Effect of language of interview on the validity and reliability of psychological well-being scales

This study tested the hypothesis that the use of different languages in a telephone survey could adversely affect the cross-cultural comparability of standardized research measures. The data used in the analysis were taken from the 1988 National Survey of Hispanic Elderly People (NSHEP) to test this hypothesis. The NSHEP survey contains data on 2,299 Hispanic persons aged 65 and older, and it was designed to study the living arrangements, health status, use of services, and attitudes toward aging problems held by the respondents. A telephone interview was used to collect the data. A majority of the respondents (86.6 percent) were interviewed in Spanish, and 13.4 percent were interviewed in English. LISREL-7 was the method of statistical analysis used to examine the factor equality of the positive affect and negative affect scales between the English and Spanish language respondents. Our findings indicate that there was no significant difference between the factor structure of the negative affect scale between the two groups; however, the factor structure associated with the positive affect scale demonstrated marked factorial differences between the two groups.