The Effect of School Funding on Opportunity Gaps for Students With Disabilities: Policy and Context in a Diverse Urban District

Abstract

School finance reform has recently centered on providing schools with more equitable access to resources to reduce opportunity gaps for students. Although special education is often a prominent part of larger equity conversations, special education funding is commonly excluded from school funding reform initiatives. Given the costly nature of special education programs, it is imperative that scholars and policy makers understand the effects of funding changes on outcomes for these students. In this study, we examined the effect of California’s Local Control Funding Formula, in addition to school context and student compositional characteristics, to identify changes in special education students’ achievement rates. Using a combination of publicly available data sources and local district data, we assessed differences in academic outcomes (i.e., achievement scores) between elementary students with and without disabilities in both high- and low-poverty schools, given increases in spending for special education programs.

Keywords

school finance reform special education policy school funding opportunity gaps equity

Often, educational inequalities are described through the language of achievement gaps, highlighting disparities that manifest through individual outcomes (e.g., test scores). This framework problematically links educational disparities to individual learning capacities. Conversely, recent research has identified how schools, and society more generally, provide students with adequate or inadequate opportunities to learn, which influences academic outcomes and manifests in consequential educational opportunity gaps (Milner, 2013; Tate, 2008). Although school policy reform has recently centered on reducing inequities in academic outcomes, enduring income and racial segregation among neighborhoods and schools has increased in the past few decades (Orfield et al., 2016). This has meant a greater likelihood of differential access to resources, ensuring gaps in opportunity for many students (Darling-Hammond, 2015), and that these gaps vary substantially by geography (Owens, 2018; Reardon et al., 2019).

Increased and flexible school funding has emerged as a potential, albeit controversial (e.g., Coleman et al., 1966; Hanushek et al., 1996), solution for bridging opportunity gaps as schools with increased resources can hire and retain quality teachers and school leaders, offer more comprehensive curriculum and extracurricular materials, and maintain higher quality facilities (Bischoff & Owens, 2019). Studies show that gaps in test score proficiency between students from different socioeconomic strata are smaller in states and districts where funding reform initiatives based on equity have taken root (Card & Payne, 2002; Lafortune et al., 2018), and that policies aiming to equalize financial resources between students with high- and low-income backgrounds have thus far been effective (Bischoff & Owens, 2019). School finance reform, constructed through equity-based funding formulas such as California’s Local Control Funding Formula (LCFF), has the potential to address some inequities that exist in urban schools (Vasquez Heilig et al., 2017).

The LCFF was developed, in part, to address the needs of students with greater academic needs, assuming they could benefit from additional financial supports, and to provide flexibility to local education agencies in determining how to provide updated curriculum and access to high-quality teachers (Humphrey et al., 2017). In this way, the funding policy change might produce meaningful impacts on the persistent racial and class stratification that exists across schools and within society, and, in turn, affect how schools and districts operate and ultimately shape educational opportunities and subsequent outcomes (Carter & Welner, 2013).

Special Education Funding

Special education is a prominent part of the equity debate in education, as the Education for All Handicapped Children Act was passed in 1975 and special education enrollment and funding have consistently increased ever since. For example, in the 1970s, just below 6% of all public school students were identified as having disabilities (National Center for Education Statistics, 1993), but between 2011 and 2018, the number of students receiving services under Individuals With Disabilities Education Act (IDEA) increased from 6.4 million (13% of total public school enrollment) to 7.1 million (14%; Hussar et al., 2020). Although tracking of per-pupil spending on students with disabilities is difficult—most states do not require districts to report on expenditures for special education services (Kolbe, 2019)—Griffith (2015) notes that data from 1999 to 2000 showed average expenditures for a general education student at US$6,556 compared with US$12,474 for a student with a disability, and more current data place the average annual cost of educating a student without disabilities at US$9,000, compared with US$26,000 for a student with a disability (Petek, 2019). California spent more than US$12 billion on special education in the 2014–2015 school year, with only a small slice of the required funds coming from the federal government (Warren & Hill, 2018).

Despite the substantial and increasing costs associated with special education, research indicates that students placed in special education often experience increased risk of negative long-term outcomes such as higher dropout rates (e.g., Thurlow & Johnson, 2011), reduced likelihood of completing postsecondary education (e.g., Newman et al., 2011), higher rates of incarceration (e.g., Bronson et al., 2015), and reduced employment rates (e.g., Newman et al., 2011), compared with peers not placed in special education. Moreover, the quality of special education services that students receive are often contingent upon resources available in the district (Warren & Hill, 2018). Thus, students with disabilities are likely subjected to inequities in addition to those caused by racial and income segregation.

Researchers have emphasized the need to better understand the “contextually dependent nature” (Sullivan & Bal, 2013, p. 491) of special education quality and state and local policies related to services (Kramarczuk Voulgarides, 2018). Hibel et al. (2010) assert that scholars “should investigate the organizational structure of low SES [socioeconomic status] and low achievement schools and how these organizational structures contribute to service delivery” (p. 328). Changes in local funding formulas provide a unique opportunity to evaluate potential changes in learning opportunities and subsequent changes in academic achievement. We have limited information on the ways in which varying resources available to schools and districts may contribute to special education service delivery and consequent achievement measure outcomes—and thus, opportunity gaps—between affluent districts and their less-resourced counterparts.

