- Measurements of social constructs that evaluate natural hazard preparedness are important to decrease natural hazard vulnerability. Preparedness reduces natural hazard impacts and human vulnerability. Investment in education and education research contribute to human sustainable development and natural hazard preparedness. Faced with other needs, people typically do not prepare enough to reduce the potential impacts of infrequent but potentially catastrophic natural hazard events. Consequently, human vulnerability increases over time and space. Specifically, natural hazards such as tsunamis and earthquakes (Colombia 1979; Sumatra (Indonesia) 2004), hurricanes (Katrina 2005, U.S.), volcanoes (Nevado del Ruiz 1985, Colombia), floods (Venezuela 1999), and fires (Southern California 2007, U.S.) can exact a tremendous toll on both lives and property (United Nations: International Strategy for Disaster Reduction (ISDR), 2009). Yet, most damage from natural hazards can be mitigated or even avoided in great extent with proper preparedness. Why, then, do so few people take precautions? Furthermore, how can we induce major levels of preparedness? This research postulates a conceptual framework that takes into account the learning theory of Paulo Freire. The main objective of this research is to establish the initial validity and reliability of a new social construct measurement (TOTPI) that is accountable for tracing transformative optimism beliefs that impact behaviors related to tsunami hazard preparedness on the United States Pacific Northwest Coast, using exploratory structural equation models (ESEM). The final SEM model for TOTPI, with maximum likelihood parameter estimator, had a multiple correlation of R=0.543 (R²=0.30) between the predictors Transformative Optimism (T.O.), Resilient Optimism (R.O.), and Fatalistic Optimism (F.O.); and the current tsunami preparedness (criterion), this means that 30% of the current tsunami preparedness variance is explained by the T.O., R.O. and F.O. factors. The Chi-square is 355.26 with 126 degrees of freedom. The Comparative Fit Index (CFI) is 0.92; TLI is 0.91; the Root Mean Square Error of Approximation (RMSEA) is 0.060 with RMSEA 90% C.I. between 0.053 and 0.068; and the Standardized Root Mean Square Residual (SRMR) is 0.051. This appears to be a good fit to the observed data according to Hu & Bentler (1999) as cited in Kline (2005, p. 140), and according to Browne & Cudek (1993, p.144) as cited in Brown (2006, p. 87).