The goals of the present study were to construct a short scale of global identity that would have cross-cultural validity, and examine the latent factor structure of this scale. Four studies were conducted. Based on the literature review (Study 1) and data collected from students from several countries in open-ended questions (Study 2), an extensive item pool was produced. The item pool was then evaluated by experts (Study 3), resulting in total trial item pool of 110 Likert-type items. The questionnaire was then administered together with a social desirability scale to three different samples of students: Norwegian, Turkish, and American (Study 4). Items were ranked based on their performances on 22 psychometric criteria. The criteria were items standard deviations and item-total correlations for each sample, the length of items in word and characters for both the English and the Turkish version, items frequency of omission in each sample, items correlations with a multicultural index, social desirability, and cosmopolitan behavior in each sample. The highest 24 ranked items were subjected to item analysis, resulting in a 12 item Global Identity Scale that had high Cronbach alpha across samples. Factor analysis of the 12 item scale yielded a 2 factor orthogonal structure that was stable across the three samples of the study. One of the subscales of the Global Identity Scale represented lack of national attachment, and the other represented cosmopolitan orientation.
Key words: cosmopolitanism, globalization, identity, internationalism, scale development