Abstract
Previous research on sensation seeking (SS) was dominated by a variable-oriented approach indicating that SS level has a linear relation with a host of problem behaviors. Our aim was to provide a person-oriented methodology—a probabilistic clustering—that enables examination of both inter- and intra-individual differences in not only the level, but also in the pattern of SS. We have applied model-based clustering to a four-semester long longitudinal high school survey (N = 3334) and to a cross-sectional university survey (N = 438). The results indicated that impulsive patterns are linked to negative outcomes whereas non-impulsive patterns are associated with positive outcomes. Our study aims to serve as a methodological example on how to apply model-based clustering to examine different types of sensation-seeking patterns. This modern clustering method allows for probabilistic categorization, with continous typicality scores besides cluster membership variables. These typicality scores turned out to have higher temporal stability than simple categorical membership variables did.
Get full access to this article
View all access options for this article.
