Development of an Affective State Scale for Sports Performance Prediction for University Student-Athletes

This study aimed to develop an Affective State Scale for Sports Performance Prediction (ASSSPP) for university student-athletes in specific situations such as before a competition or carrying out a play. Questionnaire surveys designed for university student-athletes were performed to select and narrow down possible affective terms that could be the items for the scale. A factor structure was then constructed by carrying out factor analyses to confirm the reliability and validity of the scale. Based on the circumplex model of affect, the created ASSSPP was composed of 16 items of affective terms, which were distributed two-dimensionally into four domains by pleasant/unpleasant and high/low activation, each domain corresponding to four factors (activated-pleasant, unactivated-pleasant, activated-unpleasant, and unactivated-unpleasant) to evaluate affection states that are at a high level of abstraction. The reliability of the ASSSPP was confirmed from the viewpoints of internal consistency (Cronbach′s α) and stability (test-retest). Sufficient criterion-related validity was also confirmed by correlation analyses with the Japanese version of Positive and Negative Affect Schedule (PANAS) and performance level. Additionally, construct validity was confirmed for the verification model that set the affective states in the pleasant/unpleasant domains as the independent variable and the performance level in competitions as the dependent variable by covariance structure analyses. Therefore, it was suggested that ASSSPP has a sensitive and specific reactivity with athletes′ affective state in a specific situation of sports competitions and thus it is able to predict their subsequent performance level.