An Illustration of how the Effects of Serial Dependencies are Handled in Analyses of Shine's Pure and Actualized Single-Subject Behavior Functions

Abstract

Separate analyses of Shine's pure and actualized single-subject behavior functions are presented for two simple artificially generated data examples, the purpose being to illustrate how the effects of serial dependencies in the data are handled. The first example involves a first order moving averages time-series model. The second example involves a first order autoregressive time-series model. Both examples involve level effects only. In the analyses of the pure behavior function, it is demonstrated how the effects of serial dependencies in the data are carried in the sample autocorrelations of the resulting estimated residual function. In the analyses of the actualized behavior function, it is demonstrated how the effects of serial dependencies in the data are carried as direct effects in the resulting estimated actualized behavior function.