Attitudes toward illegal immigration:: A cross-national methodological comparison

This research is an examination of the generalizability of a Likert-type scale originally devised to measure attitudes toward illegal immigrants (IA) in the United States. The current authors administered this scale across 4 national samples using several methodological procedures. Undergraduate students (63 1) responded to the IA scale (R. Ommundsen & K. S. Larsen, 1997) at the University of Oslo, Oregon State University, the University of Copenhagen, and Vrije University of Amsterdam. The authors' main purpose was to evaluate the adequacy of the 20-item IA scale by examining possible problems with method and translation. A translation study carried out with the Danish, Norwegian, and Dutch samples (N = 299) showed that the various national versions were fairly accurate and contained largely the same meanings. The use of procrustes analysis of the IA scale yielded 3 factors in all 4 national samples. The coefficient of congruence of these 3 orthogonally rotated factor matrices, with the U.S. factor matrix as target, varied from .80 to .95, supporting the cross-national robustness of the scale. In the search for a more economical cumulative scale, a Mokken analysis yielded a 5-item scale that represented the aforementioned 3 factors and was stable across national samples.