The Role of State Education Regulation: Evidence From the Texas Districts of Innovation Statute

Abstract

School districts are traditionally subject to a variety of state regulations on educational inputs. Absent regulations, policymakers fear that districts will make inappropriate decisions. However, it is also possible that regulations hinder schools from optimizing student learning. This article tests the salience of these hypotheses by estimating the impact of the Texas District of Innovation statute, which allows districts to opt out of regulations on inputs like teacher certification and class sizes. Using web-scraped implementation data, I document widespread exemptions and variation in regulatory preferences. However, staggered difference-in-differences analyses demonstrate a limited impact of deregulation within 4 years, suggesting that deregulation alone is a relatively weak lever for spurring innovation and changing the state of education.

Keywords

economics of education educational policy educational reform governance quasi-experimental analysis econometric analysis

What is the role of state legislatures in improving public education? Is it to provide funding and standards for student outcomes and then step aside? Or should legislatures and state education agencies also govern how districts educate students? In practice, states hold school districts accountable for academic achievement while also requiring them to prove compliance with a variety of regulations on educational inputs (Cohen et al., 2017). Commonly regulated inputs include minimum teacher qualifications, minimum class time, and maximum class sizes (Education Commission of the States, 2005, 2017, 2019). Until the end of the 20th century, these input regulations formed the core of state education programs (Hanushek, 1996). Yet, even as state accountability statutes came to the forefront of education policy following the No Child Left Behind Act of 2002, input regulations have remained (Cohen et al., 2017).

Prior to 2015, this was the state of affairs for traditional public school districts in Texas; schools in the state were required to comply with teacher certification standards, class size limits, calendar restrictions, and limitations on staff contracts, among other requirements, all while meeting academic accountability standards. But in 2015, the traditional approach to input regulation was upended. Today, any Texas district with an acceptable academic and financial rating (a standard met by more than 99% of traditional public school districts) can declare itself a District of Innovation and opt out of any regulation that does not apply to the state’s charter schools. As of June 2022, nearly 90% of Texas districts have claimed innovation status, exempting an average of eight regulations apiece, including teacher certification requirements, maximum class sizes, and minimum instruction time. These districts are still held accountable to student outcomes, but may meet those academic standards however they choose. Seventy-eight percent of Texas students now attend a District of Innovation.

Such an approach to education policy is lauded by many accountability policy advocates; accountability statutes were often imagined as an alternative to the more traditional regulatory approach to state policy. Those in favor of accountability-based policy argue that “the new role of states is to promote and encourage experimentation and implementation of new incentive systems” and that states should “remove unproductive input regulations and certification standards” (Hanushek, 1996, p. 48). Underlying this argument is the theory that regulations restrict schools from tailoring inputs to their unique circumstances and from serving their student populations as those circumstances require (Elmore & Fuhrman, 1995; Hanushek, 2003). On the other hand, many fear that without input regulations, some district leaders would make inappropriate decisions and thereby harm students and increase inequality (Elmore & Fuhrman, 1995). These fears were evident following the passage of the District of Innovation statute as the state’s teachers’ unions issued warnings to their members (Texas Classroom Teachers Association, 2017) and argued that it is possible to be innovative “without lifting legal protections that . . . protect students and educators” (Texas American Federation of Teachers [AFT], 2021).

It is an open question whether either fear or optimism is an appropriate response to the District of Innovation statute; empirical evidence of the impact of deregulation on traditional public schools is limited. Much of this evidence has come from the charter literature, but charter schools differ from traditional public schools in their governance structure, their political context, and their competition for students (Education Commission of the States, 2018; National Center for Education Statistics, 2019). There is also a long line of research assessing the impact of individual input regulations on student achievement (Boyd et al., 2006; Hanushek, 1986; Kane et al., 2008). This research is helpful in determining which inputs states should consider regulating but does not answer the larger question of how districts would respond if the sum total of input regulations passed by state legislatures was lifted. Indeed, it is an open question which, in any, regulatory policies districts would prefer to be lifted. While survey data may reveal stated preferences, the District of Innovation statute provides the opportunity to observe revealed preferences.

To provide empirical evidence on the role of state education regulation in determining school decisions and student outcomes, this article uses the Texas District of Innovation statute to shed light on (1) which regulations districts are most enthusiastic to remove, (2) the extent to which deregulation changes local decisions on educational inputs, and (3) the extent to which deregulation impacts student outcomes. Recognizing that equity is of particular concern with deregulation, this article focuses on gaps in educational inputs and academic outcomes under the District of Innovation statute for rural students, economically disadvantaged students, and students of color.

Through analyzing District of Innovation plans scraped from district websites, I find that Texas school districts have responded enthusiastically to the opportunity to opt out of input regulations. Nearly 90% of districts have claimed District of Innovation status, and of these, 87% have exempted teacher certification requirements, 44% have exempted elementary class size maximums, and 42% have exempted the minimum class time requirement. Regulatory preferences are also highly correlated with district geography; rural districts are particularly eager to exempt teacher certification requirements and restrictions on teacher contracts, while urban districts are particularly eager to exempt restrictions on student attendance and responses to student behavior. These distinct responses suggest that district leaders feel that input regulations do not address contextual needs.

Yet, despite enthusiastic adoption, I find that regulatory freedom under the District of Innovation statute caused limited changes in observable school inputs. Taking advantage of staggered adoption of regulatory flexibility, I use an adapted generalized difference-in-differences (DID) strategy recommended by Callaway and Sant’Anna (2020) to estimate the impact of regulatory freedom on 12 key educational inputs, including the percentage of uncertified teachers, out-of-field teachers, average elementary class sizes, and student–teacher ratios. Each of these impact estimates is a relatively precisely estimated zero within the first 4 years of implementation. Furthermore, although there are substantial gaps in educational inputs among average schools and those serving high proportions of Hispanic, Black, rural, and economically disadvantaged students, the District of Innovation statute did little to either exacerbate or lessen these inequities in the medium-term. Finally, I find a statistically insignificant and insubstantial 3-year effect of the District of Innovation statute on student achievement: less than 0.01 school-level standard deviations (SDs) in mathematics and reading.

Taken together, these results indicate that although school boards are generally in favor of increased regulatory flexibility, this flexibility did not dramatically alter school inputs within 4 years or school outcomes within 3 years. Notably, the previous regulatory regime never resulted in perfect compliance. For example, the rate of uncertified teachers in Texas has hovered around 2% since 2015. Furthermore, like most state education departments, the Texas Education Agency does not inspect public schools to the same degree that a health department might inspect restaurants (Elmore & Fuhrman, 1995). Because of this limited enforcement, districts know that if a regulation is too burdensome, they can ignore it with few repercussions (Elmore & Fuhrman, 1995). Thus, while the District of Innovation statute dramatically changed the state education code, this change may not have ultimately been necessary for districts to maintain inputs at their preferred levels. On the other hand, the Texas District of Innovation statute is a large nearly state-wide experiment in deregulating input requirements within traditional public schools. The limited medium-term impact of the statute on school inputs and student outcomes suggests that deregulation alone is a relatively weak lever for spurring innovation and changing the state of education.

