AUTHOR : 신택수
INFORMATION : page. 1~28 / 2014 Vol.21 No.3
This study explained characteristics of longitudinal data and analytic tools. It also described how various modeling techniques can be applied to longitudinal study. Among the several longitudinal data methods, the study focused on latent growth modeling (LGM) based analytic approaches. LGM is modeled as a function of an underlying growth process. It also explores effects of specific factors on individual variation in the growth characteristics. In the unconditional analysis, the growth trajectory of mathematics achievement was followed by nonlinear shape (i.e., concave shape). For the analysis of the conditional model, gender differences were found in terms of both initial status and growth. Although female students reported lower initial scores, the growth rate was significantly faster in females than in males. Additionally, low SES students repeatedly reported lower scores across years. In the school level, although significant differences were found on the initial status, the initial status and the growth were not significantly related, suggesting school gap sustained. Lastly, reading ability would have a positive influence on mathematic achievement and the proper number of latent classes was deemed to be 4. Other pertaining issues were also discussed.