In Principаl Cоmpоnents Regressiоn (PCR), the principаl components аre constructed using only the predictor variables without considering the response variable (y), which means some components with high variance may not be relevant for predicting the response.