Background

Theoretical Justifications for and Against Input Regulations

As accountability statutes like No Child Left Behind and its state-level predecessors came to the forefront of education policy, many imagined them as an alternative to the more traditional input-focused approach. Advocates of accountability argued that if states were to hold districts accountable for student performance, then they also should remove unproductive regulatory barriers that may make it more difficult for districts to meet their goals (Hanushek, 1996; Wöbmann et al., 2007). This view is at least partially bolstered by evidence. When it comes to easily regulated educational inputs, research has shown that “more is not necessarily better” (Cohen et al., 2017, p. 204). Given the difficulty of creating optimal policy, advocates of deregulation argue that students are better served when their education is designed by people within their community, rather than by policymakers in a distant state capital. They argue that school board members, superintendents, principals, and teachers all have greater familiarity with local resources, needs, preferences, priorities, and ideologies than state governments (Fuhrman & Elmore, 1995; King & Ozler, 1998; Moe, 2003).

Furthermore, there remains a question of whether state officials have the capacity and motivation to set meaningful input policies. Elected officials represent a variety of powerful education constituents and often attempt to satisfy them by setting mandates low enough that most districts exceed them with little effort (Elmore & Fuhrman, 1995). Furthermore, states rarely have centralized education offices with the ability to inspect, monitor, and enforce regulatory compliance. Without inspection ability, agencies instead rely on districts to honestly document compliance on paper, but, for the most part, districts know that if they find a regulation too burdensome, they can often ignore it with few repercussions (Elmore & Fuhrman, 1995).

On the other hand, proponents of regulation consider the goal behind input regulations to be noble; education regulations are intended to assure minimum standards throughout the state regardless of demographics, wealth, size, or capacity. Given persistent inequalities in resources, teacher quality, and programmatic offerings (Darling-Hammond & Berry, 2006; Jennings et al., 2015; Lankford et al., 2002), it is difficult to argue for the removal of minimum quality standards. Furthermore, input regulations may protect against inappropriately designed incentive systems. Although accountability policies commonly increase average student performance (Carnoy & Loeb, 2002; Dee et al., 2010; Figlio & Loeb, 2011), they can also cause schools to shift resources away from nontested content and from support for students on either end of the achievement distribution (Dee et al., 2010). Input regulations can ensure that even nontested subjects and students who are unlikely to pass state exams are taught by a certified teacher making at least some minimum salary in a classroom with a reasonable number of students.

Previous Experiments With Deregulation

The most common contemporary experiment with education deregulation comes in the context of charter schools. While state legislation often holds charter schools accountable for student outcomes, charter schools are rarely required to comply with regulations that are not related to health, safety, and civil rights (S. Cohodes, 2018). Perhaps due in part to this flexibility, the teaching workforce in charters differs from that of traditional public schools; charter teachers are commonly younger, less experienced, less likely to be certified, and less likely to be offered tenure by the school (Angrist et al., 2013; Bruhn et al., 2020; S. R. Cohodes & Parham, 2021; Preston et al., 2012). Despite some shared characteristics of the teaching workforce, research has demonstrated the highly variable effectiveness of charter schools (Chabrier et al., 2016; Chingos & West, 2015). While on average charters are no more or less effective than traditional public schools, a subset of urban charters and charters serving high proportions of disadvantaged students have demonstrated significant, positive impacts for students, enough to make substantial gains toward closing the gap in test scores between advantaged and disadvantaged students (S. Cohodes, 2018).

In particular, researchers have homed in on a set of charters subscribing to a “No Excuses” model, like the Knowledge is Power Program. No Excuses schools have some of the most impressive effect sizes across the sector, often nearing half a standard deviation (S. Cohodes, 2018). Importantly, the No Excuses practices that researchers have identified as likely contributing to these schools’ differential success above other charters do not necessarily require deregulation, but a shift in culture and investment. The most common practices include an increased emphasis on behavioral and academic expectations, tutoring and small-group instruction, increased feedback for teachers, and an extended school day (Angrist et al., 2010; Baude et al., 2020; S. R. Cohodes & Parham, 2021; Dobbie & Fryer, 2013, 2020).

Although these practices can be injected into traditional public schools with success (Fryer, 2014), the circumstances surrounding charters may make it easier. Unlike traditional public schools, charter schools are not beholden to an elected school board, cultural norms surrounding traditional public schools, or teachers’ unions. Charter schools also face heightened incentives to experiment with practices that can increase student success; they face not only competition for student attendance, but also the threat of closure if they fail to produce acceptable outcomes (Abdulkadiroǧlu et al., 2011; Lubienski, 2003).

Like other traditional public school districts, Districts of Innovation do not face substantial competition for students or threat of closure. They do, on the other hand, face political and cultural barriers to change, including the presence of traditional hierarchies, bureaucracies, local politics, and school boards. The question, then, is how do traditional public school districts respond to regulatory freedom? Two examples may be of use here. First, compared with charter schools, Districts of Innovation are more similar to a much smaller experiment with autonomy: Boston pilot schools. Boston’s pilot schools have the same regulatory independence as the city’s charter schools, but remain within the Boston Public School district, and so are beholden to district policies and union contracts. Compared with local charter schools, Boston’s pilot schools implement few changes in inputs and have a substantially smaller (and less positive) impact on students (Abdulkadiroǧlu et al., 2011), demonstrating that regulatory autonomy alone does not result in academic gains for students. Second, previous experiments in fiscal deregulation demonstrate the stickiness of business as usual. When California deregulated US$4.5 billion in education funding, allowing districts flexibility in how to allocate these previously earmarked resources, proponents of decentralizing school finance argued that flexibility would allow districts to better align resources with local priorities and optimize for instructional gains (Fuller et al., 2011). Yet, in practice, increased fiscal flexibility resulted in little innovation or advancement of new initiatives. Instead, school districts commonly swept the funds into their General Fund and continued to balance the budget as usual (Fuller et al., 2011).

The Texas Context

The Texas District of Innovation statute was passed in 2015 and was followed by a steady but enthusiastic response from districts. By the 2016–2017 school year, 178 districts had claimed District of Innovation status. An additional 509 districts followed suit in the 2017–2018 school year, along with 119 districts in 2018–2019 and another 102 districts between 2020 and 2022. Although districts may still claim Innovation status, adoption has slowed. In 2022, there were only 16 new Districts of Innovation. As of June 2022, 90% of school boards have voted to become a District of Innovation.

The process of claiming District of Innovation status involves five simple steps wherein the school board (a) resolves to become a District of Innovation, (b) appoints an Innovation Committee to identify regulations the district would like to exempt, (c) votes on the final plan listing the exemptions, (d) posts the innovation plan on the district website, and (e) notifies the Education Commissioner of their new status. Districts of Innovation are exempt from as many regulations as they choose. The term length of the designation as a District of Innovation cannot exceed 5 years, but it may be renewed as long as the district continues to meet academic and financial accountability standards.

Ineligibility for District of Innovation status is rare, ranging from just 5 to 16 ineligible districts per year. Thus, the population of 127 non–Districts of Innovation is primarily made up of districts who have (so-far) opted out of deregulation despite eligibility. Compared with Districts of Innovation, these districts are more urban, have a student body that is more Hispanic and more economically disadvantaged, and have lower state standardized test scores in the pre-deregulation period (see Supplemental Table A1 in the online version of the journal).

