Psychometric properties and concurrent validity of two exercise addiction measures: A population wide study

Objectives: The existence of exercise addiction has been examined in numerous studies. However, none of the measures developed for exercise addiction assessment have been validated on representative samples. Furthermore, estimates of exercise addiction prevalence in the general population are not available. The objective of the present study was to validate the Exercise Addiction Inventory (EAI; Terry, Szabo, & Griffiths, 2004), and the Exercise Dependence Scale (EDS; Hausenblas & Downs, 2002b), and to estimate the prevalence of exercise addiction in general population. Design: Exercise addiction was assessed within the framework of the National Survey on Addiction Problems in Hungary (NSAPH), a national representative study for the population aged 18-64 years (N = 2710). Method: 474 people in the sample (57% males; mean age 33.2 years) who reported to exercise at least once a week were asked to complete the two questionnaires (EAI, EDS). Results: Confirmatory Factor Analysis (CFA) indicated good fit both in the case of EAI (CFI = 0.971; TLI = 0.952; RMSEA = 0.052) and EDS (CFI = 0.938; TLI = 0.922; RMSEA = 0.049); and confirmed the factor structure of the two scales. The correlation between the two measures was high (r = 0.79). Results showed that 6.2% (EDS) and 10.1% (EAI) of the population were characterized as nondependent-symptomatic exercisers, while the proportion of the at-risk exercisers were 0.3% and 0.5%, respectively. Conclusions: Both EAI and EDS proved to be a reliable assessment tool for exercise addiction, a phenomenon that is present in the 0.3-0.5% of the adult general population. (C) 2012 Elsevier Ltd. All rights reserved.