IMPACT OF DATA COLLECTION MODE ON LOW-LITERACY SUBJECTS

点击翻译按钮获取中文摘要

It is estimated that 47% of the U.S. adult population (90 million adults) have low literacy skills. Studies conducted in the United States and in developing countries have found that populations with poor literacy skills tend to have poorer health as compared to adults with higher levels of literacy. This association remains when other socioeconomic factors are controlled in analysis. In order to further examine this relationship and why it may exist, investigators need to be assured that the modes of data collection being used are appropriate for respondents with low literacy skills. Few research studies have been conducted to assess the impact of data collection mode with subjects who are low literate. The proposed randomized study will examine the impact of three modes of data collection on data quality with a sample from this population. The three modes to be studied are interviewer administered (IA), paper-and-pencil self- administered (PPSA), and audio computer-assisted self-interviews (ACASI). Two measurement instruments frequently used in health research will be used in this study: the Short-Form Health Survey (SF-36) and the Center for Epidemiologic Studies Depression Scale (CES-D). Because poor literacy is more prevalent among minority and low income populations, subjects will be low income African- American women. To be eligible, subjects must be between the ages of 35-64 years and grade equivalent reading abilities at or below the 6th grade level, as measured by the Rapid Estimate of Adult Literacy in Medicine (REALM). The specific aims of the study are to 1) examine the effects of mode of data collection on data quality, characterized by level of reporting, rates of missing values, scale internal consistency; and response bias; and 2) to examine the relationship between the cognitive burden imposed by the measurement instruments and subjects cognitive abilities.