Background: Publicly funded four-year-old preschool (or 4K) programs have expanded over the past several decades (Friedman-Krauss et al. 2021). Many high-quality preschool programs lead to gains in academic and social skills for students (Cascio and Schanzenbach 2013; Gormley and Gayer 2005; Weiland and Yoshikawa 2013; Wong et al. 2008), particularly for low-income students (Weiland and Yoshikawa 2013) and for Black students and Hispanic students (Bassok 2010; Gormley and Gayer 2005). However, effects across public programs vary in magnitude (Wong et al. 2008), and not all programs yield positive outcomes (Durkin et al. 2022). Understanding variations in program features is essential to unpacking differences in efficacy. One key dimension by which public preschool programs vary is length of the school day (Friedman-Krauss et al. 2021). Several studies identify positive effects of attending full-day 4K relative to half-day 4K (Atteberry, Bassok, and Wong 2019; Herry, Maltais, and Thompson 2007; Reynolds et al. 2014; Robin, Frede, and Barnett 2006), though others find no differences in outcomes (Leow and Wen 2017; Valenti and Tracey 2009). We add to this body of literature, investigating differences in learning in literacy, numeracy, and executive functioning in full- and half-day 4K classrooms in one medium-sized, urban school district in the Midwest. Question: Are there differences in learning gains at the end of 4K and at the end of five-year-old Kindergarten (5K) for students enrolled in full-day classrooms relative to half-day classrooms? The additional school time should, in theory, allow for additional instructional time, leading full-day attendees to enjoy greater learning gains over the course of the year. If we identify gains in learning, we expect these gains to be sustained through 5K. Setting: The Madison Metropolitan School District (MMSD) is a mid-sized, urban school district in the Midwest. MMSD has offered free, universal half-day 4K since 2011, serving just under 2,000 students per year. In fall of 2021, MMSD introduced free, full-day 4K in eight of its 26 elementary schools. In fall of 2022, MMSD expanded its full-day program to four additional schools. Subjects: This study consists of two cohorts of 4K students. In the Year 1 cohort (2021), we sampled 168 students across 16 full-day and 238 students across 27 half-day classrooms. In Year 2 (2022), we sampled 212 students in 23 full-day and 162 students in 12 half-day classrooms. We also followed a sample of 122 of the 2021 cohort through 5K. (See Table 1 for descriptive characteristics.) Intervention: We investigate differences between students enrolled in full-day, ~7 hours per day, and half-day, ~3 hours per day, 4K classrooms. Evaluation design: As part of an equity strategy to address persistent gaps in outcomes, MMSD chose to initially offer full-day 4K programming at schools that serve relatively high numbers of students of color and students from families that are economically constrained. At the beginning of each school year, we selected schools offering half-day 4K that served similar populations of students as the schools offering full-day 4K. From each selected classroom, we sampled 12 students to participate in the evaluation. We use a difference-in-differences design to measure the effect of full-day relative to half-day 4K on gains in literacy, numeracy, and executive functioning, controlling for student background characteristics. We also conduct a triple-difference-in-differences analysis for literacy, for classrooms with teachers who taught in the same school for at least one year prior. We measure the difference between full- and half-day achievement gains relative to students in the same classrooms in years prior (results forthcoming). Data collection and analysis: In the fall and spring of each school year, a team of graduate assistants assessed students on the Woodcock-Johnson IV-Applied Problems subtest, measuring early numeracy, and on the Head-Toes-Knees-Shoulders (HTKS) assessment, measuring executive function. Teachers also administered the Phonological Awareness Literacy Assessment (PALS), measuring literacy. The latter assessment is given to all students in the district as part of routine practice. For the 2021 cohort, our main analytical model is: where full-day is a dummy indicating whether (1) or not (0) the student was in the full-day class, spring is a dummy indicating whether the assessment was done in the spring (1) or fall (0), and X are covariates, including student sex, race/ethnicity, parent education, free- and reduced-price lunch status, English proficiency, and special education designation. Data collection for Year 2 is ongoing and results are forthcoming. Findings: In Year 1 (2021-22) of the study, we do not find a difference in achievement gains in literacy, numeracy, or executive function between those in full- and half-day 4K classrooms, conditional on student background (Table 2). We do not find heterogeneous effects across subgroups (Table 3). Conclusions: Contrary to expectations, students enrolled in full-day 4K did not see learning gains relative to their peers in half-day classrooms. We offer several plausible explanations for this null results. It is possible that teachers in the full-day classrooms were unable to take full advantage of the instructional time allotted. This could be due to challenges arising from the COVID-19 pandemic or point to the need for additional professional development to support teachers to best utilize time afforded by the extended school day. Tracking the second year of the full-day program, along with following a cohort into 5K, will deepen our understanding of these findings. We acknowledge that our identification strategy is imperfect. While we match as best we can on classroom attributes and control for many student characteristics that might affect a student's achievement growth over the course of the year, unobserved characteristics of teachers and students may affect our ability to compare across these groups. However, we believe our difference-in-differences strategy, and our more exploratory triple-difference analyses, address these concerns to some extent. In the absence of random assignment, we argue that these results offer the best evidence available for effects of full-relative to half-day 4K in contexts like that of MMSD.