Purpose of the Study

In 2013, California passed the LCFF to address resource inequities for K–12 schools by allocating budgets based on how many “disadvantaged” students each district serves (Warren & Hill, 2018, p. 31). The LCFF provides districts with extra funding for Target Student Population (TSP; that is, English language learners, low-income students, foster and at-risk youth, as outlined in the LCFF). This flexible spending, however, did not extend to special education. In the state of California, special education funds are distributed based on average daily attendance (ADA) for all students in Special Education Local Planning Areas (SELPAs) to avoid creating a financial incentive for overidentifying students as disabled (Warren & Hill, 2018). Given that special education represents the largest categorical portion of California’s K–12 education budget (Hill et al., 2016), and that special education is an area for which many districts are not showing improved outcomes (Fuchs et al., 2018), understanding local trends in special education funding can contribute to current knowledge about how dedicated funding can reduce opportunity gaps for already marginalized students.

Because students with lower socioeconomic status are more likely to be placed into special education than their more affluent peers (Kincaid & Sullivan, 2017), many students with disabilities are already considered high-need by the LCFF and generate extra per-pupil funding; however, these funds are not meant to specifically support students with disabilities (Warren & Hill, 2018). As special education funding is not directly targeted toward high-need students in high-poverty schools, it is likely that programs serving economically disadvantaged students in special education are underfunded, creating consequential opportunity gaps for students with disabilities being served in economically disadvantaged locations.

This study uses administrative data from the Los Angeles Unified School District (LAUSD) to address the following research questions:

Research Question 1 (RQ1): How did funding per pupil on special education programs vary by the proportion of TSP students served in LAUSD elementary schools during the initial 5 years of implementation of the LCFF?

Research Question 2 (RQ2): How do changes in academic achievement measures of students placed into special education differ by school poverty levels?

Research Question 3 (RQ3): What were the changes, if any, in academic achievement measures between elementary students with and without special education services as spending for special education programs increased in LAUSD?

Research Question 4 (RQ4): How does the impact of expenditure increases on student achievement differ according to the overall availability of fully credentialed special education teachers within schools?

We used a two-stage least squares regression model to examine the impact of state-level policy changes on student achievement measures for students receiving special education services and those not receiving services situated in low- and high-poverty schools. We also examined the interaction between special education spending and the percentage of TSP students within the school, and the interaction between special education spending and the percent of special education teachers with a full credential in the school, to understand the ways in which these crucial funding differences mattered in terms of student outcomes.

Method

Data

Total school site spending

Our primary source for school-level finance data was the LAUSD school spending report for each of the 5 years following the enactment of the LCFF. This report provides spending amounts by major program group and it comprises school-level resources that support all operations (e.g., teacher salaries and instruction-specific costs). School-level spending reports exclude all sources held by the district’s central office (e.g., capital funds, debt service funds, and internal service funds). We considered school site resources (on average, 88% of LAUSD’s unrestricted total budget expenditures) as a critical mechanism for delivering improved opportunities to students.

Instructional and special education spending

We extracted the direct school-level instructional spending amount from within total school budgets. Our composite of all instructional elements included teaching positions (e.g., certificated teachers, long-term substitutes), specific instructional programs and services (e.g., special education staff, tutors for English learners, and support staff for college preparatory courses), and all other support for instructional operations (e.g., books, supplies, and extra resources to reduce class size). These represent funds meant to directly provide instructional supports and thus learning opportunities.

Targeted student population fund

Created in 2014–2015, this designated fund aimed to help meet the state proportionality requirement of increased support for weighted (TSP) students, proportional to the new revenue these pupils generate for the district. The TSP fund was created in direct response to the inflow of concentration grants from the state and is exogenous to collateral policy change. The fund funneled US$541 million directly to schools in 2017–2018 and represented the third largest contributor to total instructional spending, supporting more than 45 initiatives, including expanding college preparatory course offerings and instructional services designed for English language learners.

School context and student outcomes

We combined the school-level spending data with two school-level outcomes: credentialed teachers employed and pupil achievement (and gaps) on state tests. We compiled these data from the California Longitudinal Pupil Achievement Data System (CALPADS; California Department of Education, 2019), which provides annual data on teacher demographics and standardized test scores by student subgroup. Excluding charter and nontraditional schools, our data set included school-by-year panel data spanning from 2013–2014 through 2017–2018, representing 2,275 school-by-year observations of 455 traditional public elementary schools in LAUSD.

Variables

The primary dependent variable for this study was a continuous indicator of the percentage of students who met or exceeded achievement standards on the Smarter Balanced state assessments, in both math and English language arts (ELA), retrieved from CALPADS. Our key explanatory variable, special education spending, was generated from the aforementioned school-level instructional spending data. We extracted all elements related to special education programs and services among the instructional spending items and generated the school-level expenditures to support special education programs within schools, all held in constant 2016 dollars. We included primary school explanatory variables, including a continuous predictor for aggregated percentage race/ethnicity for each school’s student population; school-aggregated percentage of teachers with full, district intern, or university intern credentials; and percentage of special education teachers with full, district intern, or university intern credentials.

Analysis

A concern in estimating the effects of school spending on student achievement is that school spending may endogenously stem from potential confounding factors. For example, if urban income segregation leads to a higher number of underprivileged students in lower quality schools, and finance initiatives award more funding to schools with such students, the estimated effects may stem from unobserved student or family attributes, rather than from the treatment of interest. Thus, we aimed to identify the exogenous portion of variation in school spending produced solely by the school finance policy.

