Latent Growth Curve Modeling

A fundamental observation in the social and behavioral sciences is that people change over time, but not necessarily in the same way or at the same rate. For example, verbal ability increases steadily throughout the elementary school years, but it does not increase at the same rate for all students. There are individual differences in the rate and direction of change in many contexts, and these individual differences in change are often of scientific or practical interest. Change over time can be measured in seconds, as in studies of cardiac reactivity, or in decades, as in life span development studies. Marital arguments, for example, may cause significant endocrinological change over the course of only a few minutes, but these changes can be quite different for husbands than for wives (Kiecolt-Glaser et al., 1997). Flora and Chassin (2005) examined growth in drug use among adolescents as a function of parent alcoholism, following participants over a number of years through young adulthood. Early approaches to investigating change were very limited in that (a) they focused exclusively either on group-level or on individual-level growth and (b) they addressed only two occasions of measurement, resulting in data too impoverished to allow examination of some of the most basic and interesting hypotheses about change over time.