Psychometric Evaluation of the Powe Fatalism Inventory.

Abstract

Background and Purpose: Powe conceptually defined "cancer fatalism" and developed the Powe Fatalism Inventory (PFI) to operationalize cancer fatalism. Researchers report disparate underlying factor structures, and sparse evidence supports the validity and reliability of the PFI. Therefore, the purpose of this study was to examine the psychometric properties of the PFI. Specifically, we aimed to examine its (a) underlying dimensions, (b) internal consistency, and (c) construct validity. Methods: We recruited 400 post-menopausal women, 50-64 years old, for a study on mammographic breast density. Women completed the 15-item PFI and the 8-item Champion Breast Cancer Fear Scale (CBCFS). We conducted item analyses and exploratory factor analysis and evaluated different factor structures. We estimated internal consistency and conducted Pearson correlations between PFI and CBCFS scores to examine construct validity. Results: We found a two-factor solution. Factor 1, Predetermination, had an eigenvalue of 5.2 and explained 43% of the variance with factor loadings ranging from -0.59 to -0.83. Factor 2, Pessimism, had an eigenvalue of 4.5 and explained 15.2% of the variance with factor loadings ranging from 0.63 to 0.77. Both factors together explained 58.2% of the variance. There were no cross-loading items and no item loadings below 0.4. The two subscales both had alphas of .89. Cancer fatalism scores were positively related to fear scores (r =317, p < .001, 95% CI: 0.222, 0.406). Conclusion: Using PFI responses from postmenopausal women, we determined that the two-factor solution was the most parsimonious yet theoretically sound factor structure underlying the 15 items of the PFI. The subscales Predetermination (Factor 1; six items) and Pessimism (Factor 2; nine items) were internally consistent with the evidence of the construct validity.