We relied on the exogeneity of special education spending reflected in California Assembly Bill 602’s (AB 602, 1997; California Department of Education, 2018) funding formula that governs funding to SELPAs. The formula provides between US$500 and US$600 in state special education funds per student in each district, based on each SELPA’s ADA. The number of students with disabilities in a district does not affect state funding amounts; the amount is affected mainly by the change in ADA and by a cost of living adjustment. Importantly, the AB 602 funding formula does not incentivize districts to classify more students with disabilities, however, this also means that the formula does not adequately address cost differences incurred with special education services between districts with high and low proportions of students with disabilities. We constructed a simulated instrumental variable, defined as the policy-intended amount of special education resources (subsequently referred to as dosage), animated by the AB 602 funding formula. The funding formula-based changes in school-level special education spending are assumed to be conditionally exogenous to changes in unobserved socioeconomic factors conditional on the change in the level of ADA in each school.

To examine the effect of policy-induced special education funding increases on changes in our outcome variables—based on the dosage—we build on Lee and Fuller’s (2020) estimation technique by using the following equations required for a two-stage least squares regression model:

\begin{matrix} First stage : \ln {(P P E)}_{s t} = α + β . D o s a g e_{s t} \\ + {\tilde{X}'}_{s t} Γ + μ_{s} + λ_{t} + ε_{s t}, \end{matrix}

Second stage : Y_{s t} = θ + δ . \ln {(\hat{P P E})}_{s t} + {\tilde{X}'}_{s t} Π + ρ_{s} + τ_{t} + υ_{s t} .

where our endogenous treatment of interest, $\ln {(P P E)}_{s t}$ is the natural log of per-pupil special education spending for elementary school s and year t. We include the key variable $D o s a g e_{s t}$ to contain the source of exogeneity, and this variable represents the policy-intended amount of special education funding for elementary school s in year t, generated from the AB 602 funding formula. ${\tilde{X}}_{s t}$ is a matrix that includes mean-centered school-level time-varying confounders (i.e., logged student enrollment, percentage of TSP students, and pupil racial composition). We included school fixed effects $μ_{s}, ρ_{s}$ and year fixed effects $λ_{t}, τ_{t}$ in both equations to account for any underlying differences across schools and years and to exploit only variation within school-by-year cells, and we included $ε_{s t}$ and $υ_{s t}$ as random error terms.

The first-stage regression shows how the state-level special education funding formula altered the level of special education spending at the school level. In the second-stage regression, we used only the portion of school-level special education spending increases that can be explained by the policy-intended funding generated by the funding formula. We were primarily interested in estimating the regression coefficient δ of the instrumented per-pupil school spending $\ln {(P P E)}_{s t}$ . This represents the effect of policy-induced increases in school special education spending per student on the change in outcome $Y_{s t}$ after controlling for unobserved time-invariant school and year fixed effects and observed time-varying school effects.

This strategy depends on the critical identification assumption that the simulated instrument $D o s a g e_{s t}$ only impacts the outcome $Y_{s t}$ through the treatment $\ln {(P P E)}_{s t}$ . In other words, there is no direct path from instrument to outcome, except through treatment (i.e., exclusion restriction). One might argue that a direct path operates between the policy-intended amount, $D o s a g e_{s t}$ , to the school-level organizational changes as $D o s a g e_{s t}$ is a function of school enrollment. To address this, we included ${\tilde{X}}_{s t}$ (i.e., school-level time-varying covariates, such as logged student enrollment, percentage of TSP students, and racial composition of schools) in the first- and second-stage models. This accounted for their potential direct impact on school-level outcomes, thus controlling for other processes that might have otherwise linked the $D o s a g e_{s t}$ directly to the outcomes. In addition, inclusion of school and year fixed effects eliminated additional potential bias, such as time-invariant differences in school quality and any business-cycle-related statewide changes.

To discern differences in the effect of LCFF-related school spending in high-need schools compared with low-need schools, and in schools with and without high shares of fully credentialed special education teachers, we included interaction terms between $\ln {(P P E)}_{s t}$ and school disadvantage level $U P P_{s t}$ and between $\ln {(P P E)}_{s t}$ and the share of special education teachers with full credential $F u l l C r e d_{s t}$ , tested in the second-stage model. However, because the interaction between an endogenous predictor ( $\ln {(P P E)}_{s t}$ ) and an exogenous covariate ( $U P P_{s t}$ ) was itself potentially endogenous at this second stage, we fitted two first-stage models, one for each endogenous second-stage predictor: $\ln {(P P E)}_{s t}$ and $\ln {(P P E)}_{s t} U P P_{s t}$ . We regressed each endogenous predictor on the full set of instruments and covariates, $D o s a g e_{s t}$ and $D o s a g e_{s t} U P P_{s t}$ , and we only included significant interaction terms in the final second-stage model.