Conceptual Model

As pictured in the conceptual model in Figure 1, this study hypothesizes that for deregulation to impact students, (a) school boards must have an interest in deregulation and act on that interest by exempting input regulations that are theoretically related to student experiences and outcomes; (b) schools must meaningfully change their inputs into the education process; and (c) schools must meaningfully change their inputs in ways that matter for student achievement. Importantly, if this process is moderated by contextual factors that are related to student needs and characteristics (pictured in boxes in Figure 1, with bolded factors representing factors that are addressed in analyses), then the District of Innovation could result in greater inequity across the state.

Figure 1.

District of Innovation Conceptual Model Guiding Study

Key research questions flow from this conceptual model:

Following the District of Innovation statute, to what extent do districts demonstrate an interest in exemptions and which regulations do those districts exempt?

To what extent does District of Innovation status impact school-level inputs, particularly in schools serving high proportions of students of color, economically disadvantaged students, and rural students?

To what extent does District of Innovation status impact school-level student achievement, including the academic achievement of Black students, Hispanic students, economically disadvantaged students, and rural students?

Data and Measures

District of Innovation Plans

This study relies on key information found in district innovation plans: including the regulations from which each district will be exempt and the period during which the innovation plan will be active. Like many other online policy documents, district innovation plans contain rich data, but they are found in disparate locations, are stored using diverse media, and require the capacity to turn natural language into structured data to extract value. To address these challenges, I use a combination of web-scraping, text classification, and natural language processing. A comprehensive overview of these methods is beyond the scope of this article, but readers may turn to Fesler et al. (2019) for an introduction to text classification and natural language processing for education research and to Landers et al. (2016) for an introduction to web-scraping. Additional details on how these methods may be applied within the context of policy documents, including a more comprehensive explanation of the data collection methods used in this article, may be found in Anglin (2019).

Here, the data collection process proceeded in three steps where I (a) built a web crawler that visits every district website linked on the Texas Districts of Innovation webpage and downloads all potential documents; (b) trained a text classifier (which automatically categorizes documents) to identify innovation plans and discard irrelevant documents; and (c) parsed each document into a list of regulatory exemptions and the date of implementation. Regulatory exemptions were identified using regular expressions specifying patterns representing statutes: two to three numerals followed by a period and one to four more numerals (e.g., 25.0811 and 21.003 would both be captured by the regular expression). Then, the date of implementation was extracted using a prebuilt model for detecting dates using the SpaCy software package (Honnibal, 2017). Excerpts surrounding dates were manually reviewed to identify which date represented the District of Innovation plan start date. If no excerpt indicated the start date, the full document was manually reviewed.

These scraped data were then additionally validated using two methods. First, I compared the resulting corpus of district innovation plans with the list of District of Innovation maintained by the Texas Education Agency. If any Districts of Innovation were missing from my scraped data set, I added them manually. Second, to validate extracted regulations, I randomly selected 30 innovation plans and manually coded those documents for their list of regulatory exemptions. This hand-coded data set included 243 exempted regulations. I then compared the test set with the laws extracted using pattern matching. Of the 243 true exemptions, the automated method correctly identified 239 true positives and misclassified four false negatives and 10 false positives, resulting in a recall rate of 98% (98% of exemptions were identified) and a precision rate of 96% (of the exemptions identified, 96% were truly exempted by the district).

Inputs and Outputs

Data on school-level inputs and outcomes come from the Texas Education Agency, either from the Texas Academic Performance Reports (TAPR) or State of Texas Assessments of Academic Readiness (STAAR) aggregate data, or through a public information request for data contained in the Public Education Information System (PEIMS) from the 2011–2012 to 2021–2022 school years. An overview of these data can be found in Supplemental Appendix B (in the online version of the journal). I focus on eight school-level educational inputs: the proportion of uncertified teachers, the proportion of out-of-field teachers, student–teacher ratios, average elementary class sizes, the number of teachers, the number of new teachers, average teacher experience, and average teacher turnover. In addition, I present results for the number of school days, school start dates, the average number of minutes in a school day, and the total number of school minutes. These calendar data are only available following the 2016–2017 school year, however, and so those results should be treated as exploratory. For each input, in addition to presenting results for the average school, I focus on schools in the top quartile of the state in terms of the proportion of Hispanic students, Black students, and economically disadvantaged students, as well as rural schools.

The primary outcome of interest in this article is test scores from the STAAR exams, a set of standardized exams first implemented in the 2011–2012 school year. I use mathematics scores and reading scores in third through eighth grade as well as English I and Algebra scores in high school to measure performance. In Supplemental Appendix D (in the online version of the journal), I estimate effects on Biology and U.S. History exams. With all tests, scores are standardized within grade and subject using 2015 means and SDs. I also estimate the impact of District of Innovation status on school-level attendance rates. With academic and attendance outcomes, in addition to presenting average outcomes within a school, I also present average outcomes for each school’s Black students, Hispanic students, and economically disadvantaged students.

Note that I limit my analyses to input outcomes occurring before the 2020–2021 school year. This is because the COVID-19 pandemic presents a severe threat to the parallel trends assumption, discussed below. Researchers have demonstrated the highly variable impact of COVID-19 on student achievement and schools, with the greatest losses in learning being experienced by economically disadvantaged students and students of color (Hammerstein et al., 2021; Kuhfeld et al., 2020). Thus, even small differences in cohort characteristics could dramatically bias results. For this reason, impact estimates for academic achievement are limited to those occurring before the 2018–2019 school year, while impact estimates for teacher characteristics include the 2019–2020 school year, as school districts likely hired teachers prior to the start of the pandemic.

Empirical Strategy for Causal Questions

While the Districts of Innovation law is not a clean experiment, its passage creates a series of policy discontinuities that can be used to identify the impact of deregulation on school inputs and student achievement. The District of Innovation statute was passed in 2015, but take-up was staggered over the next several years (see Supplemental Tables A2 and A3 in the online version of the journal). One typical approach to assessing impacts with variation in treatment timing is to employ an event study or generalized DID design with two-way fixed effects. However, recent research demonstrates that weighting schemes embedded in these approaches produce biased and unintuitive impact estimates when effects vary with the length of exposure to the policy (Goodman-Bacon, 2021). I therefore follow the advice of Callaway and Sant’Anna and estimate the impact of District of Innovation status using a series of DID estimates (2020). This strategy has the advantage of producing well-understood impact estimates while capitalizing on treatment variation and controlling for time-invariant confounders (by controlling for pretreatment levels) and history effects (by controlling for comparison group trends). The key assumption in this design is that each cohort would have experienced the same change in outcomes as the comparison group, if not for changes resulting from District of Innovation status itself. This is colloquially known as the parallel trends assumption and it requires that there are no time-varying confounders that differ between the treatment and comparison group.

Defining Treatment Status

In my primary impact analyses, I define the treatment as District of Innovation status. This means that treatment is defined as access to blanket regulatory flexibility on schools within Districts of Innovation. The comparison group consists of schools within districts that have not yet claimed District of Innovation status but do so by the 2021–2022 school year. By narrowing the comparison group to districts that eventually claim District of Innovation status, I minimize differences between the treatment and comparison group, increasing the plausibility of the parallel trends assumption. In Supplemental Appendix D (in the online version of the journal), I also estimate the impact on inputs by defining treatment status as claiming a relevant exemption. For example, when estimating the impact of District of Innovation status on the number of uncertified or out-of-field teachers, this appendix limits the treatment group to those who have exempted teacher certification. The results from this specification are very similar to the results presented in the body of the study. Also in Supplemental Appendix D (in the online version of the journal), I estimate effects including all never-treated districts, which increases precision. Again, results do not qualitatively change.

