Contextually Appropriate Measurement: Content Validity in a Neighborhood of Managua, Nicaragua

Purpose: Creating culturally valid measurement is essential to research. Many studies neglect to investigate the differential psychometric properties of these scales within and between racial and ethnic minorities (Mui, Burnette & Chen, 2001). In fact, even when measures are found to be valid in a translated version, they are often not contextually appropriate because they fail to account for SES, education level, or how the meaning of concepts are constructed in different communities or cultures. Therefore, the content validity of an instrument cannot be assumed across or within cultures. While the measurement items may be translated with literal accuracy, the underlying concepts may be flawed because they may have different meanings in different contexts. Measures themselves are not valid, rather their validity lies in “the scores that they yield and the interpretation we make of them” (Sechrest, 2005, p. 1586). The absence of appropriate measurement contributes to a lack of understanding of health and mental health disparities and the ability to develop strategies to alleviate them (Ramirez, Ford, Stewart & Teresi, 2005). This paper reports on the content validity of the measures used in a study investigating predictors of mental health in a low income Nicaraguan neighborhood.

Method: Face-to-face interviews were conducted with 227 adult men and women in Barrio 3-80 of Managua. Participants were asked questions using standardized surveys which had been found to be reliable and valid among some Spanish-speaking populations: Rosenberg Self Esteem Scale, BEM Sex Role Inventory, Multidimensional Scale of Perceived Social Support, and General Health Questionnaire. Six participants were asked open-ended questions regarding gender roles and mental health. Using an item response approach, thematic analysis was used to assess content validity of the measures. Scale item distributions for quantitative data were also analyzed.

Results: Scale item distributions of quantitative data revealed a bimodal response structure, especially for the BSRI. Corroborating this finding, thematic analysis of item responses revealed that respondents had difficulty understanding the notion of a continuum of responses (i.e., the BSRI's 7 point Likert scale). Qualitative interviews found a cultural resistance toward revising traditional gender roles. Regarding mental health, 98.2% of respondents displayed psychological distress (GHQ); however qualitative interviews indicated that people viewed mental distress as a way of life, but not necessarily as a mental illness.

Implications: In Social Work, translated standardized measurements are used to assess outcomes for programs and research in diverse communities. Applying socially constructed concepts in a universal fashion needs to be critically examined. The content validity of measurements created from a Western viewpoint must be evaluated to assess if the measured constructs match the community's cultural meaning of the concept. Consequently, quantitative surveys may need revision to reflect these meanings in order to be used in these communities. In sum, the language used in measurement instruments is not universally relevant. Thus, social and cultural basis of a socially determined response to such concepts must be taken into account by social work researchers.