SMABS 2004 Jena University
SMABS 2004 Home Organization About Jena Sponsors Links Imprint / Contact SMABS Home

European Association of Methodology

Department of methodology and evaluation research

Jena University

Contributions: Abstract

Point process models for intensive longitudinal data in the Human Sciences

Stephen Rathbun
Penn State University

Human behavior may in part be characterized by repeated events that occur at discrete points in time. Examples include the lighting of cigarettes by smokers (Shiffman et al., 2002), acts of aggression (Jones et al., 1992), and misbehavior of children (Hill, 1973). The times at which such events occur forms a temporal point pattern, which is assumed to be realized from a temporal point process. Point process models are routinely used in seismology, studies of ecological phenomena and in other areas of natural sciences in which discrete events produce temporal or spatial point patterns.

This paper will introduce a number of point process models and demonstrate how they can be adapted to investigations of human behavior. In particular, these models will be illustrated using data from a study smoking behavior that utilized an Ecological Momentary Assessment, a technology enabled approach to collection of diaries using random prompts for mood self-reports and count entries for participants' substance use events. For example, in homogeneous Poisson processes are used describe the effects of day of the week, time of day, and temporally varying covariates on the intensity of smoking rates.

The random effects of individual variation are described through a Cox process model. Under a stress-release model, borrowed from seismology, the smoking of a cigarette releases stress, decreasing te likelihood that another cigarette will be lit soon after a cigarette is smoked. As time passes, however, stress increases, increasing the likelihood that another cigarette will be lit. This is consistent with a number of regulatory theories of nicotine dependence (see Velicer & Fava, 2003, for a brief review).