Errors of correlation method on the example of a specific cause parameter (love for children), her non-linear influence on others professional competencies of the teacher

The article raises the question of studying statistical dependences in psychological research. Psychologists in the vast majority on the basis of experimental data offer only the interpretation of linear dependences. In the last 20-25 years are often studied the complex processes that do not fit into the linear models. And psychologists choose (consciously or not) way of describing weak linear dependencies (where virtually not detected strong dependencies ). To do this, they have used test hypotheses (=0) for the correlation coefficients. Correlations such are indeed many. The article analyzes linear correlations for professional competencies of the teacher. Their number is determined for different intervals of the correlation coefficient. We consider combined in content, statistically interesting, non-linear dependencies. They demonstrate the gross errors of the traditional use of correlation analysis. In all dependencies, the reason (independent parameter) is “Love for children”. It is interesting that the love for children in the analysis of statistical dependencies is revealed only as a cause (independent parameter). Other parameters depend on this parameter. It itself is weakly dependent on other parameters. Statistical dependencies are predominantly one-sided.