Smoothing enhances the detection of common structure from multiple time series.

The effects of smoothing (i.e., temporal averaging) on the detection of intraindividual interdependency from between-subjects aggregated (i.e., multiple) bivariate time series were examined. A simple moving average smoother was applied to different types of simulated processes that included error. The results indicated that smoothing facilitates the detection of common intraindividual structure from multiple time series. The efficacy of smoothing was dependent on the characteristics of the underlying process. It is suggested that smoothing increases the efficiency of the detection of common structure by increasing the signal-to-noise ratio (i.e., temporal reliability) of the time series. The issue and application of smoothing is further discussed in terms of signal extraction and classical test theory.