The Relationships Among Physiological Measures of Anxiety and Psychological Measures of Anxiety: A Multivariate, Naturalistic Approach

Abstract

An examination of the hypothesis the that a significant relationships would be found between a set of four commonly employed psychological instruments for assessing anxiety and a set of seven commonly employed physiological measures of anxiety was carried out. The research was done in naturalistic setting to improve on the typical labor-story anxiety research that has, of late, been highly criticized and analyzed with a multivariate approach as advocated by many critics of traditional bi-variate research. The methodological approach of naturalistic research was operationalized by employing 54 patients suffering a first myocardial infarction admitted to a large metropolitan hospital. Statistically significant intercorrelations were found among the four psychological instruments employed: Bending short form of the Taylor Manifest Anxiety Scale (TMAS), and abbreviated versions of the State-Trait Anxiety Inventory (STAI), Multiple Affect Adjective Check List (MAACL), and Mood Adjective Check List (MACL). Two cut of the twenty-eight pair-wise correlations between these four instruments and the seven physiological measures averaged over a 24 hour period (heart rate, systolic and diastolic blood pressure, epinephrine, norepinephrine, vanillymandelic acid, and urine output) were significant; but, in general, a lot, negative pattern of pair-wise correlations was found between the two sets with the physiological set also demonstrating insignificant pair-wise intercorrelations among the seven physiological measures of anxiety. The multivariate technique of canonical correlation (external factor analysis) revealed no significant linear overlap between the psychological and physiological measures of anxiety. A principal components analysis (internal factor analysis) revealed three factors accounting for 68 percent of the total variance. All four psychological instruments showed high factor loadings, -.94 to -.80, with communalities from .93 to .66, on the first group factor which was clearly separated from the other two. The seven physiological measures formed the other two factors which demonstrated some degree of overlap between them. These results were interpreted as indicating that, for the sample employed, there was clearly a difference between what is being assessed by the psychological instruments proporting to measure anxiety and physiological measures commonly employed for assessing the same emotion. This interpretation was discussed in terms of Schacter’s theory of emotion; and implications for further research were discussed.