ABSTRACT
The study of positive youth development is inextricably tied to the analysis of longitudinal data. Most researchers working in this field have probably been exposed to methodological questions regarding the design, analysis, and/or adequate interpretation of longitudinal studies. Unfortunately, longitudinal statistical analysis can be full of pitfalls and, as apparent from the history of longitudinal statistical analysis in the social sciences, the literature is peppered with questionable recommendations and myths, resulting in inadequate conclusions, confusion about which methods to use, or even resignation by avoiding longitudinal data altogether (Cronbach & Furby, 1970; Harris, 1963; Rogosa, 1995; Voelkle & Adolf, 2015). Three particularly fundamental issues regard (a) the conception and statistical treatment of time in longitudinal models, (b) the treatment of missing values, such as missing measurement occasions, and (c) the presence of multiple cohorts in longitudinal models of youth development.
