The Behavioral Regulation in Exercise Questionnaire (BREQ-3) Portuguese-Version: Evidence of Reliability, Validity and Invariance Across Gender

The main propose of the present study was to analyse the psychometric properties of the Behavioural Regulation Exercise Questionnaire (BREQ-3) in a sample of Portuguese exercisers and invariance across gender. Two independent samples (448 calibration; 374 validation) of both gender (495 female; 327 male) were enrolled in this study, with ages between 16 and 78 years (M = 40.29; SD = 16.24). The results show that the original model (six factors; 24 items) did not fit to the data in a satisfactory way (χ² = 977.49; df = 237; B-S p < .001; SRMR = .07; NNFI = .80; CFI = .83; RMSEA = .08; 90% CI = .08–.09). After removing six items (one for each factor), the model (six factors; 18 items) adjustment improved in a satisfactory way in both samples: calibration (χ² = 331.86; df = 120; B-S p < .001; SRMR = .06; NNFI = .91; CFI = .93; RMSEA = .06; IC90% .06–.07) and validation (χ² = 254.08; df = 120; B-S p < .001; SRMR = .04; NNFI = .93; CFI = .95; RMSEA = .06; IC90% = .05–.06). Results also showed model invariance across gender (∆CFI ≤ .01). The Portuguese version of BREQ-3 (six factors; 18 items) is a valid and reliable measurement instrument to measure behaviour regulation underlying self-determination theory in the exercise domain. However, the evidence also indicated that further empirical studies are needed to address the fragilities of the original model (six factors; 24 items).