When California districts faced the Common Core adoption window a decade ago, more than half reorganized their high school math sequence from Algebra–Geometry–Algebra II to an integrated model (IM), and even more curtailed access to algebra (or its integrated equivalent) in 8th grade, rolling back a generation of "Algebra for All" policy ("rollback"). This study examines the consequences of those choices, providing quasi-experimental evidence relevant to decisions districts continue to grapple with as they respond to the 2023 California Mathematics Framework.
Districts that adopted integrated math saw small improvements in 11th grade math achievement of roughly 0.037 standard deviations on the state assessment. This finding is sensitive to pre-trend violations and may not be the sole result of the integrated sequence itself: when math-specific growth is separated from concurrent improvements in English Language Arts, the IM-attributable effect becomes small and imprecise. A more defensible interpretation, consistent with the broader literature, is that enthusiastic CCSS adopters saw small, generalized learning gains, of which integrated math adoption may have been one component.
IM-adopting districts also saw slower student progression through the core high school sequence: the cohort-tracked share reaching precalculus and above declined by approximately two percentage points in standalone models. When IM and rollback are estimated jointly (i.e., through TWFE interaction models, CSDID subgroup decompositions, and controls for the other reform) rollback emerges as the primary driver of course-taking declines, while the IM-specific contribution is smaller and less certain. Districts that adopted IM alongside rollback actually saw smaller declines in advanced course-taking than districts that rolled back alone, suggesting IM may be mildly protective of student pacing. Notably, IM-associated pacing changes did not extend to the most advanced courses: calculus, statistics, and AP/IB enrollment were unaffected. Slower pacing also does not explain the test score effect, whatever drove the small improvements likely operated through integrated (or otherwise CCSS-aligned) instruction itself.
The rollback of 8th-grade algebra access also reduced enrollment in precalculus and beyond. Defining a singular “rollback effect” is conceptually and empirically vexed, so impacts are documented across a range of binary cutpoints, a continuous-treatment estimator, and alternative control groups; estimates converge on a roughly −2 to −3 percentage point reduction in advanced course-taking among (i.e., 5 to 8 percent below baseline) districts that sharply limited 8th grade algebra. This effect is identified from a roughly 33 percentage point differential reduction in 8th-grade Math 1 enrollment between treated districts (64% to 10% between 2012 and 2018) and partially-treated comparison districts (67% to 46%). In other words, a roughly 1 percentage point decline in advanced course-taking corresponds to a 10 percentage point decline in 8th-grade algebra access. These effects are concentrated among the small-to-mid-size districts that make up most of the state California districts, suggesting larger districts may be able to buffer the downstream effects of deceleration through alternative pathways. Unlike IM, rollback reduced enrollment in the most advanced courses: calculus enrollment declined by roughly 1.2 percentage points. While modest in absolute terms, this represents roughly a 15 percent decline against baseline calculus enrollment in rollback districts.
In the statewide cohort-tracked data, the share of 11th- and 12th-grade students in advanced math courses fell about five percentage points between the 2015 and 2017 cohorts. About 60 percent of this decline is attributable to the IM and rollback adoption decisions; the remainder likely reflects a combination of broader secular trends and reform effects operating through high-school-only districts that are not directly observed in the analytic sample.
However, slowing the pace at which most students could move past middle-school coursework did not improve 11th grade test scores, and may have even reduced them in districts without the buffer of IM-associated learning. As expected, rollback reduced the share of 9th graders enrolled in Algebra I as repeaters. But it reduced the share of students still completing foundational coursework (Algebra I, Math 1) in 11th grade by less than IM and not at all in 12th grade. Furthermore, neither policy achieved the equity gains in the manner many reform districts intended. To the extent that these reforms narrowed racial gaps in advanced course access, it was through leveling down rather than lifting up: advanced course take-up fell for students from groups typically more represented in advanced courses, while students from underrepresented minority groups (URM) saw no offsetting gains.
These findings carry direct implications for California districts navigating the current California Mathematics Framework adoption window. The full report develops five implications for the current California Mathematics Framework adoption window. Three are highlighted here. First, curricular adoption windows can be powerful catalysts for structural change. If motivated, district administrators are well-positioned during these moments to implement structural changes (i.e., course sequencing, placement thresholds) that directly shape which math students can access and when. This contrasts with the frequently diluted mechanisms of standards-based reform that depend on teacher-initiated changes to classroom practice. The scale and speed of the post-Common Core transformation underscore the leverage these windows carry.
Second, while the rollback of middle-school algebra was motivated by equity and achievement goals, the evidence suggests broad middle-school deceleration does not meaningfully support either goal in the spirit intended. Supporters hoped that deferring early algebra access would narrow attainment disparities by reducing differential acceleration and Algebra I repetition rates, and that prolonging student engagement with pre-algebra material would build a deeper mathematical foundation that paid off later. As this report documents, neither materialized. Deceleration did not improve learning, and where the racial gap in advanced course-taking did narrow, it did so by lowering the ceiling rather than raising the floor. These findings point toward more nuanced approaches (e.g., automatic enrollment) that avoid all-or-nothing placement and reduce reliance on the discretion of traditional gatekeepers.
Finally, the context behind the magnitude of these findings (e.g., the 1.2-point calculus effect amounting to a 14 percent decline from baseline) is a reminder that policy debates over acceleration focus intense attention on the relatively small, though important, group of students who ever reach calculus, AP math, and other advanced courses. If California's goal is to broaden the population of students prepared for advanced coursework while also strengthening the mathematical development of all students — regardless of whether they ever reach these classes — attention must be focused primarily on the greater challenge of improving instructional quality in elementary and middle school grades.
