Serial correlation, or autocorrelation, is defined as the correlation of a variable with itself over successive observations. It often exists when the order of [Page 1353]observations matters, the typical scenario of which is when the same variable is measured on the same participant repeatedly over time. For example, serial correlation is an important issue to consider in any longitudinal designs.
Serial correlation has mainly been considered in multiple regression and time-series models. Multiple regression models are designed for independent observations, where the existence of serial correlation is undesirable. So the main focus in multiple regression is on testing whether serial correlation exists. Conversely, the purpose of time-series analysis is to model the serial correlation to understand the nature of time dependence in the data. The pattern ...
Looks like you do not have access to this content.