Results

School Characteristics

We began by grouping schools into poverty-level quartiles to examine the relationship between student academic proficiency and school poverty levels by subgroup. Table 1 reports descriptive statistics for student characteristics and teacher composition variables, including means for school enrollment, racial composition, and percentage of students designated as TSP, between the first quartile of schools (n = 114), enrolling the lowest shares of TSP students (e.g., the lowest-poverty schools), and the fourth quartile of schools (n = 113), enrolling the highest shares of TSP students (e.g., the highest-poverty schools). There were important differences between the lowest-poverty quartile of schools and highest-poverty quartile of schools in LAUSD; the lowest-poverty quartile enrolled an average of 60.9% of TSP students, whereas the highest-poverty quartile enrolled an average of 96.8% of TSP students. In low-poverty schools, 47.5% of all students were Latino, 12.4% were African American, and 24.2% were White. In contrast, in high-poverty schools, 88.9% of students were Latino, 7.8% were African American, and only 1.6% were White, highlighting the critical need for understanding school context related to funding and related outcomes. Table 1 also indicates that high-poverty schools relied more on special education teachers with district intern or university intern credentials, rather than fully credentialed teachers. The percentage of special education teachers with a district or university intern credential (compared with a full credential) was roughly 2 times higher in high-poverty schools (11.3% and 6.4%, respectively) than in low-poverty schools (5.9% and 3.0% respectively).

Table 1.

Descriptive Statistics for Student Characteristics and School Features by Low- and High-Poverty Quartile Elementary Schools in LAUSD, Pooled Data, 2013–2017.

Category	Variable	Overall	Mean by subgroup
Category	Variable	M	Q1	Q4
Student characteristics	Enrollment	552.2	462.7	634.9
	The unduplicated pupil percentage of TSP	85.3	60.9	96.8
	% White	8.3	24.2	1.6
	% Hispanic or Latino	75.8	47.5	88.9
	% African American	8.8	12.4	7.8
	% Asian	9.4	12.8	6.6
Teacher staffing	All teachers
	% Teachers with full credential	97.8	98.7	97.1
	% Teachers with district intern credential	1.4	1.0	1.8
	% Teachers with university intern credential	0.9	0.5	1.1
	Special education teachers
	% Teachers with full credential	89.6	93.7	86.1
	% Teachers with district intern credential	8.1	5.9	11.3
	% Teachers with university intern credential	4.9	3.0	6.4
Number of	Schools	455	114	113
Number of	Observations (school by year panel)	2,275	570	565

Note. The unduplicated pupil percentage of TSP includes students with free or reduced-price meal eligibility, English learners, and foster youth. Data are from the California Longitudinal Pupil Achievement Data System. Highest- and lowest-poverty schools are those in the top and bottom quartiles of school-level distributions of 5-year mean school-aggregated unduplicated TSP pupil counts, respectively. LAUSD = Los Angeles Unified School District; TSP = targeted student population; Q1 = Quartile 1 (lowest-poverty quartile); Q4 = Quartile 4 (highest-poverty quartile).

Per-Pupil Special Education Spending

We examined how special education funding per pupil varied by proportion of TSP students during the first 5 years of LCFF implementation. Figure 1 depicts TSP instructional spending from fiscal year 2013 to 2017, and Figure 2 depicts instructional spending on special education; these figures include only categories of school-level spending directly related to instruction (for comparison, total spending is depicted in the online supplemental Figure 1). Of the total instructional spending, US$1,567 million was allocated to LAUSD public schools in 2016–2017 for special education spending. We found that total instructional, TSP, and special education spending increased over the study period, with progressive targeting of high-TSP schools. The almost equal and persistent differences in TSP spending between low-, middle-, and high-poverty schools over 4 years suggest that these dollars were distributed according to the percentage of disadvantaged students among schools. However, the per-pupil spending level for special education programs in high-poverty elementary schools was similar to low-poverty schools over 5 years, likely because special education funding is based on ADA, and not the share of high-need students. The provision of special education spending, in this manner, violates the LCFF’s equity principle.

Figure 1.

Average TSP program per-pupil elementary school spending from 2013–2014 to 2017–2018.

Figure 2.

Average special education per-pupil elementary school spending from 2013–2014 to 2017–2018.

Furthermore, as noted in Table 1, schools in the highest-poverty quartile employed more special education teachers with intern credentials, rather than fully credentialed teachers, compared with schools in the lowest-poverty quartile. Figures 3 and 4 depict trends over time for these groups and, notably, changes in special education teacher staffing for LAUSD elementary schools from 2013–2014 to 2017–2018 indicate that both quartiles experienced an increase in intern credentialed special education teachers, but that the highest-poverty quartile experienced a sharp reduction in percentage of fully credentialed special education teachers over time, suggesting that both quartiles of schools were grappling with special education teacher shortages (Sutcher et al., 2016), but that more affluent schools were able to retain fully credentialed teachers at a higher rate.

Figure 3.

Change in special education teacher staffing: Credentialed teachers from 2013–2014 to 2017–2018.

Figure 4.

Change in special education teacher staffing: District intern teachers from 2013–2014 to 2017–2018.

Budget, Achievement, and Gaps Between Groups

Although academic performance was higher in low-poverty schools for every subgroup in both math and ELA, Table 2 shows substantially larger achievement gaps between subgroups within the lowest-poverty quartile compared with the highest-poverty quartile schools, considering the percentage of students who met or exceeded standards in both subjects. The math achievement gap between students with and without disabilities was 19.2 percentage points in high-poverty schools, but 30.2 percentage points in low-poverty schools; the gap for ELA was similar: 24.2 percentage points in high-poverty compared with 37.8 percentage points in low-poverty schools. This pattern holds in comparisons between English language learners and students proficient in English, as well as students who were economically disadvantaged compared with those not labeled as disadvantaged although the mean gap difference was lowest between students who were economically disadvantaged compared with those who were not disadvantaged.

