Validation of DOE Factorial/Taguchi/Surface Response Models of Mechanical Properties of Synthetic and Natural Fiber Reinforced Epoxy Matrix Hybrid Material.

A validation of the factorial, Taguchi and response surface methodology (RSM) statistical models is developed for the analysis of mechanical tests of hybrid materials, with an epoxy matrix reinforced with natural Chambira fiber and synthetic fibers of glass, carbon and Kevlar. These materials present variability in their properties, so for the validation of the models a research methodology with a quantitative approach based on the statistical process of the design of experiments (DOE) was adopted; for which the sampling is in relation to the design matrix using 90 treatments with three replicates for each of the study variables. The analysis of the models reveals that the greatest pressure is obtained by considering only the source elements that are significant; this is reflected in the increase in the coefficient of determination and in the predictive capacity. The modified factorial model is best suited for the research, since it has an R2 higher than 90% in almost all the evaluated mechanical properties of the material; with respect to the combined optimization of the variables, the model showed an overall contribution of 99.73% and global desirability of 0.7537. These results highlight the effectiveness of the modified factorial model in the analysis of hybrid materials.

Multi-Validity Process and Factor-Invariance. Perceived Self-Efficacy-Scale for the Prevention of Obesity in Preteens.

Given the lack of scales with a robust psychometric assessment of self-efficacy related to obesity in early adolescence, we aimed to obtain an instrument with high-quality validity and reliability items. Nonrandom samples (N = 2371) classified boys (1174, M = 12.83, SD = 0.84) and girls (1197, M = 12.68, SD = 0.78) from Mexico City and some cities of the Mexican Republic with obesity rates near to the national level mean. A multi-validity process and structural invariance analysis using the Perceived Self-efficacy Scale for Obesity Prevention were performed. A two-factor-physical activity and healthy eating-model with high effect-sized values-girls R2 (0.88, p < 0.01) and boys R2 (0.87, p < 0.01)-were obtained. Each factor explained more than half of the variance with high-reliability coefficients in each group and acceptable adjustment rates. The self-efficacy scale proved to have only girls, an invariant factor structure, or a psychometric equivalence between the groups. The obtained scale showed that a two-factor structure is feasible and appropriate, according to the highest quality of validity and reliability.