NONLINEAR-ANALYSIS OF DYNAMIC AND PSYCHOL OGICAL SYSTEMS .1. CONCEPTSAND METHODS

In order to understand and analyze psychological phenomena it is frequ ently necessary to observe their evolutions in time. This is accounted for by many theories which emphasize the dynamical character of psych ological systems. But on the other hand, empirical research is still c entered around cross-sectional studies which yields no information on temporal process. Additionally, if we consider recent theories of self -organizing and chaotic systems we should expect many psychosocial sys tems to be inherently non-linear. Therefore, empirical studies should over both dynamical and non-linear aspects. The present paper introduc es a methodological approach that allows for the investigation of the dynamics and non-linearity of psychological systems. This approach is implemented as a method using surrogate data sets for hypothesis testi ng; we used the goodness of fit of a non-linear forecasting algorithm as a test statistic. In artificial data sets (Henon attractor) we can show that this methodology reliably assesses non-linearity and chaos i n time series even if they are short and contaminated by noise. An app lication to empirical psychological data is provided in part II of thi s work.