Table 2.

Descriptive Statistics for Smarter Balanced State Assessment Results, Percentage Met or Exceeded Standard by Student Subgroups, LAUSD Elementary Schools, 2013–2017.

Subject	% Met or exceeded achievement standard	Overall	Mean by subgroup
Subject	% Met or exceeded achievement standard	M	Q1	Q4
English language arts	Student with no reported disability	38.6	57.2	28.0
	Student with disability	8.9	19.0	4.0
	Gap: No disability–disability	29.5	37.8	24.2
	Fluent English proficient/English only	43.0	57.0	36.2
	English learner	5.1	9.3	3.4
	Gap: Fluent English proficient–English learner	36.0	45.2	32.8
	Not economically disadvantaged	45.4	64.5	28.3
	Economically disadvantaged	32.2	44.6	24.8
	Gap: Not disadvantaged–disadvantaged	12.2	19.8	3.6
Mathematics	Student with no reported disability	32.0	48.8	23.8
	Student with disability	9.0	18.0	4.7
	Gap by disability status: No disability–disability	22.7	30.2	19.2
	Fluent English proficient/English only	35.0	48.5	29.8
	English learner	7.9	14.2	5.7
	Gap: Fluent English proficient–English learner	25.5	31.9	24.0
	Not economically disadvantaged	37.5	56.5	22.0
	Economically disadvantaged	26.5	36.9	21.4
	Gap: Not disadvantaged–disadvantaged	10.3	19.6	0.5

Note. Percentages of students meeting or exceeding state standards are averaged over the 4 years for which comparable data are available. LAUSD = Los Angeles Unified School District; TSP = targeted student population; Q1 = Quartile 1 (lowest-poverty quartile); Q4 = Quartile 4 (highest-poverty quartile), or those in the top and bottom quartiles of school-level distributions of 5-year mean school-aggregated unduplicated TSP pupil counts, respectively.

As aforementioned, there were gaps in academic performance between students classified with a disability compared with nonclassified peers in low- versus high-poverty schools. Specifically, not more than 5% of students with disabilities in high-poverty schools met or exceeded state standards on the ELA assessment, whereas 19% of students with disabilities in low-poverty schools met or exceeded state ELA standards. Students with no reported disability performed better in low-poverty schools (57.2% meeting or exceeding standards), compared with peers attending high-poverty schools (28.0%).

These results indicate that, on average, more students with disabilities in high-income schools performed at a proficient level on state standards than similar students in low-income schools. Student performance on state tests rose overall during the 5-year period, but patterns differed greatly by student and school demographics. This percentage of students meeting or exceeding the state standard for ELA increased for pupils without disabilities in both low-poverty schools (from 50.6% to 62.7%), and high-poverty schools (from 22.1% to 33.6%; see Figure 5). Students with disabilities, however—already performing lower than students without disabilities—scored lower across time, especially in high-poverty schools.

Figure 5.

Change in ELA smarter balanced state assessment results by disability status.

The number of students with disabilities meeting or exceeding ELA and math standards had less gain on average than students without disabilities regardless of poverty level. Specifically, there was a 12.1 percentage point increase for students without disabilities in ELA versus 7.8 among students with disabilities in the lowest-poverty schools, and an 11.5 percentage point gain among students without disabilities versus 2.8 meeting or exceeding state benchmarks in high-poverty schools. Trends in math performance followed a similar pattern; there was a larger gain in the percentage of students without disabilities meeting or exceeding state standards in math (10.5% in low-poverty and 8.8% in high-poverty schools), but less gain for students with disabilities (7.1% in low-poverty and 1.6 in high-poverty schools; see Figure 6). Trends in ELA and math scores by socioeconomic status can be found in Supplemental Figures 2 and 3 for comparison.

Figure 6.

Change in math smarter balanced state assessment results by disability status.

Funding Increases, School Variables, and Achievement Gaps

As shown in Table 3, we found that gains in special education spending significantly predicted higher achievement levels overall, but results varied by poverty levels. The model indicates that an estimated 10% increase in special education spending predicted a 1.87 (p < .001) percentage point increase in students with no reported disability meeting or exceeding standards in ELA when the percent of TSP students was set to the overall average (85.3%). The spending increase did not show any significant effects for students with disabilities in schools with average poverty levels.

Table 3.

Summary of Estimated Effects of Funding Increase on Smarter Balanced State Assessment Results, Percentage Met or Exceeded Standard, by Student Subgroups, 2013–2017.

