Graduation Year

2011

Document Type

Dissertation

Degree

Ph.D.

Degree Granting Department

Psychological and Social Foundations

Major Professor

Robert Dedrick

Keywords

Cross-cultural Measurement, Cross-cultural psychology, Invariance Testing, Irrational Beliefs, Translations and adaptations of psychological tests

Abstract

The Irrational Beliefs Inventory (IBI) was built to measure self-defeating beliefs as conceptualized in Rational-Emotive Behavior Therapy. The IBI has five factors: worrying, rigidity, problem avoidance, need for approval, and emotional irresponsibility. A three-phase cross-cultural study was conducted to translate and adapt the IBI from English to Icelandic, and a Confirmatory Factor Analysis (CFA) approach was used for a test of factorial validity and cross-cultural invariance.

In Phase 1, the IBI was translated from English to Icelandic, using a forward-translation and back-translation. Two forward-translators and two back-translators were recruited. In Phase 2, qualitative interview methods were used in both the U.S. and Iceland to gain insights into the meaning of the items on the IBI. In the U.S., 21 university students provided insights in a group discussion, and four students were individually interviewed in depth about individual items on the IBI. In Iceland, four university students were interviewed in depth about the meaning of individual items. Three Icelandic psychology professionals were recruited to evaluate the appropriateness of the IBI for the Icelandic culture. In Phase 3, a Confirmatory Factor Analysis (CFA) was conducted to check factorial validity and cross-cultural invariance of the IBI. The total sample size in Phase 3 was N=1547, all college students, with n=827 in the U.S. and n=720 in Iceland. Overall, the CFA did not support the fit of the IBI's original five-factor model, although the fit was slightly better in the Icelandic version. Fit indices conflicted; the chi-square and comparative fit index (CFI) showed poor fit, while the RMSEA and SRMR showed acceptable fit. Correlated error was found between 85 item pairs in the U.S. model, and between 68 item pairs in the Icelandic model. Modifications were attempted to the original model by including the correlated errors, and a multigroup CFA was conducted. Adding the correlated errors slightly improved the fit of both models, but only 11 out of the IBI's 50 items were found to have equivalent item factor loadings and intercepts between the countries. Results from the psychometric analysis and qualitative interviews indicated that the IBI needs to be rewritten if the measure is to be used for research in Iceland. The results were discussed in light of a recent analysis of REBT-based measurement instruments, and implications for cross-cultural research on highly abstract constructs such as irrationality were discussed.

Share

COinS