Impact Estimates

For each implementation cohort (2017, 2018, and 2019) and each year, I estimate a simple two-group, two-time-period DID where the cohort of interest serves as the treatment group and not-yet-treated schools serve as the comparison group. For each cohort, the difference between pretreatment and post-treatment outcomes is estimated using the last pretreatment year before that cohort’s innovation plans were activated. This limited time frame reduces the likelihood of confounding events that occur within the treatment cohort but not within the comparison group. Additional pretreatment years are only used as a validity check, testing for parallel trends in the pretreatment period.

To minimize differences between the treatment and comparison group (thereby increasing the plausibility of the parallel trends assumption), I also control for several baseline school characteristics (calculated in the 2015–2016 school year): the number of students in a school, the percentage of Hispanic students, the percentage of Black students, the percentage of economically disadvantaged students, and academic achievement (defined as the average standardized school-level test scores across all grades and subjects).

Formally, I estimate

\begin{matrix} Y_{s t} = β_{0} + β_{1} T r e a t_{s t} + β_{2} P o s t_{s t} + β_{3} T r e a t P o s t_{s t} \\ + β_{4} S t u d e n t s_{s} + β_{5} H i s p a n i c_{s} + β_{6} B l a c k_{s} \\ + β_{7} E c o n D i s a d v_{s} + β_{8} A c h i e v e m e n t_{s} + ε_{s t} \end{matrix}

where $Y_{s t}$ is the outcome of interest for school s in year t. $T r e a t_{s t}$ is an indicator for whether the school is within the cohort of interest (e.g., whether the school is a 2017 implementer), $P o s t_{s t}$ is an indicator for whether the year is post-treatment for the cohort (in the example, post-2017), $T r e a t P o s t_{s t}$ is an indicator for whether the school and year are post-treatment, $S t u d e n t s_{s}$ is a continuous variable for the number of students in the school in 2015–2016, $H i s p a n i c_{s}$ is a continuous variable for the proportion of Hispanic students, $B l a c k_{s}$ is the proportion of Black students, $E c o n D i s a d v_{s}$ is the proportion of economically disadvantaged students according to Free and Reduced Price Lunch status, and $A c h i e v e m e n t_{s}$ is the standardized baseline school-level achievement averaged across subjects. The estimated impact of District of Innovation status for the cohort and year is represented by the coefficient on $T r e a t P o s t_{s t}$ .

In total, I estimate nine DID treatment effect estimates for each input and six treatment effect estimates for each outcome: the impact of District of Innovation status for the 2017 cohort in 2017, 2018, 2019, and (for inputs only) 2020; the impact for the 2018 cohort in 2018, 2019, and 2020; and the impact for the 2019 cohort in 2019 and 2020. I also estimate 16 pretreatment effect estimates that serve as placebo tests for pretrend assumptions. Estimates are then aggregated using two schemes: (a) dynamic average treatment effect estimates after 1, 2, 3, and, in the case of select inputs, 4 years of implementation, and (b) a summary treatment effect estimate averaged across all cohorts and times, weighted by the size of the cohort. Note that because the second cohort is the largest (with 509 districts), this summary treatment effect more heavily weights times for which the second cohort is included in the treatment group. Thus, third-year effects for outcomes and fourth-year effects for inputs have less of a contribution to this summary treatment effect because the largest cohort did not have 3 years of implementation by 2018–2019.

Threats to Causal Validity

There are two primary ways that the parallel trends assumption might be violated. First, the assumption would be faulty if implementation cohorts were on different trajectories with respect to student enrollment, educational inputs, or student outcomes even before the passage of the District of Innovation statute. This might occur, for example, if early adopters were eager to claim innovation status because they were on a different student achievement trend than their peers. In reality, there are indeed substantive differences between implementation cohorts; the earliest implementers have the highest baseline STAAR performance, followed by the 2017–2018 implementers, 2018–2019 implementers, and 2019–2020+ implementers (see Supplemental Table A1 in the online version of the journal). However, Supplemental Appendix A (in the online version of the journal) demonstrates that while different cohorts have different baseline means, trends are reasonably parallel in the pretreatment period (see Supplemental Figures A1 and A2 in the online version of the journal).

I also probe the likelihood of divergent pretreatment trends by estimating the “impact” of District of Innovation status on school-level inputs and outcomes in the years before the innovation plan was implemented. Figures 5, 6, 7, and 9 show the results of these placebo tests. For all input outcomes, the model fails to identify a treatment effect for any of the pretreatment years, suggesting no significant difference in pretreatment trends. Furthermore, in Supplemental Appendix D (in the online version of the journal), I estimate the impact of District of Innovation status on the proportion of students who are Black, Hispanic, and economically disadvantaged, as well as the number of students in the school. Texas is rapidly diversifying; if changes in demographics occur differentially among implementation cohorts, this might confound the impact of the District of Innovation statute. In this case, however, there is no significant evidence of demographic changes varying by cohort.

The second way that the parallel trends assumption might be violated would be if policy or programmatic changes occurring in the post-treatment time period differentially impacted District of Implementation cohorts. Unfortunately, such a threat to validity is more difficult to probe with specification tests. Instead, understanding the plausibility of this threat requires knowledge of historical policy changes. Here, the most substantive educational policy change that occurred following the passage of the District of Innovation statute was Texas response to the Every Student Succeeds Act (ESSA). ESSA relaxed the requirement that core academic subjects are taught by “highly qualified” teachers and instead requires state to submit plans containing assurances that economically disadvantaged students and students of color are not disproportionately taught by out-of-field or inexperienced teachers. In this case, Texas passed such a plan (the Texas Equity Toolkit) in March 2018 (Texas Education Agency, 2022). Given that later cohorts have a larger proportion of economically disadvantaged students and students of color, this is a potential threat to this study’s identification strategy. However, there are two pieces of evidence that ESSA has not confounded impact estimates. First, effect estimates for the District of Innovation statute are very similar before and after the implementation of Texas’s ESSA plan. Second, effect estimates are also very similar when observing subgroup effects for schools serving high proportions of students economically disadvantaged compared with the average district.

Results

District Responses to Deregulation Under the District of Innovation Statute

Regulatory flexibility under the District of Innovation statute has proven very appealing to Texas school districts; as of June 2022, 90% percent of school boards have voted to become a District of Innovation and 78% of Texas public school students now attend a District of Innovation. On average, Districts of Innovation claim eight exemptions spanning a range of educational inputs. Exemptions commonly concern school schedules, teacher certification, class sizes, teacher contracts, and student behavior. Table 1 presents the 15 most exempted regulations and the proportion of Districts of Innovation exempting that statute, as well as the proportion of the state’s students which are in a district that has exempted the regulation.