Treatment variable (logged special educationspending) and its interaction with other variables	% of students who met or exceeded achievement standard
	ELA	Math
Subgroup: Student with no reported disability
Special ed. spending	18.67*** (7.24)	10.50** (4.59)
Special ed. spending × % TSP students	0.45*** (0.16)	0.36 (0.23)
Special ed. spending	21.03*** (8.22)	13.72** (6.45)
Special ed. spending × % Special ed. teachers with full credential	−0.52 (0.33)	−0.47 (0.26)
Subgroup: Student with disability
Special ed. spending	6.38 (3.78)	2.31 (4.26)
Special ed. spending × % TSP students	−0.36*** (0.06)	−0.30*** (0.06)
Special ed. spending	8.65 (4.42)	5.89 (3.56)
Special ed. spending × % Special ed. teachers with full credential	0.52*** (0.07)	0.49*** (0.08)
Gap variable: Student with no reported disability—Student with disability
Special ed. spending	20.95*** (8.44)	20.56* (10.05)
Special ed. spending × % TSP students	0.34*** (0.11)	0.33** (0.14)
Special ed. spending	17.93** (8.24)	18.55* (9.12)
Special ed. spending × % Special ed. teachers with full credential	−0.29** (0.13)	−0.42*** (0.15)

Note. School and year fixed effects, and time-varying covariates for $X_{t i}$ , not reported in the final table. Time-varying covariates $X_{t i}$ included in the model are logged student enrollment, percentage of TSP students, and percentages of White, Latino, African American, and Asian students. TSP = Target student population; ELA = English language arts.

.05 < p ≤ .10. **.01 < p ≤ .05. ***p ≤ .01.

Significant interaction effects of special education spending with the percentage of TSP students and percentage of credentialed special education teachers suggest that the effect of spending increases would be significantly strengthened in low-poverty schools with high shares of credentialed special education teachers, potentially reducing the large gap between subgroups. For example, in schools with a 10% increase in fully credentialed special education teachers, compared with schools with an average share of fully credentialed teachers, a 10% spending increase in special education programs resulted in an estimated 1.38 percentage point increase in students with disabilities meeting or exceeding ELA standards.

Discussion

In this study, we examined a school funding formula meant to address opportunity gaps within complex and influential school contexts. Our results provide evidence that although enduring income and racial segregation among neighborhoods and schools has increased (Orfield et al., 2016), opportunity gaps are malleable, and macro-level policy has the potential to impact structural and interactional components that can affect student outcomes. When LAUSD schools increased spending on special education programs, there was an associated increase in students with disabilities meeting or exceeding ELA standards. We also found that among students without disabilities, the percentage meeting or exceeding the state standard for ELA increased in both low-poverty schools and high-poverty schools with concurrent increases in spending for special education. However, these promising trends were not evident among students with disabilities in high-poverty schools as these students received access to fewer certificated special education teachers than their affluent counterparts. Our findings indicate that high-poverty schools were impacted by high turnover rates of special education teachers, thus reducing the likelihood that students with disabilities who live in poverty will be taught by highly qualified special educators (Darling-Hammond, 2015), and that the organizational context of high-poverty schools is likely to impact service delivery for students with disabilities.

Research consistently shows that students with disabilities are at higher risk for long-term negative outcomes, such as lower overall academic achievement (Schiller et al., 2008), higher dropout and postgraduation unemployment rates (Thurlow & Johnson, 2011), and higher incarceration rates (Bronson et al., 2015) among others (Newman et al., 2011). Our findings demonstrate that school and community contexts contribute to this inequity. In essence, we place school effects in context (Downey & Condron, 2016). In terms of equity conversations related to students identified as having disabilities, this is an area for further examination, as Tate (2008) states, “an uneven geography of opportunity, left unaddressed, generally grows” (p. 409). Debates about the impact of schools as “engines of inequality” (Downey & Condron, 2016, p. 207) and the best ways to address inequities continue among sociologists and educational economists. Downey and Condron argue that although schools reproduce or exacerbate some inequalities, they also potentially compensate for others. Similar to prior research, we found that average academic performance was strongly associated with school-level poverty (Reardon et al., 2019). Performance was highest in low-poverty schools for every subgroup in both subjects. Also, more students with disabilities in high-income schools performed proficient on state standards than similar students in low-income schools, and the gaps between groups were larger in low-poverty schools. Opportunities to learn are not equally distributed between students even within the same school and a child’s academic development depends on proximal interactions with important socializing adults in key micro-system contexts (Bronfenbrenner, 1979).

In the second largest school district in the country, our findings provide evidence that a child’s academic development and achievement are influenced by an enriched learning context (e.g., highly trained and more experienced educators), the policies that govern the learning context (e.g., school funding formulas), and the connections between schools and policies. These forces interact to mitigate disparities more effectively in some schools than others.

Limitations

LAUSD is highly diverse. For example, the 2018–2019 data indicate that 20.6% of students are English language learners (California Department of Education, 2019). Therefore, the sample in our study may not be representative of other districts in California or nationally and, therefore, the external validity of this study’s findings is limited. Nonetheless, the results provide evidence that local funding can create meaningful changes in student outcomes, but that social context continues to moderate this relationship. Specifically, gains in special education spending predict higher achievement levels overall, but the effect of spending increases varies by contextual factors, particularly poverty. Our findings indicate that a higher share of credentialed special education teachers could reduce large achievement gaps between groups. This could have implications for districts across the country, which are seeking to hire and retain qualified special education teachers, to support teachers in obtaining a special education credential and to equitably distribute credentialed teachers across schools.

The secondary data analyzed here come with the typical limitations inherent to this type of research. In particular, the variables were selected according to the data available. The policy field could particularly benefit from future studies that examine neighborhood effects, including the racial and ethnic composition of the larger community, to investigate how those with increased political and social capital might shape district policies and practices. The analytical method in this study (i.e., the simulated instrumental variable approach) protected against unmeasured confounding and reduced the problem of selection bias inherent in nonexperimental studies. This approach enabled us to establish a causal relationship and to provide meaningful evidence to guide policy makers (Bounthavong et al., 2016).

