Using Idiographic Methods to Develop a Typology of Alcohol Users

Idiographic methods focus on the time-dependent variation within a single individual (intra-subject variability) in contrast to methods which focus on group-level relationships (inter-subject variability). The method is an alternative to the nomothetic approach, which focuses on group level analysis. An inter-individual analysis may yield different results than an intra-individual analysis. Equivalent results will occur only if the two conditions specified by the Ergodic Theorems are met: (1) Each individual trajectory has to obey the same dynamic laws, and (2) Each individual trajectory must have equal mean levels and serial dependencies. In addition, idiographic methods have several advantages. Idiographic methods can: (1) Address different research questions; (2) Be used in applied settings, such as businesses, schools, clinics, and hospitals; (3) Have important advantages for investigating patterns of change across time: (4) Can be used to determine the generating function for the behavior of interest; and (5) Can address the relationship between variables over time. However, the generalizability of research findings from idiographic studies involves conceptually different models. In this paper, an example will be used to illustrate different aspects of idiographic methods. A typology of alcohol users based on an implementation of idiographic methodology to examine longitudinal patterns of alcohol use will be described. The analysis employed 180 daily observations of alcohol use in a sample of 177 alcohol users. The first order autocorrelations ranged from -.76 to .72 and seventh order autocorrelations ranged from -.27 to .79.  Times series analysis was performed at the individual level and a dynamic cluster analysis was employed  to identify homogenous longitudinal patterns. Eight distinct profiles of alcohol users were identified, each characterized by a unique configuration of first and seventh autoregressive terms and longitudinal trajectories of alcohol use. The findings demonstrate that individuals with very similar levels of alcohol use can have different temporal patterns of drinking.