The purpose of this study was to conduct further investigations of the latent structure underlying the Behavioral and Emotional Screening System Teacher Rating Scale-Preschool (BESS TRS-P) by examining latent variable differences due to demographic factors of the preschool children being rated. A U.S. representative sample of teacher ratings of more than 1,200 preschoolers was used to examine the latent structure of the BESS TRS-P. A multiple indicator-multiple causes (MIMIC) model used with a bifactor model showed differences on selected latent factor means relative to a child's age and gender, but no relationship with child's race. Use of age and/or gender scoring norms may produce a more accurate estimate of a child's latent score. (author abstract)