Table 1

Proportion of Districts of Innovation Exempting the Most Common Regulations and Proportion of the State’s Students Within Exempting Districts

Regulation	Proportion District of Innovation	Proportion of students impacted
School schedules
25.0811—Earliest first day of instruction	0.98	0.76
25.081—Minimum minutes of operation	0.41	0.33
25.0812—Earliest last day of instruction	0.28	0.18
25.082—Pledge of allegiance and minute of silence	0.25	0.20
Certification
21.003—Teacher certification required	0.87	0.61
21.057—Notice of uncertified teacher	0.32	0.14
21.053—Presentation of teacher certificates	0.30	0.21
Professional contracts
21.102—Maximum probationary contract length	0.52	0.21
21.401—Minimum service days required for teachers	0.36	0.06
21.352—Teacher evaluation requirements	0.19	0.16
Class size
25.112—Elementary class size maximum	0.43	0.29
25.113—Notice of class size	0.36	0.22
Attendance, transfers, and behavior
25.092—Minimum attendance for class credit	0.26	0.30
37.0012—Campus behavior coordinator	0.22	0.23
25.036—Student transfer requirements	0.24	0.05

Note. Statistics are as of June 2022. Data were scraped from District of Innovation plans posted on school district websites. The table presents the top 15 most exempted regulations. The proportion of students is calculated by dividing the number of students within an exempting district by the number of students in the state.

By far, the most popular exemption is the requirement that school districts not begin instruction before the fourth Monday in August; 98% of Districts of Innovation have exempted this requirement, allowing them to start the school year earlier. On the other hand, other popular scheduling exemptions give districts the freedom to shorten the school year. For example, 28% of districts exempted the requirement that the last day of school not occur before May 15, and 41% of districts exempted the requirement that schools operate for at least 75,600 minutes each year (an average of 7 hours a day in a 180-day instructional year).

The second-most popular category of exemptions concerns teacher certification and teaching conditions. Notably, 87% of Districts of Innovation are no longer required to hire certified teachers or to ensure that teachers are certified in the field they teach. Furthermore, 43% of Districts of Innovation have exempted the requirement that elementary class sizes do not exceed 22 students and that teachers be fired or tenured within 3 years. In addition, 36% of Districts of Innovation have exempted the requirement that teacher contracts extend for at least 187 days a year, and 19% have exempted the requirement that teachers are evaluated at least once a year using performance criteria developed by the education commissioner or by the district themselves.

A final category of popular exemptions generally pertains to student attendance, transfers, and behavior infractions. The most popular exemption in this category, with 26% of Districts of Innovation exempting, is the regulation that requires a 90% attendance rate for students to receive course credit. A fifth of districts have also exempted the requirement that schools designate a single person as the campus behavior coordinator who handles behavior referrals, and a quarter have exempted requirements regarding the acceptance of student transfers.

Table 2 disaggregates the percentage of districts exempting each regulation by urbanicity. (Supplemental Table C1, in the online version of the journal, provides additional descriptive statistics of teacher and student characteristics for exempting Districts of Innovation compared with nonexempting Districts of Innovation). Table 2 demonstrates that some categories of exemptions are universally popular, like early school start dates, while others show a clear divide in popularity between urban and rural districts. Compared with urban and suburban districts, rural districts are more likely to exempt regulations related to teacher certification, teacher tenure, and minimum service days required for teachers. For example, while 91% of rural districts have excepted teacher certification requirements, only 78% of urban districts have done the same. Even more striking, while 55% of rural districts have exempted the minimum service days required for teachers, only 2% of urban districts claimed the same exception. These patterns suggest that rural districts struggle to recruit highly effective teachers and so may wish to use regulatory freedom to consider nontraditional uncertified applicants, to entice new applicants by decreasing workdays, and to reduce the number of teachers they are forced to fire or provide with tenure. Conversely, urban districts are less likely to exempt statutes concerning hiring and contracts and more likely to exempt statutes concerning expectations for students. These exemptions allow urban schools to lower attendance standards and to task multiple personnel with responding to behavior infractions.

Table 2

Proportion of Districts of Innovation Exempting Regulations by Urbanicity

Description	Urban	Suburb	Town	Rural	p
School schedules
25.0811—Earliest first day of instruction	1	0.97	0.97	0.98	.51
25.081—Minimum minutes of operation	0.45	0.38	0.35	0.45	.11
25.0812—Earliest last day of instruction	0.22	0.24	0.27	0.31	.21
25.082—Pledge of allegiance and minute of silence	0.2	0.19	0.26	0.25	.25
Class size
25.112—Class size maximum	0.38	0.42	0.44	0.43	.85
25.113—Notice of class size	0.35	0.35	0.38	0.35	.83
25.111—Maximum student–teacher ratio	0.18	0.16	0.15	0.14	.78
Teacher certification
21.003—Teacher certification required	0.78	0.8	0.87	0.91	<.01
21.053—Presentation of teacher certificates	0.32	0.24	0.32	0.31	.21
21.057—Notice of uncertified teacher	0.3	0.25	0.35	0.35	.07
Professional contracts
21.102—Maximum probationary contract	0.2	0.4	0.58	0.58	<.01
21.401—Minimum service days for teachers	0.02	0.14	0.29	0.55	<.01
21.352—Teacher evaluation requirements	0.22	0.2	0.19	0.19	.93
Attendance, transfers, and behavior
25.092—Minimum attendance for credit	0.5	0.34	0.28	0.18	<.01
37.0012—Campus behavior coordinator	0.35	0.27	0.26	0.15	<.01
25.036—Transfers	0.02	0.1	0.26	0.36	<.01

Note. Statistics are as of June 2022. Data were scraped from District of Innovation plans posted on school district websites. The table presents the top 15 most exempted regulations.

Educational Inputs Under the District of Innovation Statute

Figures 2, 3 and 4 display average school-level inputs across the period of study for public schools across the state, excluding charters. From 2016 to 2021, the proportion of uncertified and out-of-field teachers increased across all schools, while class sizes and student–teacher ratios decreased. The number of instructional days and minutes demonstrated substantial variability, but instructional minutes were at their peak in the 2019–2020 school year (and decreased following the pandemic). Importantly, the graphs demonstrate key gaps in educational inputs. Rural schools have a substantially higher proportion of uncertified and out-of-field teachers than the average district, as do schools serving high proportions of Black students (although this gap is smaller). Rural schools and schools serving high proportions of Black students also face higher teacher turnover. On the other hand, rural schools have smaller class sizes and a smaller student–teacher ratio than the average district, likely reflecting a smaller student body overall. Rural schools also have fewer, but longer instructional days, resulting in a greater number of instructional minutes overall, while schools with high numbers of Black students and Hispanic students have the fewest number of instructional minutes.

Figure 2.

Patterns in Educational Inputs Across Texas Public Schools, Excluding Charters

Figure 3.

Patterns in Teacher Characteristics Across Texas Public Schools, Excluding Charters

Figure 4.

Patterns in School Calendars Across Texas Public Schools, Excluding Charters

Yet, despite these inequitable gaps in educational inputs, input inequality was neither exacerbated nor lessened by the District of Innovation statute within the first 4 years of implementation. Table 3 displays the aggregate impact estimates across years and implementation cohorts for each of these 12 educational inputs. Figures 5, 6 and 7 display impact estimates disaggregated by year and school subgroup. Aggregate estimates are small (e.g., a decrease in 0.003 for the number of uncertified teachers and a decrease in less than a tenth of a student for student–teacher ratio) and only reach statistical significance in one case; the statute caused school districts to begin the school year earlier but without increasing the total number of school days or instructional minutes. Furthermore, no school subgroups experience a statistically significant effect on teacher qualifications, class sizes, student–teacher ratios, or school time within 4 years of implementation.

