Examining the relationship between worry and trait anxiety

College Student Journal, Sept, 2004 by William E. Kelly

The relationship between worry and trait anxiety was explored. University students (N = 218) completed the trait form of the State-Trait Anxiety Inventory (STAI) and the Penn State Worry Questionnaire (PSWQ). The STAI was found to include four factors: Negative Affect, Worry, Fatigue, and Avoidance. STAI Worry was the best predictor of PSWQ scores. However, Negative Affect and Fatigue also significantly contributed to PSWQ scores. It was concluded that inclusion of worry in the STAI may inflate correlations between trait anxiety and worry, but cannot completely account for the relationship between trait anxiety and worry.

**********

It is commonly assumed that worry and anxiety are synonymous. However, there is growing evidence that anxiety and worry are separate, yet related, constructs (i.e., Davey, Hampton, Farrell, & Davidson, 1992; Davey, 1993; Gana, Martin, & Canouet, 2001 ). Worry, is generally defined as a stream of negative thoughts (Kelly & Miller, 1999). Anxiety, however, tends to include somatic tension, fear, and a subjective sense of unease (Barlow, 2002). One of the most common methods of operationally defining anxiety, is the construct of trait anxiety, as measured by the trait form of the State-Trait Anxiety Inventory (STAI; Spielberger, Gorsuch, & Lushene, 1970).

Research has found a consistent, robust relationship between the STAI and worry. For instance, across six articles identified for the present study (Constans, 2001; Davey, 1993; Davey, 1994; Davey et al., 1992; Russell & Davey, 1993; Wells, 1994), 11 correlations between trait anxiety and various measures of worry were reported ranging from .48 to .74, with an average of .61.

While trait anxiety is purported to represent anxious responses to stressful situations (Gregory, 2000), the STAI seems to reflect a higher-order factor of negative emotional experience and sub-factors of depression and anxiety (Bieling, Antony, & Swinson, 1998). An inspection of the STAI items reveals that the items thought to measure anxiety often reflect elements of worry (i.e., "I worry too much over something that really doesn't matter"). It is possible, therefore, that the STAI includes a worry component, in addition to anxiety. If so, this is a potentially important methodological problem which could contribute to inflated findings of the relationship between anxiety and worry. The purpose of this study was to explore: 1. whether there is a worry component measured by the STAI, and 2. which components of the STAI most influence the relationship between trait anxiety and worry.

METHOD

Participants and Procedure

After obtaining informed consent, 218 (151 female) university students enrolled in undergraduate psychology courses at a small public university in the southern U.S. were administered scales measuring trait anxiety and worry. The average age of the sample was 25.9 (SD = 9.6).

Measures

State-Trait Anxiety Inventory--Trait (STAI). As a measure of trait anxiety, the trait form of Spielberger's State-Trait Anxiety Inventory, Form X-2 (STAI; Spielberger et al., 1970) was utilized. The STAI is a twenty-item self-report instrument designed to assess trait anxiety. Participants indicated their agreement with each item on a Likert scale ranging from 1 = "not at all" to 4 = "very much so." The STAI has good reliability and validity (Spielberger et al., 1970). The 20 items are summed to produce a total score with higher scores indicating more trait anxiety.

Penn State Worry Questionnaire (PSWQ). Worry was assessed using the 16-item Penn State Worry Questionnaire (PSWQ; Meyer, Miller, Metzger, & Borkovec, 1990). Participants respond to each item using a Likert scale ranging from 1 = "not at all typical" to 5 = "very typical." Good reliability and validity have been reported (Meyer et al., 1990). Responses to each item are summed to produce a total PSWQ score with higher scores indicating more worry.

RESULTS AND DISCUSSION

The Coefficient Alpha of the: total PSWQ and STAI were .94 and .91, respectively. As in previous research, the two scales correlated significantly, r = .75, p < .0001. To explore the potential inclusion of a worry component in the STAI, an exploratory principal components factor analysis with a varimax rotation was calculated on STAI items. Four factors were extracted with eigenvalues greater than 1 accounting for a total of 57.8% of the variance in responses. In attempts to provide some "practical significance" of findings, a stringent factor loading cut-off criteria of .50 was chosen (Hair, Anderson, Tatham, & Black, 1995). Using this criteria, Factor 1 included eight items ([alpha] = .87) accounting for 21.5 % (eigenvalue = 7.6) of the variance reflecting lack of contentment, insecurity, and unpleasant affect. Thus, this factor was named Negative Affect. Factor 2 included four items ([alpha] = .79) accounting for 15.3% (eigenvalue = 1.8) of the variance reflecting unpleasant thinking and resulting tension. It was termed Worry. Factor 3 included two items ([alpha] = .69) accounting for 10.8% (eigenvalue = 1.1)of the variance reflecting tiring easily and not being rested. Thus, it was named Fatigue. Factor 4 included three items ([alpha] = .56) accounting for 10.2% (eigenvalue = 1.1) of the variance reflecting avoidance and resulting disappointment. It was termed Avoidance. Three items reflecting depressive experience did not load on factors (1).