Although, a few limitations existed; sample sizes were relatively low, most research was limited in degrees of hearing losses or failed to account for demographic variables associated with language development. Yoshinaga, et. al., worked to avoid such limitations by gathering data from a large sample (150 children), not constricted to high risk criteria, which allowed for the full spectrum of hearing loss (mild to profound) and in analyzing the data, they accounted for demographics such as, cognitive ability, age at testing, mode of communication, gender, degree of hearing loss, minority status, socioeconomic status, and whether additional disabilities were active. However, to acquire definite results, the researchers would have obtained a random sample of children, arbitrarily assigned to each group, early or later identified. In doing so, the later identified group would be withheld early intervention services due to later identification resulting in unethical practice, therefore, early and later identified groups in this population have to be naturally occurring in order to