Table 3

Aggregate Impact Estimates Across Years and Implementation Groups

Input or outcome	Estimate	SE	Lower 95% CI	Upper 95% CI
Proportion uncertified teachers	−0.002	0.004	−0.010	0.005
Proportion out-of-field teachers	−0.001	0.002	−0.006	0.003
Average elementary class size	−0.199	0.180	−0.553	0.154
Student–teacher ratio	−0.022	0.107	−0.232	0.188
Number of teachers	0.070	0.292	−0.502	0.642
Number of new teachers	−0.222	0.180	−0.575	0.132
Percent teachers turning over	0.280	0.893	−1.470	2.031
Average teacher experience	0.083	0.157	−0.225	0.390
Instructional days	0.575	0.505	−0.416	1.565
Days before third weeks	0.272	0.033	0.207	0.337
Minutes per day	−1.247	0.999	−3.204	0.710
Total number of minutes	−198	239	−667	271
Standardized math scores	0.004	0.029	−0.052	0.060
Standardized reading scores	0.004	0.022	−0.038	0.047
Student attendance rate	0.043	0.057	−0.070	0.155

Note. Treatment effects are estimated using a simple difference-in-differences strategy where schools within Districts of Innovation that have not yet implemented their Innovation plan are treated as the comparison group for those that have. Overall estimates are an average of all post-treatment years and implementation cohorts weighted by the size of the treatment group, thus giving greatest weight to years for which all cohorts are included in the estimate. Standard errors are in parentheses below effect estimates and are bootstrapped and clustered at the district level. CI = confidence interval.

No estiamtes are statistically significant at alpha threshold of 0.05.

Figure 5.

The Impact of District of Innovation Status on Key Educational Inputs

Figure 6.

The Impact of District of Innovation Status on Teacher Characteristics

Figure 7.

The Impact of District of Innovation Status on School Calendars

Furthermore, Supplemental Appendix C1 (in the online version of the journal) contains boxplots showing the number of outliers in each year, disaggregated by District of Innovation status and traditional public schools. The plots demonstrate that there is no consistent pattern in outliers, providing evidence against the hypothesis that District of Innovation status increases the probability of extreme input disparity across schools. In Supplemental Appendix D (in the online version of the journal), I test this hypothesis more directly by estimating the impact of District of Innovation status on the probability that a school has a proportion of uncertified teachers higher than 0.05, a proportion of out-of-field teachers higher than 0.05, average elementary class sizes greater than 25, or a student–teacher ratio greater than 18. None of these estimates are statistically significant, nor do they trend in a concerning direction in the initial 4 years following implementation.

Academic Outcomes Under the District of Innovation Statute

Figure 8 displays average school-level outcomes across all Texas public schools (including Districts of Innovation but excluding charters) from the 2015–2016 school year to the 2018–2019 school year. While in the previous section graphs were disaggregated by school-level characteristics, these graphs are disaggregated by within-school student characteristics, except for rurality which is a district-level variable. In this time period, average academic outcomes were generally on an upward trend but with achievement gaps that showed little sign of decreasing. In math, rural students, Black students, Hispanic students, and economically disadvantaged students all had average test scores below the state average across the time period. In reading, rural schools tracked the state mean, but, as with math, Hispanic students, Black students, and economically disadvantaged students had test scores below the state average. (Following the COVID-19 pandemic, test scores plummeted—see Supplemental Appendix C2, in the online version of the journal.)

Figure 8.

Patterns in Student Outcomes Across Texas Public Schools, Excluding Charters

However, the District of Innovation statute did little to impact either average outcomes or inequitable gaps in outcomes among Hispanic students, Black students, and economically disadvantaged students. Figure 9 demonstrates that there are no significant effects 1, 2, or 3 years post implementation, with aggregate impact estimates of less than 0.01 SD in mathematics and less than 0.01 SD in reading, both insignificant. The treatment effect for student attendance rate is 0.05 (a 5 percentage point increase in attendance), but the confidence interval on that estimate is quite large, ranging from −0.07 to 0.16. In Supplemental Appendix D (in the online version of the journal), I show the average impact estimates for individual tests, including Biology and U.S. History. After Bonferroni adjustment for multiple hypothesis testing, no subjects show statistically significant impacts, nor does there appear to be any pattern to the magnitude of effects.

Figure 9.

The Impact of District of Innovation Status on Student Outcomes

There is also no evidence of a differential effect of District of Innovation status on the outcomes of Black, Hispanic, or economically disadvantaged students within 3 years. The impact on rural schools, on the other hand, trends in a positive direction: 0.21 SD impact in the third year in mathematics (with a 95% confidence interval ranging from −0.11 to 0.52) and 0.16 SDs (with a confidence interval ranging from −0.10 to 0.43) in reading. Note, however, that these are school-level effect sizes. In Texas, school-level SDs are twice as large as student-level SDs. The third-year effect of the District of Innovation statute in mathematics for this subset of early rural implementers, then, is roughly 0.10 student-level SDs, roughly a third of the treatment effect size in mathematics for Boston charter schools. However, these effect estimates are not statistically significant, have different from zero, have wide-ranging confidence intervals, and are estimated for a small sample of just 112 rural schools in 56 districts that adopted District of Innovation status in the 2016–2017 school year.

Discussion

Regulation advocates and critics offer two competing hypotheses about the role of state regulations in public education. In one camp, experts argue that regulations hinder schools from tailoring inputs to student needs and from serving students to the best of their ability (Fuhrman & Elmore, 1995; Hanushek, 2003). In another camp, experts argue that regulations ensure some minimum standard of quality and protect students against negative outcomes (Darling-Hammond & Berry, 2006; Fuhrman & Elmore, 1995). Considering these hypotheses, the limited impacts of the Texas District of Innovation statute are perhaps surprising; although the statute resulted in more than 5,000 regulatory exemptions in staffing, school time, and human resources, this regulatory flexibility did not meaningfully impact the measured school-level inputs within 4 years of implementation. Importantly, this is also true for demographic subgroups, including rural schools and schools serving high proportions of Black, Hispanic, and economically disadvantaged students. When coupled with descriptive evidence of variability in inputs even before deregulation, these results suggest that, in Texas, school-level decisions are not limited by the necessity of proving compliance with state input regulations. While some Districts of Innovation did indeed change their inputs (e.g., at the maximum, one school increased its rate of uncertified teachers by 39% within 2 years of implementing their District of Innovation plan), non–Districts of Innovation also changed their input levels throughout the period (see Supplemental Appendix C, in the online version of the journal). Importantly, the previous regulatory regime in Texas never resulted in perfect compliance, as the descriptive results of this article demonstrate. Experts have long noted that district leaders can ignore regulations when they need to without severe consequences (Fuhrman & Elmore, 1995). Thus, given limited enforcement of regulations and previous loopholes, the District of Innovation statute may not have been necessary for districts to set their inputs at preferred levels.

