ERIC Number: ED478198
Record Type: RIE
Publication Date: 2002-Nov
Reference Count: N/A
Measuring Cognitive Function: An Empirical Investigation of the Psychometric Properties of a Cognitive Measure.
Witta, E. Lea; Sivo, Stephen A.
Herzog and Wallace (A. Herzog and R. Wallace, 1997) discussed a measure designed to assess the cognitive functioning of older adults who participated in the study formerly known as the Asset and Health Dynamics among the Oldest Old (AHEAD). The measure derived from four well-known tests of cognitive functioning, but improves on them by combining elements of each emphasizing those aspects most relevant to the cognitive changes in the gerontological population. This measure promises to allow researchers to identify cognitive changes that may lead to dementia more effectively. While this measure has been used to assess large numbers of people, it has not been scrutinized empirically as an evaluative tool to assess the internal and external structural validity evidence of the scores produced. To understand this underlying factor structure of the instrument better, longitudinal congeneric, tau-equivalent, and parallel models were fit using five waves of the Health and Retirement study data (previously the AHEAD study) obtained from the University of Michigan. There were 2,681 male respondents and 3,841 females. The final three survey years provided surprisingly consistent models of the cognitive indicators. The first two survey years did not. Results indicate that the three measures used as indicators of cognition (immediate recall, delayed recall, and reverse 7s) are neither tau-equivalent nor parallel. Immediate and delayed recall, however, are equivalent measures of cognition when the word list contains 10 words, but are not parallel. (Contains 2 figures, 3 tables, and 9 references.) (Author/SLD)
Publication Type: Reports - Research; Speeches/Meeting Papers
Education Level: N/A
Authoring Institution: N/A
Note: Paper presented at the Annual Meeting of the Ameri