This article presents the process of building a dataset for evaluation of the accessibility of 368 web pages, beginning with Webometrics rankings, the WAVE tool was used in the evaluation of the web pages. The dataset documents data on repeated errors with higher frequency, in such a way that they alert the web developers, supporting them in creating more inclusive and accessible websites for all types of people, including users with disabilities. The data show that university websites have frequent problems related to the lack of alternative text linked to images. Some of the university websites included in this dataset were found to violate web accessibility requirements based on the Web Content Accessibility Guidelines 2.0 and 2.1. Therefore, this data has been shared to allow replication of the experiment, and serve as an input to future studies related to web accessibility. The dataset is hosted, with public access, in the Mendeley Dataset Repository.