Comparing Cognitive Interviewing and Psychometric Methods to Evaluate a Racial/Ethnic Discrimination Scale
Dados Bibliográficos
| AUTOR(ES) | |
|---|---|
| AFILIAÇÃO(ÕES) | Harvard School of Public Health, Department of Society, Human Development and Health, Boston, MA, USA, University of Washington, Center for Multicultural Mental Health Research, Cambridge Health Alliance & Harvard Medical School, Somerville, MA, USA, Applied Research Program, National Cancer Institute, National Institutes of Health, Bethesda, MD, USA, UCLA School of Public Health, Los Angeles, CA, USA, The University of North Carolina at Chapel Hill, Cancer Prevention Institute of California, Fremont, CA, USA, Health Services and Economics Branch, Applied Research Program, National Cancer Institute, National Institutes of Health, Bethesda, MD, USA, Westat, Rockville, MD, USA |
| ANO | 2011 |
| TIPO | Artigo |
| PERIÓDICO | Field Methods |
| ISSN | 1525-822X |
| E-ISSN | 1552-3969 |
| EDITORA | Annual Reviews (United States) |
| DOI | 10.1177/1525822x11416564 |
| CITAÇÕES | 5 |
| ADICIONADO EM | 2025-08-18 |
| MD5 |
4bac815fd706afdd8f3fcc3c197bfd82
|
Resumo
Proponents of survey evaluation have long advocated the integration of qualitative and quantitative methodologies, but this recommendation has rarely been practiced. The authors used both methods to evaluate the 'Everyday Discrimination' scale (EDS), which measures frequency of various types of discrimination, in a multiethnic population. Cognitive testing included 30 participants of various race/ethnic backgrounds and identified items that were redundant, unclear, or inconsistent (e.g., cognitive challenges in quantifying acts of discrimination). Psychometric analysis included secondary data from two national studies, including 570 Asian Americans, 366 Latinos, and 2,884 African Americans, and identified redundant items as well as those exhibiting differential item functioning (DIF) by race/ethnicity. Overall, qualitative and quantitative techniques complemented one another, as cognitive interviewing findings provided context and explanation for quantitative results. Researchers should consider further how to integrate these methods into instrument pretesting as a way to minimize response bias for ethnic and racial respondents in population-based surveys.