OBJECTIVETo develop a suitable physical exercise behavior-psychological scale for the Chinese adolescents and to evaluate its validity and reliability.

METHODSA total of 3600 junior students were recruited and tested through multistage sampling method. Reliability was assessed, using Cronbach's α and split-half reliability; while exploratory factor analysis and confirmatory factor analysis were used to test its validity, with entries-dimension correlation coefficient (IIC), correlation coefficient between the scores and the dimension, the dimension of correlation coefficient test content validity.

RESULTSThe internal consistency reliabilities for total scale score (Cronbach's α = 0.888), Cronbach's α of four domains were 0.880, 0.706, 0.552 and 0.839, respectively. The four-domain split-efficacy reliabilities were ranged between 0.559 and 0.876. Data from the exploratory factor analysis revealed the following dimensions: the entries were all inclusive; the cumulative contribution rate was more than 40%.

RESULTSfrom the confirmatory factor analysis showed that the absolute fit measures of GFI = 0.972, AGFI = 0.958, RMSEA = 0.032 (90%CI: 0.030 - 0.034) could satisfy the condition while the model fit was better. The results of IIC showed: in the four dimensions, and the contained entries between Spearman correlation coefficient range were 0.396 - 0.700, 0.470 - 0.709, 0.696 - 0.771 and 0.665 - 0.813, while other dimensions from weak to moderate relationships, the r value range was from 0.386 to 0.935. With the change of the phase changing process, decision balance, self-efficacy scores were differently prompting the health-related intervention periodically and pertinently.

CONCLUSIONThe reliability and validity of the 'Adolescent physical exercise behavior-psychological scales (Chinese version) ' were good, and could be used to measure the adolescent's physical training behavior-psychological conditions.