Demographics and Sample Bias Estimates of the Deaf and Hard of Hearing School Age Population

Abstract

Demographics and Sample Bias Estimates of the Deaf and Hard of Hearing School Age Population Donald F. Moores Editorial Recently while organizing some material for publication, I was struck by how much we depend on data that either are very limited on the amount of information presented or that provide extensive information on samples that may not be representative of the population of deaf and hard of hearing students in the United States as a whole. As most readers are probably aware, the Child Count data mandated by the Individuals With Disabilities Education Act theoretically enumerates every deaf and hard of hearing child in the country who has an Individual Education Plan (IEP). However, the information reported is very limited; there is not even delineation between children who may be categorized as deaf or hard of hearing, let alone detailed reports on cause of hearing loss, communication mode in instruction, or support services, for example. The Annual Survey of Deaf and Hard of Hearing Children and Youth, conducted over a period of more than 30 years by the Gallaudet Research Institute (GRI), on the other hand, is a rich resource for the students that are included in their surveys. I am included with the many professionals in our field who have used the GRI data on several occasions, however with the knowledge that the samples do not necessarily represent the entire population of deaf and hard of hearing children served under IDEA. In other words, there are sampling biases. This, of course, has been acknowledged by researchers affiliated with the annual surveys and GRI. As long ago as 1986, Ries argued that there was a firmer basis for drawing conclusions for students with profound hearing losses in full-time programs than for students with lesser degrees of hearing loss or those receiving part-time services. There have been strong trends since 1986 toward educational placement of deaf and hard of hearing students in part-time placement or in full inclusion classrooms with all hearing classmates. Despite more efficient data-gathering techniques and improved technology it is clear that the task of identifying a representative sample of deaf and hard of hearing children is becoming more difficult each year. We are faced with a dilemma. There is no indication that IDEA will mandate more information from Child Count and the approximately 6,000,000 students with IEPs. As increasing numbers of deaf and hard of hearing students are educated in mainstream or inclusive environments they will become increasingly difficult to identify. This is particularly true because Child Count is federally mandated and participation in the annual survey is voluntary. Factoring in the massive amounts of paper work required by No Child Left Behind legislation, leaves motivation to respond the annual surveys a distant third. The question is whether or not existing data can be modified in some way to provide a more accurate representation of the American deaf and hard of hearing school age population. Mitchell (2004), at that time a member of the GRI, using data from the 1999–2000 annual survey, examined the extent to which the survey adequately represented the larger Child Count population and reported sampling biases. The Child Count for same year reported 71,695 deaf and hard of hearing students from ages 6 to 22 as compared to 31,493, approximately 44% of the child count total, in the annual survey. He reported that, among other things, the survey over-sampled Hispanic students, students with severe and profound hearing losses, students being taught in sign-only classes, and students with at least one deaf parent. Conversely, the survey under-sampled white students, students with mild and moderate hearing losses, students being taught in speech-only classes, students with no deaf or hard of hearing parents, and students receiving itinerant services. Mitchell even found that in 5 of 10 geographic regions of the United States the proportional representation on the annual survey was statistically different from that of the IDEA Child Count, with the west, south, and central cluster of states substantially over-represented while Middle Atlantic region was appreciably under-represented. Obviously, statements made on the basis of the annual survey data regarding racial/ethnic...