Study finds limited, mostly non‑significant links between school waivers and outcomes; resource waivers linked to small spending uptick

APA consultants updated the Senate Education Committee on their analysis of state waiver policies and school finances, focusing on schools that accessed waivers under the Act 1240 waiver pathway. Amanda Brown said the team compared 2015–16 (pre‑waiver) and 2018–19 (post‑waiver) data for waiver schools against a control group of non‑waiver schools and used regression models to control for baseline differences such as free and reduced‑price lunch (FRL) share, special education, English‑learner status, school type and enrollment.

The consultants reported that most instructional and resource waiver areas produced non‑statistically significant results and generally trended positive. "Most of those instructional waivers were non‑significant and mainly trending in the positive direction," Michaela Tonkin said, summarizing the aggregate results on achievement measures. She added that the majority of findings for individual waiver categories were not statistically significant.

The presentation identified a few specific signals. For resource waivers, the team found an association with about a $613 increase in total expenditures per pupil; one regression result showed about a 1.2‑percentage‑point change in literacy achievement associated with resource waivers. Amanda Brown and Michaela emphasized that those are correlations and urged caution: "Even with statistically significant difference, that doesn't tell us whether the waiver caused the change in outcome," Michaela said.

Committee members questioned data choices and interpretation. Representative De La Rosa asked whether ACT Aspire data were comparable across the years; Amanda Brown replied that ACT Aspire results for 2015–16 are present on the MySchoolInfo site and that the team used the available test data. Senator Chesterfield and others pressed how FRL and other controls were used; Brown said the regression models control for FRL and other demographic and school characteristics, while noting the analysis is at the school level and cannot identify effects at the individual classroom level.

The consultants also noted sample and scope limitations: they excluded conversion and open‑enrollment charters (which had waivers in prior decades) and schedule‑change waivers (very common), and they removed schools with missing data from the regression dataset. Justin Silverstein said the sample of analyzed schools was drawn to be reflective of the state while focusing on the waiver pathway that most districts used.

Next steps: presenters said they will provide the committee with fuller descriptive tables, explain which waiver subcategories were included, and — if the committee wants — run additional subgroup analyses (for example, FRL students in individual waiver categories). The committee requested supplemental materials and clarified that the team should avoid overstating causal claims.