One of the most vexing challenges in early childhood developmental research is the “fade out” of preschool benefits: preschool attenders tend to outperform non-attenders on basic literacy and math outcomes at the end of preschool, but scores between the two groups often converge by the end of kindergarten. There are two shortcomings of this research: first, most studies have relied on a relatively narrow band of literacy and math measures to assess preschool impacts, leaving out a broader range of cognitive skills and ignoring the domain of self-regulation entirely. Second, the few studies that have included broader-band outcomes have only followed children to first grade, with one exception that followed children to third grade, but produced a confusing set of mixed findings and did not assess self-regulation. Thus, a complete picture of the potential impacts of preschool attendance remains unknown. The current study tests for sustained preschool benefits in third grade (N=694, Mage at 3rd=8.5 years) on an unprecedented set of directly assessed language and math skills as well as multiple dimensions of self-regulation. In our racially and ethnically diverse sample of low-income children attending the nation’s two largest public preschool programs – Head Start and school-based public pre-k – in Tulsa, OK, preschool attenders outperformed non-attenders on numeracy and self-regulatory skills in third grade. Consistent with prior studies, the preschool advantage on basic literacy disappeared, and the advantage on applied problem-solving and language skills weakened to marginal levels of significance. (author abstract)