The State-Trait Anxiety Inventory (STAI) (6) has been used extensively in recent years as both a clinical and research tool. The instrument was designed to measure two separate components of general anxiety : state anxiety (A-State) and trait anxiety (A-Trait).
A-State is conceptualized as a transitory condition that varies in intensity and fluctuates in time in reaction to circumstances that are perceived as threatening. A-Trait, on the other hand, is conceived of as a relatively permanent personality disposition that tends to remain stable over time and across situations. Theoretically, persons who are high in A-Trait, or "anxiety proneness," will exhibit A-State elevations more often and more intensely than persons low in A-Trait because they tend to perceive a wider range of situations as threatening.
Although an impressive amount of literature exists in support of Spielberger's ( 6 ) theoretical conceptualizations of A-Stat,e and A-Trait, recent investigations by N e ~m a r k ' ~, 5 , and Wadsworth, Barker, and Wilson (') have demonstrated that A-Trait scores are not invariable and are amenable to therapeutic intervention. Such findings raise questions about the validity of the constructs that underlie the two separate scales. For example, do the A-State and A-Trait scales actually measure two different dimensions of anxiety? In addition, if two separate factors are reflected by scores on the state and trait measures, are the factors of about equal substantive value or is one so pervasive that the other has insignificant value?
A survey of recent literature indicated that the validity of the state and trait concepts has been examined by one previous factor analysis. Bartsch and Nesselroade ( l ) employed a manipulative, factor-analytic design and obtained results that seem to support the State-Trait distinction. However, their methodology was applied to a reduced set of Spielberger's anxiety items and also eliminated the standard instructions that normally accompany the STAI. The importance of the standard instructions has been emphasized ( 6 , p . 3 , :
"The validity of the STAI rests upon the assumption that the examinee has a clear understanding of the 'state' instructions which require him t o report how he feels at this moment and the 'Trait' instructions which ask him to indicate how he generally feels. The examinee's attention should be directed to the fact that the instructions are different for the two parts of the inventory, and he should be told that both sets of instructions must be read carefully."
Thus it appears that the factor structure of the STAI, when administered in its entirety and under standard instructions, has not been explored.
This study was designed to evaluate the State-Trait anxiety distinction by applying a factor-analytic strategy to A-State and A-Trait scores obtained under standard procedures of scoring and test administration.