Contradictions in the linear analysis of psychological data for strong dependencies with a maximum

Abstract

Two dependencies have been considered in comparison, which are similar in form, representing dependencies with a maximum of approximately the same values in standard scores or comparative weights, but at the same time completely different in linear Pearson correlation within the accepted rules of its interpretation. These are the dependence of the Sensitivity (SMI-2) variable of the Smishek questionnaire on the Respect for others – self-esteem (25F-10) variable of the five-factor personality questionnaire with a negative (– 0.218) significant correlation and the dependence of the Guilt Feeling (AGR-8) variable of the A. Bass and A. Darki methodology on the Pedantry (SMI-7) variable of the Smishek questionnaire with a positive (0.225) significant correlation. These correlations are significant at the level of p = 0.01, that is, according to the unspoken rules accepted by the majority of the psychological scientific community, they deserve a lot of attention in order to be presented as the study result. With this approach, two strong dependencies with a maximum should be considered as two opposite relationships: one as a direct relationship (increasing dependence), the other as an inverse relationship (decreasing dependence), which qualitatively changes both the interpretation of the results and possible practical recommendations. For clarity, all information has been illustrated with graphical representations of the analyzed dependencies.