equation modeling (SEM) with the SKI*HI National Data Set did not result in models that adequately fit the underlying data. As such, these methods did not result in findings from which we can draw strong conclusions regarding the relationship between age at identification of hearing loss or intervention and child outcomes. To conclude, we know too little about whether earlier identification and intervention is better for children born with hearing loss or who acquire it at young ages. In addition to stronger research designs with sufficient sample sizes, use of reliable measures to collect a broader a [...]