The CRAVE and ARGE scales for motivation states for physical activity and sedentarism: Brazilian Portuguese translation and single-item versions.

Abstract

Motivation states for physical activity and sedentarism potentially vary from moment to moment. The CRAVE scale (Cravings for Rest and Volitional Energy Expenditure) was developed to assess transient wants and desires to move. Three studies were conducted with the aims of: (1) translating and validating the scale in Brazilian Portuguese, (2) examining changes with exercise, and (3) determining the best single-item for Move and Rest subscales for English and Portuguese. In Study 1, six bilingual speakers translated the scale into Brazilian Portuguese [named Anseios por Repouso e Gastos com Energia (ARGE)]. The ARGE had good content validity coefficients across three dimensions (0.89-0.91), as determined by three independent, bilingual referees. 1,168 participants (mean age = 30.6, SD = 12.2) from across Brazil completed an online version of the ARGE. An Exploratory Factor Analysis found two clear, oblique, and inversely related factors (Move and Rest; GFI = 1.00, RMSR = 0.03). Reliability was good (Cronbach α's: 0.93 and 0.92). Two models of the scale (10 vs. 13 items) were compared with Confirmatory Factor Analysis. The previously validated version using 10 scored items (GFI = 1.00, RMSEA = 0.07, RMSR = 0.02) outperformed the version scored with 13 items. State anxiety and exercise behavior had small associations with Move and Rest (-0.20 to 0.26). In Study 2, ARGE Move scores had high correspondence post-session (ICC = 0.83) for 9 women performing short Sprint Interval Training (sSIT; 6 sessions). Large, but non-significant, effects were detected for changes in motivation states with sSIT. In Study 3, IRT analyses found that for the United States sample, "be physically active" and "be still" were the most representative items for Move and Rest, respectively, while for the Brazil sample they were "exert my muscles" and "be a couch potato." Overall, it was found that: (A) the ARGE scale demonstrated good psychometric properties, (B) the original scoring (with 10 items) resulted in the best model, (C) it had small associations with exercise behavior, and (D) the subscales were reduced to single items that varied by country, indicating potential cultural differences in the concept of motivation states for physical activity.