Study Objective: The purpose of this study was to develop a new test of expressive language and evaluate it for evidence of culture bias in African-American, Hispanic American and European American older adults
Design: This was a psychometric study of a new instrument in which internal consistency, degree of difficulty of individual items and the test as a whole were compared across cultural groups.
Population, Sample, Setting: Fifty-three cognitively impaired African American, Hispanic American and European American residents of 5 long term care facilities and 120 unimpaired older adults from 6 senior centers were tested. The sample was 28% male and 72% female. They ranged in age from 54 to 98 (mean 77) and had an average Mini Mental Status score of 21, range 0 (lowest) to 30 (highest possible score).
Concepts: In this study, bias was defined as systematic error distortion in the measurement process. To evaluate an instrument for culture-related bias, individual items and the measure as a whole are analyzed to determine if they conform to a given set of psychometric rules in the same manner for all examined, regardless of cultural or ethnic group membership.
Methods: Following consent, sociodemographic information (age, gender, ethnic group membership, education and so forth) was obtained. All subjects completed the Mini-Mental State Examination and the 92 item (long) version of the newly developed Miami Naming Test of expressive language in individuals with Alzheimer's Disease and related dementias. This test version of the Miami Naming Test consisted of 92 original black and white, 8 1/2" X 11" drawings of items ranging in difficulty from dog to satellite dish. Each stimulus item was reviewed by a panel of experts on cognitive testing and minority aging prior to this testing.
Data Analysis: Means were compared across ethnic groups to evaluate difficulty and analysis of variance was used to compare total scores across ethnic groups controlling for mental status. Cronbach's alpha was used to evaluate internal consistency.
Findings: Significant differences in mean scores across the three ethnic groups were found in the long (92 items) version (f 6.25 p=.0024). Altogether 44 items were deleted on the basis of differences in difficulty or low alphas. No significant differences across ethnic groups were found on this shortened 48 item version. Cronbach's alpha of this final 48 item version was .97.
Conclusion: The Miami Naming Test may be useful across ethnic groups to assess the anomia found in Alzheimer's and related dementias.
Implications: Although it has long been recognized that cognitive tests are troublesome and potentially biased cross-culturally, these tests are administered to individuals from different cultural and ethnic groups without sufficient attention to their differential reliability and validity across cultural and ethnic groups.
Back to Research Instruments and Practices
Back to The Advancing Nursing Practice Excellence: State of the Science