Is psychology based on a methodological error?

It is believed a proven fact that variables in social and personality psychology match to normal distribution with its single peak. Multiple peaks are explained by independent variables. However, after a comprehensive data analysis of more than 8.000 patients and on the basis of a bio-psycho-social model with 27 scales, we arrived at the conclusion that normal distribution and the psychometric error theory cannot withstand critical analysis in large samples. Beyond the "truth" that is proved by distribution-dependent statistical inferences, there exists another "truth" that is denied by the empirical doctrine. This "truth" is influenced by compensatory belief systems and explains paradoxes in quality of life research. We hypothesize that items, referred to life risks are micro-stressors, triggering self-regulatory processes as a humanly inherent response, deeply anchored in human evolution. Especially when exposed to threatening experiences, self-focused attention generates amplified multimodal distributions and subverts the methodological premises by an ambivalence-bias between thrill and threat, hopes and fears, pleasure and pain, success and failure, etc. In this article we want to focus attention to the incommensurability between test theoretical axioms and the way people usually respond to self-focused items. We discuss basic distribution patterns and approach to an evolutionary theory of fluctuation of validity.