Evaluating Discrepancies in Print Reading Disability Statistics through Cognitive Interviews

According to the 1991-92 Survey of Income and Program Participation (SIPP), 9.7 million people 15 years old and older have difficulty reading newspaper print, and 1.6 million of those cannot read newspaper print at all. According to combined 1977 and 1984 National Health Interview Survey (NHIS) data, 4.2 million people 15 years old and older cannot read newspaper print. The attached report, Evaluating Discrepancies in Print Reading Disability Statistics through Cognitive Interviews, (prepared by Paul Beatty, National Center for Health Statistics, and Wendy Davis, Bureau of the Census) investigates reporting and questionnaire differences as possible reasons for the different estimates produced by the two surveys. Cognitive interviews were conducted with 30 respondents, selected to represent persons who were borderline in how difficult it was for them to see newspaper print sufficiently to read it. The results of this small study cannot support general conclusions, but the interviews do suggest that differences in question wordings and context contribute to the discrepant estimates. (There are other contributing factors as well.) The authors conclude that: < the NHIS and SIPP questions do appear to tap the intended construct of " being able to see to read ordinary newspaper print. " < cognitive interviews confirm that the three questions differ in the severity of the disability required to give a " disabled " response. From the least severe ('difficulty seeing words and letters ") in SIPP, we infer 9.7 million disabled; from the intermediate question (can't " see well enough to read ") in NHIS, we infer 4.2 million; and from the most severe (" cannot see words and letters at all ") in SIPP, we infer 1.6 million. NHIS and SIPP obtain different estimates because the questions are different. < the NHIS question, but not the SIPP questions, is preceded by other questions on vision which, the evidence suggests, may influence respondents to give more sanguine reports about their ability to see newspaper print. < the NHIS question does not as clearly establish as the SIPP question does that the question asks about respondents' ability to see with glasses (if they ordinarily wear them). It would be desirable to replicate this study in a larger scale field experiment, and to compare and evaluate the quality of the data produced by the different questions. If compatible with NHIS and 2 SIPP objectives, it would be desirable to use more comparable measures in both surveys …