Furthermore, Districts of Innovation are still held accountable to student outcomes. Thus, if principals believe previously regulated inputs are important for producing student achievement, they are unlikely to make dramatic changes. And, even if school leaders did believe innovating would be helpful in their circumstances, they may be hesitant to break social norms (Deephouse et al., 2017). Taken together, these two forces—acceptance of some noncompliance, on one hand, and social norms, inertia, and outcomes accountability, on the other—likely result in a sort of stable state that is relatively unimpacted by changes in state regulation.

Limitations

These findings have four key limitations. First, this study only sheds light on the relatively short-term impacts of deregulation: 3 years for student outcomes and 4 years for school-level inputs. Many of the now optional regulations had been in place for over a decade before the District of Innovation statute. The teacher certification and elementary class size requirements, for example, were put in place in 1995 (Education Code Chapter 21 Educators, n.d.; Education Code Chapter 25 Admission, Transfer and Attendance, n.d.). Thus, it is reasonable to assume that there may be some initial inertia in response to deregulation, but that schools will eventually change their practices given enough time. Unfortunately, estimating long-term impacts of the statute is infeasible given the COVID-19 pandemic spurred massive education disruptions and new state-level responses. Second, a key limitation of these analyses is that they cannot distinguish between the input levels of student subgroups within schools. Thus, if schools used regulatory flexibility to shift resources away from some students and toward others (e.g., increasing class sizes for students of color while decreasing class sizes for white students), this would not be identified in these data. One piece of evidence against this scenario, however, is that regulatory flexibility did not cause a detectable impact on the academic achievement of Black students, Hispanic students, or economically disadvantaged students (outcomes for which I am able to disaggregate within schools). Third, we may be most concerned about the impact of deregulation on outcomes for which schools are not held accountable. Changes in student experiences in untested subjects and student’s socioemotional learning, for example, are not accounted for in the current measures. Finally, it should be noted that the null impact of deregulation does not necessarily imply that the original regulations had no role in influencing district behavior or student outcomes. The original regulations may have impacted cultural norms and the standard to which schools hold themselves so that, even after their removal, they continue to impact student experiences.

Conclusion

Despite these limitations, as the first study to estimate the impact of state education deregulation for a near-statewide population of traditional public schools, these results have useful policy implications. While previous studies have examined the impact of deregulation for a select subset of schools, the District of Innovation statute provides the opportunity to study the impact of deregulation at scale. By analyzing the impact of the statute on key educational inputs such as teacher certification and class sizes, this article documents the limited impact of regulatory flexibility on school-level decisions within 4 years following deregulation. Many districts in this context articulate the desire for certain regulatory freedoms by explicitly naming exempted inputs (e.g., teacher certification), but school leaders largely do not act on these freedoms (e.g., by increasing the number of uncertified teachers). These results suggest that if policymakers want to encourage innovation or increase student achievement, deregulation alone is unlikely to be effective.

Supplemental Material

sj-pdf-1-epa-10.3102_01623737231176509 – Supplemental material for The Role of State Education Regulation: Evidence From the Texas Districts of Innovation Statute

Supplemental material, sj-pdf-1-epa-10.3102_01623737231176509 for The Role of State Education Regulation: Evidence From the Texas Districts of Innovation Statute by Kylie Anglin in Educational Evaluation and Policy Analysis

Footnotes

Acknowledgements

The author would like to thank Vivian Wong, Beth Schueler, and James Wycoff at the University of Virginia for their helpful comments throughout the development of the article. She is also appreciative of feedback received by conference attendees at Association for Public Policy Analysis and Management (APPAM) and Association for Education Finance and Policy (AEFP) and from three anonymous reviewers.

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) disclosed receipt of the following financial support for the research, authorship, and/or publication of this article: The research reported here was supported by the National Academy of Education and the National Academy of Education/Spencer Dissertation Fellowship Program.

ORCID iD

Kylie Anglin

Supplemental Material

Supplemental material for this article is available online.

Author

KYLIE ANGLIN, PhD, is an assistant professor at the University of Connecticut. Her research uses natural language processing and causal inference to improve program and policy evaluation.

References

Abdulkadiroǧlu

Angrist

J. D.

Dynarski

S. M.

Kane

T. J.

Pathak

P. A.

(2011). Accountability and flexibility in public schools: Evidence from Boston’s charters and pilots. Quarterly Journal of Economics, 126(2), 699–748. https://doi.org/10.1093/qje/qjr017

Anglin

K. L.

(2019). Gather-narrow-extract: A framework for studying local policy variation using web-scraping and natural language processing. Journal of Research on Educational Effectiveness, 12(4), 685–706. https://doi.org/10.1080/19345747.2019.1654576

Angrist

J. D.

Dynarski

S. M.

Kane

T. J.

Pathak

P. A.

Walters

C. R.

(2010). Inputs and impacts in charter schools: KIPP lynn. American Economic Review, 100(2), 239–243. https://doi.org/10.1257/aer.100.2.239

Angrist

J. D.

Pathak

P. A.

Walters

C. R.

(2013). Explaining charter school effectiveness. American Economic Journal: Applied Economics, 5(4), 1–27. https://doi.org/10.1257/app.5.4.1

Baude

P. L.

Casey

Hanushek

E. A.

Phelan

G. R.

Rivkin

S. G.

(2020). The evolution of charter school quality. Economica, 87(345), 158–189. https://doi.org/10.1111/ecca.12299

Boyd

Grossman

Lankford

Loeb

Wyckoff

(2006). How changes in entry requirements alter the teacher workforce and affect student achievement. Education Finance and Policy, 1(2), 176–216. https://doi.org/10.1162/edfp.2006.1.2.176

Bruhn

J. M.

Imberman

S. A.

Winters

M. A.

(2020). Regulatory arbitrage in teacher hiring and retention: Evidence from Massachusetts charter schools. National Bureau of Economic Research.

Callaway

Sant’Anna

P. H. C.

(2020). Difference-in-differences with multiple time periods. Journal of Econometrics, 225(2), 200–230. https://doi.org/10.1016/j.jeconom.2020.12.001

Carnoy

Loeb

(2002). Does external accountability affect student outcomes? A cross-state analysis. Educational Evaluation and Policy Analysis, 24(4), 305–331. https://doi.org/10.3102/01623737024004305

10.

Chabrier

Cohodes

Oreopoulos

(2016). What can we learn from charter school lotteries? Journal of Economic Perspectives, 30(3), 57–84. https://doi.org/10.1257/jep.30.3.57

11.

Chingos

M. M.

West

M. R.

(2015). The uneven performance of Arizona’s charter schools. Educational Evaluation and Policy Analysis, 37(1S), 120S–134S.

12.

Cohen

D. K.

Spillane

J. P.

Peurach

D. J.

(2017). The dilemmas of educational reform. Educational Researcher, 47(3), 204–212. https://doi.org/10.3102/0013189X17743488

13.

Cohodes

(2018). Charter schools and the achievement gap [The future of children, policy]. https://futureofchildren.princeton.edu/sites/g/files/toruqf2411/files/resource-links/charter_schools_compiled.pdf

14.

Cohodes

S. R.

Parham

K. S.

(2021). Charter schools’ effectiveness, mechanisms, and competitive influence. https://blueprintlabs.mit.edu/research/charter-schools-effectiveness-mechanisms-and-competitive-influence/

15.

Darling-Hammond

Berry

(2006). Highly qualified teachers for all. Educational Leadership, 64(3), 14–20.

16.