Future Directions

Although we examined indirect relationships of school funding and opportunity through student achievement, we continue to lack knowledge about direct special education service delivery quality, given more nuanced context variables. Future studies that examine the effect of funding increases on curricular structure and quality special education services more directly are needed. Structural and interactional social forces are interdependent and reinforce each other within schools to generate patterns of inequality. These social forces are spatially distributed across and within social institutions and collectively afford and constrain educational opportunity through uneven geographies of opportunity (Tate, 2008). Tefera et al. (2017) suggested that studying uneven geographies of opportunity can contextualize educational inequities; a spatial lens situates how structural and interactional opportunity gaps manifest through both enabling and disabling geographies of opportunity. Future research should examine the spatial distribution of resources and opportunities in neighborhoods and communities to “create maps that capture the geographies of opportunity in which students and families live” that identify both “traditional neighborhood boundaries with families’ subjective neighborhood borders” (Tefera et al., 2017, p. 202). This perspective accounts for the dynamic interplay between structural and interactional educational opportunity gaps that generate inequities and can help scholars and policy makers best track both intended and unintended consequences of policy and funding shifts on student outcomes.

Footnotes

Authors’ Note

Joon-Ho Lee is now affiliated with the University of Alabama, Tuscaloosa, USA.

Declaration of Conflicting Interests

The author(s) declared no potential conflicts of interest with respect to the research, authorship, and/or publication of this article.

Funding

The author(s) received no financial support for the research, authorship, and/or publication of this article.

ORCID iD

Rebecca A. Cruz

References

Assembly Bill 602. Special Education Reform Act to amend Cal. Educ. Code § 56836.08. (1997). http://www.leginfo.ca.gov/pub/97-98/bill/asm/ab_0601-0650/ab_602_bill_19971010_chaptered.html

Bischoff

Owens

(2019). The segregation of opportunity: Social and financial resources in the educational contexts of lower-and higher-income children, 1990–2014. Demography, 56(5), 1635–1664. https://doi.org/10.1007/s13524-019-00817-y

Bounthavong

Adamson

Basu

(2016). Instrumental variables: A tool to reduce bias in non-randomized studies. Value and Outcomes Spotlight, 2, 24–25.

Bronfenbrenner

(1979). The ecology of human development. Harvard University Press.

Bronson

Maruschak

L. M.

Berzofsky

(2015). Disabilities among prison and jail inmates, 2011–12. U.S. Department of Justice Bureau of Justice Statistics. https://www.bjs.gov/content/pub/pdf/dpji1112.pdf

California Department of Education. (2018). Special education AB 602 apportionment. https://www.cde.ca.gov/fg/fo/profile.asp?id=5094&recID=5094

California Department of Education. (2019). The California longitudinal pupil achievement data system (CALPADS) [Staff data file]. https://www.cde.ca.gov/ds/sd/df/

Card

Payne

A. A.

(2002). School finance reform, the distribution of school spending, and the distribution of student test scores. Journal of Public Economics, 83(1), 49–82. https://doi.org/10.1016/S0047-2727(00)00177-8

Carter

P. L.

Welner

K. G.

(Eds.). (2013). Closing the opportunity gap: What America must do to give every child an even chance. Oxford University Press.

10.

Coleman

J. S.

Campbell

E. Q.

Hobson

C. J.

McPartland

Mood

A. M.

Weinfeld

F. D.

York

R. L.

(1966). Equality of educational opportunity (OE-38001; Superintendent of Documents Catalog No. FS 5.238:38001). U. S. Government Printing Office.

11.

Darling-Hammond

(2015). The flat world and education: How America’s commitment to equity will determine our future. Teachers College Press.

12.

Downey

D. B.

Condron

D. J.

(2016). Fifty years since the Coleman report: Rethinking the relationship between schools and inequality. Sociology of Education, 89(3), 207–220. https://doi.org/10.1177/0038040716651676

13.

Education for All Handicapped Children Act of 1975, Public Law 94–14220 U.S.C. § 1400. (1975).

14.

Fuchs

L. S.

McMaster

K. L.

Lemons

C. J.

(2018). Students with disabilities’ abysmal school performance: An introduction to the special issue. Learning Disabilities Research & Practice, 33(3), 127–130. https://doi.org/10.1111/ldrp.12180

15.

Griffith

(2015). The progress of education reform: A look at funding for students with disabilities. Education Commission of the States. https://files.eric.ed.gov/fulltext/ED560999.pdf

16.

Hanushek

E. A.

Rivkin

S. G.

Taylor

L. L.

(1996). The identification of school resource effects. Education Economics, 4(2), 105–125. https://doi.org/10.1080/09645299600000012

17.

Hibel

Farkas

Morgan

P. L.

(2010). Who is placed into special education? Sociology of Education, 83(4), 312–332. https://doi.org/10.1177/0038040710383518

18.

Hill

Warren

Murphy

Ugo

Pathak

(2016). Special education finance in California. Public Policy Institute of California. https://www.ppic.org/content/pubs/report/R_1116LHR.pdf

19.

