The aim of this article was to independently test the factor structure of the Early Childhood Environment Rating Scale, 3rd Edition (ECERS-3). Using a sample of 148 independent observations, Standard and Satorra--Bentler confirmatory factor analysis were used to determine if the ECERS-3 conformed to the structure published in its manual using a community sample of urban classroom observations from Rochester, NY. When it was found it did not, exploratory factor analyses of various types, and various number of factors were extracted and compared. In our sample, the ECERS-3 does not conform to the published factor structure. Published subscales have relatively low internal reliability. Exploratory factor analyses suggest that there are three clusters of items (learning activities and environment, interaction, gross motor related) with a number of items failing to load with any construct. Caution should be used when employing the published ECERS-3 subscales for making decisions regarding early childhood programming.