Dee

Jacob

Hoxby

Ladd

H. F.

(2010). The impact of no child left behind on students, teachers, and schools. Brookings Papers on Economic Activity, 41, 149–207.

17.

Deephouse

D. L.

Bundy

Tost

L. P.

Suchman

M. C.

(2017). Organizational legitimacy: Six key questions. The SAGE Handbook of Organizational Institutionalism, 4(2), 27–54.

18.

Dobbie

Fryer

R. G.

(2013). Getting beneath the veil of effective schools: Evidence from New York city. American Economic Journal: Applied Economics, 5(4), 28–60. https://doi.org/10.1257/app.5.4.28

19.

Dobbie

Fryer

R. G.

(2020). Charter schools and labor market outcomes. Journal of Labor Economics, 38(4), 915–957. https://doi.org/10.1086/706534

20.

Education Code Chapter 21 Educators. (n.d.). TEC 21.003. https://statutes.capitol.texas.gov/Docs/ED/htm/ED.21.htm

21.

Education Code Chapter 25 Admission, Transfer and Attendence. (n.d.). TEC 25.112. https://statutes.capitol.texas.gov/Docs/ED/htm/ED.25.htm

22.

Education Commission of the States. (2005). State class-size reduction measures.

23.

Education Commission of the States. (2017). Teacher license reciprocity: All data points for all states.

24.

Education Commission of the States. (2018). 50 state comparison: Charter school policies.

25.

Education Commission of the States. (2019). 50-state comparison: Minimum number of days or hours per school year.

26.

Elmore

R. F.

Fuhrman

(1995). Opportunity to learn and the state role in education. https://ncela.ed.gov/files/rcd/BE022357/Opportunity_to_Learn.pdf

27.

Fesler

Dee

Baker

Evans

(2019). Text as data methods for education research. Journal of Research on Educational Effectiveness, 12(4), 707–727. https://doi.org/10.1080/19345747.2019.1634168

28.

Figlio

Loeb

(2011). School accountability. In E. A. Hanushek, S. Machin & L. Woessman (Eds.) Handbook of the economics of education (3rd ed., pp. 343–421). Elsevier. https://doi.org/10.1016/B978-0-444-53429-3.00008-9

29.

Fryer

R. G.

(2014). Injecting charter school best practices into traditional public schools: Evidence from field experiments. The Quarterly Journal of Economics, 162(2), 1355–1407. https://doi.org/10.1093/qje/qju011

30.

Fuhrman

S. H.

Elmore

R. F.

(1995). Ruling out rules: The evolution of deregulation in state education policy. Consortium for Policy Research in Education, 708(2), 1–34.

31.

Fuller

Marsh

J. A.

Stecher

B. M.

Timar

(2011). Deregulating school aid in California: How 10 districts responded to fiscal flexibility, 2009–2010 [Research report number 2]. RAND Corporation.

32.

Goodman-Bacon

(2021). Difference-in-difference with variation in treatment timing. Journal of Econometrics, 225(2), 254–277. https://doi.org/10.1016/j.jeconom.2021.03.014

33.

Hammerstein

König

Dreisörner

Frey

(2021). Effects of COVID-19-related school closures on student achievement-a systematic review. Frontiers in Psychology, 12, Article 4020.

34.

Hanushek

E. A.

(1986). The economics of schooling: Production and efficiency in public schools. Journal of Economic Literature, 24(3), 1141–1177. https://doi.org/10.2307/2725865

35.

Hanushek

E. A.

(1996). Outcomes, costs, and incentives in schools. In Hanushek

E. A.

Jorgenson

D. W.

(Eds.), Improving America’s schools: The role of incentives (pp. 29–52). National Academic Press.

36.

Hanushek

E. A.

(2003). The failure of input-based schooling policies. The Economic Journal, 113, F64–F98.

37.

Honnibal

(2017). SpaCy: Industrial-strength natural language processing in Python. https://spacy.io/

38.

Jennings

J. L.

Deming

Jencks

Lopuch

Schueler

B. E.

(2015). Do differences in school quality matter more than we thought? New evidence on educational opportunity in the twenty-first century. Sociology of Education, 88(1), 56–82. https://doi.org/10.1177/0038040714562006

39.

Kane

Rockoff

Staiger

(2008). What does certification tell us about teacher effectiveness? Evidence from New York city. Economics of Education Review, 27(6), 615–631. https://doi.org/10.1016/j.econedurev.2007.05.005

40.

King

Ozler

(1998). What’s decentralization got to do with learning: The case of Nicaragua’s school autonomy reform [Research paper]. World Bank.

41.

Kuhfeld

Soland

Tarasawa

Johnson

Ruzek

Liu

(2020). Projecting the potential impact of COVID-19 school closures on academic achievement. Educational Researcher, 49(8), 549–565.

42.

Landers

R. N.

Brusso

R. C.

Cavanaugh

K. J.

Collmus

A. B.

(2016). A primer on theory-driven web scraping: Automatic extraction of big data from the internet for use in psychological research. Psychological Methods, 21(4), 475–492. https://doi.org/10.1037/met0000081

43.

Lankford

Loeb

Wyckoff

(2002). Teacher sorting and the plight of urban schools: A descriptive analysis. Educational Evaluation and Policy Analysis, 24(1), 37–62. https://doi.org/10.3102/01623737024001037

44.

Lubienski

(2003). Innovation in education markets: Theory and evidence on the impact of competition and choice in charter schools. American Educational Research Journal, 40(2), 395–443. https://doi.org/10.3102/00028312040002395

45.

Moe

T. M.

(2003). Politics, control and the future of school accountability. In Peterson

P. E.

West

M. R.

(Eds.), No child left behind? The politics and practice of school accountability (pp. 80–106). Brookings Institution Press.

46.

National Center for Education Statistics. (2019). Public charter school enrollment. U.S. Department of Education, Institute of Education Sciences.

47.

Preston

Goldring

Berends

Cannata

(2012). School innovation in district context: Comparing traditional public schools and charter schools. Economics of Education Review, 31(2), 318–330. https://doi.org/10.1016/j.econedurev.2011.07.016

48.

Texas American Federation of Teachers. (2021, April 1). HFT members and allies defeat district of innovation plan. https://www.texasaft.org/membership/local/houstonsupport/hft-members-and-allies-defeat-district-of-innovation-plan/

49.

Texas Classroom Teachers Association. (2017). Districts of innovation: The reality. Texas Classroom Teachers Association. https://tcta.org/node/14394-districts_of_innovation_the_reality

50.

Texas Education Agency. (2022, August 4). Every Student Succeeds Act. Texas Education Agency. https://tea.texas.gov/about-tea/laws-and-rules/essa/every-student-succeeds-act

51.

Wöbmann

Lüdemann

Schütz

West

M. R.

(2007). School accountability, autonomy, choice, and the level of student achievement. https://www.oecd-ilibrary.org/content/paper/246402531617

Supplementary Material

Please find the following supplemental material available below.

For Open Access articles published under a Creative Commons License, all supplemental material carries the same license as the article it is associated with.

For non-Open Access articles published, all supplemental material carries a non-exclusive license, and permission requests for re-use of supplemental material or any part of supplemental material shall be sent directly to the copyright owner as specified in the copyright notice associated with the article.

0.00 MB

0.92 MB