Humphrey

Koppich

Lavadenz

Marsh

O’Day

Plank

Stokes

Hall

(2017). Paving the way to equity and coherence? The local control funding formula in year 3. Policy Analysis for California Education, Stanford Graduate School of Education. https://edpolicyinca.org/sites/default/files/LCFFRC_04_2017.pdf

20.

Hussar

Zhang

Hein

Wang

Roberts

Cui

Smith

Bullock Mann

Barmer

Dilig

(2020). The Condition of Education 2020 (NCES 2020-144). U.S. Department of Education, National Center for Education Statistics. https://nces.ed.gov/pubs2020/2020144.pdf

21.

Kincaid

A. P.

Sullivan

A. L.

(2017). Parsing the relations of race and socioeconomic status in special education disproportionality. Remedial and Special Education, 38(3), 159–170. https://doi.org/10.1177/0741932516671199

22.

Kolbe

(2019). Funding special education: Charting a path that confronts complexity and crafts coherence. National Education Policy Center. http://nepc.colorado.edu/publication/special-ed

23.

Kramarczuk Voulgarides

(2018). Does compliance matter in special education? IDEA and the hidden inequities of practice. Teachers College Press.

24.

Lafortune

Rothstein

Schanzenbach

D. W.

(2018). School finance reform and the distribution of student achievement. American Economic Journal: Applied Economics, 10(2), 1–26. https://doi.org/10.1257/app.20160567

25.

Lee

J. H.

Fuller

(2020). Does progressive finance alter school organizations and raise achievement? The case of Los Angeles. Educational Policy, 36(3), 587–623. https://doi.org/10.1177/0895904820901472

26.

Milner

H. R.

(2013). Rethinking achievement gap talk in urban education. Urban Education, 48(1), 3–8. https://doi.org/10.1177/0042085912470417

27.

National Center for Education Statistics. (1993). 120 years of American education: A statistical portrait. U.S. Department of Education, Office of Educational Research and Improvement. https://nces.ed.gov/pubs93/93442.pdf

28.

Newman

Wagner

Knokey

A. M.

Marder

Nagle

Shaver

Wei

(with Cameto, R., Contreras, E., Ferguson, K., Greene, S., & Schwarting, M.) (2011). The post-high school outcomes of young adults with disabilities up to 8 years after high school: A report from the National Longitudinal Transition Study-2 (NLTS2) (NCSER 2011-3005). SRI International. https://ies.ed.gov/ncser/pubs/20113005/pdf/20113005.pdf

29.

Orfield

Frankenberg

Siegel-Hawley

(2016). Brown at 62: School segregation by race, poverty and state. Civil Rights Project/Proyecto Derechos Civiles. https://civilrightsproject.ucla.edu/research/k-12-education/integration-and-diversity/brown-at-62-school-segregation-by-race-poverty-and-state/Brown-at-62-final-corrected-2.pdf

30.

Owens

(2018). Income segregation between school districts and inequality in students’ achievement. Sociology of Education, 91(1), 1–27. https://doi.org/10.1177/0038040717741180

31.

Petek

(2019). An overview of special education in California. Legislative Analysist’s Office. https://lao.ca.gov/reports/2019/4110/overview-spec-ed-110619.pdf

32.

Reardon

S. F.

Kalogrides

Shores

(2019). The geography of racial/ethnic test score gaps. American Journal of Sociology, 124(4), 1164–1221.

33.

Schiller

Sanford

Blackorby

(2008). A national profile of the classroom experiences and academic performance of students with LD: A special topic report from the Special Education Elementary Longitudinal Study. SRI International. http://www.seels.net/info_reports/national_profile_students_learning_disabilities.htm

34.

Sullivan

A. L.

Bal

(2013). Disproportionality in special education: Effects of individual and school variables on disability risk. Exceptional Children, 79(4), 475–494. https://doi.org/10.1177/001440291307900406

35.

Sutcher

Darling-Hammond

Carver-Thomas

(2016). A coming crisis in teaching? Teacher supply, demand, and shortages in the US. Learning Policy Institute. https://learningpolicyinstitute.org/product/coming-crisis-teaching-brief

36.

Tate

W. F.

IV . (2008). “Geography of opportunity”: Poverty, place, and educational outcomes. Educational Researcher, 37(7), 397–411. https://doi.org/10.3102/0013189X08326409

37.

Tefera

A. A.

Aguilar

C. R.

Artiles

A. J.

Voulgarides

C. K.

Vélez

(2017). Developing a critical space perspective in the examination of the racialization of disabilities. In Ares

Buendía

Helfenbein

(Eds.), Deterritorializing/reterritorializing: Critical geography of educational reform (pp. 191–207). Brill/Sense.

38.

Thurlow

M. L.

Johnson

D. R.

(2011). The high school dropout dilemma and special education students [The California Dropout Research Project]. University of California. http://cdrpsb.org/dropouts/researchreport18.pdf

39.

Vasquez Heilig

Romero

L. S.

Hopkins

(2017). Coign of vantage and action: Considering California’s local accountability and school finance plans for English learners. Education Policy Analysis Archives, 25, 1–24.

40.

Warren

Hill

(2018). Revisiting finance and governance issues in special education [Technical report: Getting Down to Facts II]. Stanford University: Public Policy Institute of California. https://gettingdowntofacts.com/sites/default/files/2018-09/GDTFII_Brief_SpecialEducation.pdf