New Statistical Developments for the analysis of changes in longitudinal studies of medical cohort patients: The particular aspect of Structural Equation Modeling
(2002) Acta Obstet Gynecol — Vol. 81, p. 581-587 (2002)
Files
No attached file found for this publication.
Details
- Authors
Lehert, Philippe 
- Abstract
- (en) A comprehensive review of techniques to study cohorts exposed different types of techniques, including Summary statistics techniques, Multivariate analysis of Variance, Autocorrelation Time series and ARMA techniques, and Simultaneous Equation Models. a) Summary statistics are probably the most simple and evident techniques to assess menopause effects in comparing states between pre-peri and post menopausal phases. b) Although ARMA like time series techniques proved very useful, , their use remains was found limited in this application where only 8 years are available, and where only rend trend tendency is expected. b) Multivariate Technique such as MANOVA are very performant when within time subject follows the experiment. Conclusion: The Partial First Order Instantaneous Markov model is probably the most appropriat e, and is characterized by the simple equation yt=ayt-1 + ?bj xit + e. Thus, Yt is influenced by the last value yt-1 and a combination of variables but only at time T. Th
- Affiliations
Louvain School of Management
Citations
Lehert, P., & Dennerstein, L. (2002). New Statistical Developments for the analysis of changes in longitudinal studies of medical cohort patients: The particular aspect of Structural Equation Modeling. Acta Obstet Gynecol, 81, 581-587. https://hdl.handle.net/2078.5/129521 (Original work published 2002)