Hershberger SL, Fisher DG, Archuletta E, Rodriguez J; International Conference on AIDS (15th : 2004 : Bangkok, Thailand).

Int Conf AIDS. 2004 Jul 11-16; 15: abstract no. D10725.

Cailfornia State University, Long Beach, Long Beach, United States

Background: Psychometric studies, comparing the measurement characteristics of the English and Spanish Language versions of the SOCRATES 8-S (for sex behavior), have not been reported. Test equating methods based on item response theory (IRT) can be used to verify that the psychometric properties of the SOCRATES 8-S, which include item reliability, item difficulty, and item discrimination, are identical for the English and Spanish versions. Although the data suggest that the English language version is reliable, comparable data for the Spanish language version are unavailable. Methods: The SOCRATES was administered on two occasions 48-hours apart. At the first administration there were 89 English-speaking and 60 Spanish-speaking participants, at the second, 69 and 51. Results: The 48-hour test-retest reliability of the SOCRATES, combining the English and Spanish language versions, was.868; Cronbach's alpha, for the first occasion of measurement, was.931, and for the second occasion,.941. Comparable numbers for only the English language version were.878,.939, and.923; for the Spanish version,.863,.922, and.932. PARSCALE (Muraki & Bock, 1997) was used to estimate the item parameters simultaneously for the two test versions, conditional on the equality of their item parameters. The chi-square goodness-of-fit test for the model of equality was not rejected: chi-square (23) = 14.292, p =.918. The mean item reliability was.699; the mean item difficulty was 1.340; and the mean item discrimination was.832. Conclusions: The psychometric characteristics of the two versions of the SOCRATES are identical: Both versions are highly reliable, have items of above average "intensity," and items which discriminate well among different levels of behavior. Based on these results, the versions do not have to be formally equated. Their measurement characteristics are already equivalent.

Publication Types:

• Meeting Abstracts

Keywords:

• Bias (Epidemiology)
• Chi-Square Distribution
• European Continental Ancestry Group
• Factor Analysis, Statistical
• Hispanic Americans
• Language
• Psychometrics
• Research Design

Other ID:

• GWAIDS0033608

UI: 102277824

From Meeting Abstracts