Fixed Effects Regression Methods for Longitudinal Data Using SAS

6.6. Summary and Conclusion

Linear, fixed effects or random effects regression models for quantitative response variables can be estimated with PROC CALIS to yield the same results as those obtained using other SAS procedures as described in chapter 2. This method requires a different data structure, however, with one record containing all the measurements for each individual or cluster, and with the multiple measurements coded as distinct variables. In PROC CALIS, a separate equation is specified for each response variable at each point in time, and the coefficients are typically constrained to be the same across equations. The random or fixed effect is specified as a latent variable that is common to all the equations. In the fixed effects ...