Another look at the psychometric properties of the Anxiety Control Questionnaire

https://doi.org/10.1016/S0005-7967(98)00206-XGet rights and content

Abstract

This study examined the psychometric properties of the Anxiety Control Questionnaire (AxCQ), a recently developed 30-item instrument intended to assess an individual's perceived level of control over anxiety related events. The sample consisted of 303 college students (189 females, 110 males, 4 no response). Although the scale is reported to have a two factor solution, results of an exploratory factor analysis in this sample indicated a three factor solution was more appropriate. Significant correlations were found between the three factor scales (r's ranged from 0.28 to 0.62). Internal consistency appeared good for the total AxCQ and for the factor scales derived from both the two and three factor models. There were no significant gender differences. We conclude that the psychometric properties of the scale look promising, but recommend that the factor structure of the AxCQ continue to be explored.

Introduction

Recent years have witnessed a resurgence of interest in perception of control as a mediator of psychological distress (Barlow, 1988, Barlow, 1991, Rapee et al., 1996). Early research into the importance of perceived control focused mainly on the Rotter (1966) locus of control construct and his internal–external (I–E) locus of control scale. The I–E scale was designed to assess the extent to which an individual believes an outcome is a function of (or dependent upon) his or her own behavior rather than being dependent upon outside factors (e.g. luck, fate, chance). The popularity of the locus of control construct, and the I–E scale, are affirmed by the prolific amount of research into the topic.

Although the I–E scale was designed to measure a unidimensional construct, psychometric investigations indicated substantial item heterogeneity and that the scale was, in fact, not unidimensional Klockars & Varnum, 1975, Lange & Tiggeman, 1981. Apart from psychometric considerations, the scale was also criticized for its generalized nature and content (Rapee et al., 1996). These concerns led investigators to suggest that the construct of perceived control might be better conceptualized as multidimensional and attempts were thus made to examine more specific domains of perceived control Coan, Fairchild & Dobyns, 1973, Rotter, 1975, Harper, Oei, Mendalgio & Evans, 1990. While the theoretical importance of perceived control was maintained, the construct was more narrowly focused on a specific area of functioning or behavioral response. As a result, a variety of instruments have been developed to measure the impact of perceived control on domains such as social-political influence, impulse control, behavioral problems and health concerns Kirscht, 1972, Reid & Ware, 1974, Craig, Franklin & Andrews, 1984.

The locus of control construct has also been applied to affective disorders. Most notably, a perceived lack of control is often associated with the development of depression. People who have an external locus of control and believe that they are not able to control or manage their environment have an increased vulnerability to depression (Billings & Moos, 1985). Conversely, those who have an internal locus of control and feel a sense of mastery over their environment have an increased resistance to depression and stress Johnson & Sarason, 1978, Pearlin & Schooler, 1978, Kobasa, 1979.

Until recently, the contribution of perceived control to the development of affective disorders, other than depression, had been largely neglected. Although researchers have noted the comorbidity between anxious and depressive disorders, and have suggested that the association between the disorders may reflect the common contribution of a perceived lack of control, little empirical research has examined the contribution of perceived control to the development of anxiety disorders Alloy, Kelly, Mineka & Clements, 1990, Barlow, 1991. This is particularly surprising given that Barlow has suggested that a psychological vulnerability to early uncontrollable events may have etiological significance in the development of anxiety disorders. Thus, events that are perceived as out of one's control may not only lead to depressive symptoms but also to anxious symptomatology. To examine the relationship between perception of control and anxious symptomatology, Rapee et al. (1996) recently developed the Anxiety Control Questionnaire (AxCQ).

The AxCQ was designed to assess perceived control over ‘internal emotional reactions’ as well as ‘externally threatening events’ (Rapee et al., 1996, p. 279). In developing the AxCQ, the authors selected 53 items on the basis of face validity. To avoid response bias, approximately half of the items were reverse worded and scored. A 6-point Likert-type rating scale was used to force respondents to make some judgment as to whether they agreed or disagreed with the items. This initial version of the scale was administered to 250 individuals with an anxiety disorder of clinical severity. After an examination of redundant items and internal consistency, 15 items were deleted from the scale thus reducing the number of items to 38.

When the factor structure of the AxCQ was examined, 11 factors had eigenvalues greater than one; however, the authors reported that examination of the scree plot indicated a one, two or three factor solution (Rapee et al., 1996). After examination of the two to four factor solutions, the authors determined a two factor solution to be optimal in terms of scale interpretability. The two factors were interpreted as indices of perceived control over external events (external events) and perceived control over internal emotional reactions (internal reactions). At this stage in scale development an additional 8 items were removed because they either double-loaded or did not meet the salient loading criteria of 0.30 or greater. Thus, the final version of the AxCQ contains 30 items.

