Abstract
Objective: The purpose of this study is the validation of a proposed additional item for the PWI-7 scale for measuring sexual satisfaction as a dimension of Personal Well-being.
Methodology: An adaptation of the PWI-7 questionnaire was administered to adult inhabitants of urban areas of Santiago, Chile. Analysis consisted of exposition of descriptive statistics, item-scale correlation, item-item correlation, multiple linear regression with the Overall Life Satisfaction Scale (OLS), moderation analysis and, confirmatory factor analysis. All analyses were divided by gender due to significant differences in sexual satisfaction found in literature reviewed.
Results: All items were significantly and positively associated with the PWI. Internal consistency was satisfactory for Cronbach’s alpha (α = 0.884 for males and α = 0.877 for females). After conducting a Confirmatory Factor Analysis using maximum likelihood (ML) as estimator, adequate levels of adjustment were obtained.
Discussion: Results indicate that adding a new item on sexual satisfaction might be a contribution for the measurement of life satisfaction. The scale shows adequate internal consistency once the item is added and the new item on sexual satisfaction has a better fit than the one regarding satisfaction with spiritual life. Factor structure was invariant for males and females.
Notes
1 Kolmogoronov tests on distribution of items were significant, between .208 and .137, all p < .0001, suggesting that answers were not normally distributed and were positively skewed as usual in subjective wellbeing studies. Positively skewed distributions are often achieved especially on quality of life and life satisfaction topic (Bowling, Citation2014). Evidence suggests that treating them as normally distributed, in subsequent analysis, does not lead to significant biases: the practice is indeed common in the analysis of subjective well-being data, and there appear to be few differences between the conclusions of research based on parametric and nonparametric analyses. In fact, Pearson and Spearman correlations showed similar effect sizes (Diener & Tov, Citation2012; Ferrer-i-Carbonell & Frijters, Citation2004).
2 Both Pearson and Spearman correlations were carried out and results (effects sizes) were similar. Because parametrical analyses are robust when assumptions are not fulfilled, Pearson correlations were reported. See criterion validity correlations for comparison of r and rho correlations.