Issues in Use of Standardized Assessment Instruments

Appropriateness of English Version

Educational Level: Results of the test may be influenced by variables other than culture, such as educational background or socioeconomic status. For example, high false-negative rates of cognitive impairment have been reported among highly educated subjects, whereas high false-positive rates of cognitive impairment have been reported among less educated clients.

Established Population Norms

  • Norms for validated measures are usually available only for the largest of the specific language groups
  • Importance of using versions adapted for subgroup of language category (e.g., Puerto Rican, as opposed to generic Spanish)

  Back to Top |

Appropriateness of Format

Level of Difficulty

response format (e.g. true false/multiple choice, Likert, semantic differential, or visual analogue may be more difficult for elders not familiar with that type of format) vs. a simpler yes/no format.

Verbal Expression of Feelings

Some elders or family members (especially those from Asian backgrounds) may be more reluctant to express their feelings in interviews than written self-report, and/or may tend be more likely to respond in accordance to social desirability rather than their true feelings/attitudes in interviews compared to self-report format.

Font Size

Some recommendations are 16 point font with capitals for older adults:


Reading Level

  Back to Top |

Translated Instruments

Adequacy of translated version depends on the following domains:

Content Equivalence: to ensure the content in each item in the instrument has consistent cultural relevance.

Semantic Equivalence: to ensure the meaning of each item remains conceptually and idiomatically the same.

Technical Equivalence: to ensure that the methods of assessment (interviews, observation, self-report) elicit comparable data.

Criterion Equivalence: to establish the normative interpretation of the variable.

Conceptual Equivalence: to ensure the same theoretical construct is being measured in each culture.

  Back to Top |

Culturally and Linguistically Appropriate Measures

Cognitive Screening and Dementia Assessment

The Cognitive Assessment Screening Instrument (CASI) (Teng, et al, 1994) was specifically designed for easy cross-cultural adaptation and has been adapted and validated in the following populations English speaking North Americans, English speaking Chamorros in Guam, Chinese elders in Kimmen and Taiwan, and Japanese elders in Seattle, Honolulu and Japan. It is continually being adapted and used with other populations. The CASI-short (Teng et al.,1998) consists of four items and is particularly good for screening elders with very low levels of education (Dick et al, 2006).

For more comprehensive cognitive evaluation, the Cross-Cultural Neuropsychological Test Battery (CCNB) (Dick et al., 2002) includes the CASI and 10 additional tests to assess six cognitive domains. It has been used successfully with African American, Caucasian, Chinese, Hispanic, and Vietnamese elders (Dick et al., 2006).

The Spanish and English Neuropsychological Assessment Scales (SENAS) (Mungas et al., 2005) was developed with the goal of creating psychometrically matched English and Spanish language neuropsychological measures of cognitive abilities. The scales assess seven domains and include both verbal and nonverbal domains (Mungas, 2006).

For discussion of assessment of cognition and dementia in American Indian and African American elders, see Hargrave, 2006 and Jervis, Cullum, & Manson, 2006.

Measures of Depression

The Geriatric Depression Scale (GDS) has been translated into at least 30 languages, many with multiple versions.

Ada Mui developed a revised version of the GDS short form in Chinese, that was found to have greater internal consistency among Chinese American elders than the original GDS shortform (Mui, 1996). She and colleagues subsequently used it successfully also with Filipino, Asian Indian, Japanese, Korean, and Vietnamese elders in New York.

For an excellent review of validation studies of the GDS with diverse populations as well as description of validation of the Mui GDS short form in multiple languages see Mui et al, 2003.