After initial examination of the factor structure in a clinical sample the authors then examined the structure of the AxCQ in a nonclinical sample (Rapee et al., 1996). An exploratory factor analysis (EFA) was performed with this sample, although a two factor solution was sought at the outset. The authors reported good internal consistency and factor loadings similar to those found in the clinical sample. However, the authors state that “because the evidence for the clear existence of a second factor is not strong, we would suggest the use of the scale in most situations as a single measure” (Rapee et al., 1996, p. 289).

The authors conducted further studies to examine test–retest reliability, convergent and divergent validity, as well as sensitivity to change in psychopathology during the course of treatment (Rapee et al., 1996). The AxCQ continued to display good psychometric properties. Gender differences were also examined and the authors found no gender differences in total scale score in either anxiety patients or normal control participants.

Clearly, Rapee et al. (1996) were very thorough in their development of the AxCQ. However, as noted by the authors, the psychometric properties of the scale have yet to be replicated by independent researchers. Replication in support of their findings would strengthen confidence in the questionnaire. Thus, the present study sought to examine the factor structure, as well as other scale characteristics, in an independent sample.

Section snippets

Participants and procedure

A total of 316 students enrolled in introductory psychology classes at Texas Tech University participated in the study. They were tested in large groups during one of six data collection sessions. After elimination of 12 protocols that contained missing and/or incorrect data (i.e. out of range on the Likert scale) and one multivariate outlier, the final sample size was 303 (189 females, 110 males, 4 did not respond). Age was gathered by categories. Thus, the mean and standard deviation cannot

Data suitability

Prior to conducting the factor analyses, the data were screened for normality. 18 of the 30 items showed substantial positive kurtosis and negative skew, indicating that many participants endorsed those items in a manner indicating the perception of control. This was not unexpected given the use of nonclinical subjects. Due to the negative skew, all nonnormal items were reflected and, depending upon the severity of the skew, either a square root or log transformation was then applied. However,

Discussion

These data provide mixed support for the psychometric characteristics of the AxCQ. As noted, the means observed in this sample are comparable to those found in the Rapee et al. (1996) non-clinical student sample and thus provide supporting evidence of a normative score in college-age non-clinical samples. In addition, internal consistency appears quite good for both the full scale and the subscales derived from the various factor structures (Nunnally & Bernstein, 1994). The factor structure,

References (21)

There are more references available in the full text version of this article.

Cited by (26)

  • Beliefs over control and meta-worry interact with the effect of intolerance of uncertainty on worry

    2012, Personality and Individual Differences
    Citation Excerpt :

    Participants respond on a 6-point Likert Scale. We restricted our analysis to the ACQ total score because the psychometric properties of the total score are better than those of the two subscales (Antony et al., 2001; Zebb & Moore, 1999). The total score has demonstrated strong internal consistency and high test-retest reliability as well as the ability to discriminate between anxious and non-anxious individuals (Rapee et al., 1996).

  • Does intolerance of uncertainty predict anticipatory startle responses to uncertain threat?

    2011, International Journal of Psychophysiology
    Citation Excerpt :

    Descriptive statistics for the IUS, PSWQ, and ACQ are presented in Table 1. The means and standard deviations were consistent with previous investigations using the same scales (Buhr and Dugas, 2002, 2006; Carleton et al., 2007; Dugas et al., 2001; Meyer et al., 1990; Rapee et al., 1996; Zebb and Moore, 1999). There was a significant positive correlation between the IUS and PSWQ, r(57) = .64, p < .001, and significant negative correlations between the ACQ and IUS, r(64) = −.53, p < .001, and PSWQ, r(57) = −.59, p < .001.

  • Low perception of control as a cognitive factor of eating disorders. Its independent effects on measures of eating disorders and its interactive effects with perfectionism and self-esteem

    2008, Journal of Behavior Therapy and Experimental Psychiatry
    Citation Excerpt :

    The total score was internally consistent, with high test–retest reliability and the ability to discriminate between anxious and non-anxious individuals (Rapee et al., 1996). In addition, Zebb and Moore (1999) failed to replicate the two-factor structure observed by Rapee et al. (1996). Thus, using the two ACQ subscales would have added complication to statistical analyses and lessened the psychometric qualities of the instrument.

  • Perceived anxiety control as a mediator of the relationship between family stability and adjustment

    2008, Journal of Anxiety Disorders
    Citation Excerpt :

    Higher scores reflect a greater degree of perceived control. The ACQ has internal consistency scores of Cronbach's α = 0.87–0.89 (Rapee et al., 1996; Zebb & Moore, 1999). Test–retest reliability at 1-week and 1-month intervals are r = 0.88 and 0.82, respectively (Rapee et al., 1996).

View all citing articles on Scopus
View full text