In we write SAP for clinical trials, we often are required to specify a primary analysis (model) to identify the covariates in the model. It is important to build a model with the important covariates included. It is controversial how to evaluate the study center effect and whether it should be included in the model: as fixed effects (with or without interaction term), random effects. This roundtable intends to have a discussion on this topic based on participants' practical experiences. Three questions to discuss: 1. In what situations should we include study center as a covariate in the model? What's the advantages and disadvantages? 2. How should study center be integrated, as fixed or random effects? 3. What's your experience when evaluating study center effect?