Relation of Anxiety to Locus of Control

Several recent studies have reported significant relationships between various measures of general anxiety, test anxiety, and the locus of control scale (LC). Watson (1967) concluded that the correlation between locus of control and anxiety supports Mandler and Watson's (1966) hypothesis that actual or appraised lack of control (high external locus of control) produces anxiety. In order to determine whether there exists an anxiety factor within the locus of control scale, the Manifest Anxiety Scale (MAS), the Test Anxiety Scale (TAS), and the LC scale were administered to two samples of introductory psychology students at Vanderbilt University (N, = 323; Nq = 303). In an item analysis every item on the LC scale answered in exteral locus of control direction correlated postively with total score (high anxiety) on the MAS and TAS. However, the correlation of only 3 items from the LC scale actually reached significance ( p < .05) with the TAS and MAS in both samples (the significant IS ranged from . I 2 to .34). Total scores on the LC scale and the MAS were significantly correlated in both samples (sl = .40, rz = .30; p < .01). The LC scale and the TAS were also significantly correlated (sl = .22, r:! = .21; p < .01). A factor analysis using varimax rotation demonstrated that these significant correlations could not be explained in terms of an anxiety factor composed of highly correlated individual items on the LC scale which, when removed, would reduce the degree of relationship between the anxiety scales and the remaining locus of control items. These findings support the proposition that the anxiety scales and the LC scale are measuring conceptually separate variables which correlate with each other and that this correlation is not due to a hidden anxiety factor within the locus of control scale. A feeling of lack of control over the environment and the outcome of one's actions is associated with anxiety. Whether this appraised lack of control produces anxiety, or vice versa, cannot, of course, be determined from correlational studies of this type.