Incorporating Stability Information into Cross-Sectional Estimates.

Psychological researchers are often in a position to use data collected at a singletime point (cross-sectional data) to make inferences about processes that unfoldover time (longitudinal processes). It has been well documented that regression andcorrelation coefficients based on cross-sectional data are not unbiased estimates oftheir corresponding longitudinal effects. In this paper, we introduce the StabilityInformed Model, that can yield more accurate longitudinal estimates by augmentingcross-sectional data with information about the longitudinal stability of thevariables and the stationarity of the process. We explore via simulation the limitations of this model and discuss potential extensions to the Stability Informed Model.