Side Navigation

Research Article

Psychometric Properties of the State Anxiety Inventory (SAI) in the Context of Acute Aerobic Exercise: Another Look

Ekkekakis, P., Hall, E.E., & Petruzzello, S.J., University of Illinois

The SAI (Spielberger, 1983) has been criticized for exhibiting reduced internal consistency in the context of acute aerobic exercise and for confounding activation with affective valence, producing potentially invalid total state anxiety scores during and post-exercise. Two studies evaluated the psychometric properties of the SAI under such conditions. In Study 1, 45 female undergraduate volunteers participated in a 50-min session of aerobics at an intensity of 59% HRmax reserve. They completed an 8-item version of the SAI before, 25-min into, and immediately following the exercise bout. During exercise, coefficient alpha and mean inter-item correlation were reduced (.71 from .87 pre and .23 from .46 pre, respectively). Although the total SAI score remained unchanged from pre to during exercise, 2 items increased and 4 decreased. A measure of activation (Svebak & Murgatroyd, 1985) accounted for 7% (p<.05) of variance in total SAI score during and 26% (p<.001) following the exercise bout, beyond the variance accounted for by a measure of valence (38% and 16%, respectively). In Study 2, 69 undergraduate volunteers completed a 10-item version of the SAI before and after running on a treadmill for 30 min at 75% VO₂max. The total SAI score was increased (p<.001). Of the items, 5 indicative of activation showed increases, but 2 indicative of apprehension showed decreases. Internal consistency was again reduced. Concurrent assessment of Energetic Arousal (positively valent activation) also showed a significant increase. These findings indicate that the reliability and validity of the SAI as a measure of state anxiety are severely compromised in the context of acute exercise.