ECMT Submission 3

Abstract Although parents want the best for their students they may not know what school is a quality school. By looking at enrollment and expenditure it becomes possible to determine how parents choose schools. The objective of this paper is to investigate the link between enrollment and per student spending on extracurriculars. Data on enrollment and spending of charter schools from the Texas Education Agency and population data is used to run an OLS regression. A positive correlation is found between enrollment and extracurricular spending however when controls are accounted for there is a loss in significance caused by multicollinearity bias. Statement of Contribution This paper was a solo project.

Introduction According to the National Center for Education Statistics, the number of charter schools in the United States grew by 2,500, while the number of traditional public schools fell by 2,100 between the years 2011 and 2022. As the popularity of charter schools rises, it is important to look at why charter schools appear to be more desirable. For this project, I am looking at how charter schools spend their budget on extracurriculars and how that impacts student enrollment. Clarity on what types of programs attract students can guide education spending on what is important to parents. Even if extracurriculars do not directly impact academic performance, there can be some unobservable benefit to participation that parents may consider when choosing a school. When parents choose a school for their children, many factors go into that decision, including a school's academic prestige, neighborhood, the types of students that attend, and other aspects of a school's programs. Parents may not entirely know what school is best for their child and may use facilities as a deciding factor. One of these facilities is extracurricular activities. While after-school programs are not likely to be the primary reason a student is enrolled, they can be a factor that causes a parent to choose one school over another. This is why it can be expected that there is a correlation between extracurricular spending and enrollment. My X variable is the per student spending on extracurriculars, and the Y variable is the number of students enrolled. There are many X’s that can affect enrollment, including the demographics of parents, the quality of surrounding public schools, and, importantly, the population of the city. It can also be true that the amount a school spends on per student

extracurriculars and the number of students enrolled. Conversely, parents who are higher income are more likely to have their child enrolled in one or more extracurriculars but are also more likely to be worried that their children’s schedules are overloaded (Pew Research Center, 2015). This could cause parents to overlook extracurriculars as an important aspect when choosing a school. Parents who would take the initiative to enroll their children in charter schools might already have their child in an extracurricular outside of school and not value the schools extracurriculars. Model Specifications The dependent variable for my analysis is the number of students enrolled at the school. The independent variable of interest is the amount per student a school spends on extracurriculars. The three control variables are the population, and the amount the school spends on instruction, and on athletics. x 1 is extracurricular spending per student, found by dividing total expenditure on extracurriculars by the number of students enrolled. A similar process is used for x 3 instructional spending and x 4 , athletic spending. Population is simply the population of the city in which the school is located. These three other xs account for omitted variable bias. The number of students enrolled, extracurricular, instructional, and athletic spending are all from the Texas Education Agency. The population is from the US census bureau. The data used from the Texas Education Agency is a panel dataset that contains information on all charter schools in Texas from the years 2007 to 2022. This analysis only considers the academic year of 2015-2016 and is a cross sectional analysis. The descriptive statistics are shown in the table below.

Variable Names Observation Mean Std. Dev Min Max Enrollment (y) 144 1131.06 1732.24 60 14031 Extracurricular (x1) 144 90.19 116.71 0 661.99 Population (x2) 144 915315.2 798986.3 814 2302878 Instruction (x3) 144 4834.31 1496.6 2627.77 14456.47 Athletic (x4) 144 37.40 68.01 0 384.5056 Because the Y variable is a count data and the model is an OLS. The OLS is run in Stata and the equation is shown below: Enrollment= B 0 + B 1 x 1 + B 2 x 2 + B 3 x 3 + B 4 x 4 The Null Hypothesis is H0: B1 = B2 = B3 = B4 = 0 and my alternative hypothesis is that at least one of the xs has a significant impact on the dependent variable enrollment or that at least one beta is not equal to zero. Empirical Results The scatter plot below shows the relationship between our X of interest, extracurricular spending, and enrollment. For this analysis two outliers are removed, both outliers spend more than 10,000 per student on extracurriculars. The scatterplot shows a positive correlation between enrollment and extracurricular expenditure.

This leads to a multicollinearity bias which is the cause of drop in statistical significance. This correlation shows that schools that spend more on instructional and athletic funding also spend more on extracurricular funding. Many schools that are well funded have high funding in all aspects. Regression 1: Enrollment= B + ₀ B 1 x 1 + ε Regression 2: Enrollment= B + ₀ B 1 x 1 + B 2 x 2 + B 3 x 3 + B 4 x 4 + ε Enrollment (1) (2) Extracurricular ( x 1 ¿ 2.90 ¿ ∗ ¿¿ (1.22) 3.08 (1.52) Population( x 2 ¿ 0.000043 (0.00018) Instructional Spending ( x 3 ¿ -0.142 (0.097) Athletic Spending( x 4 ¿ -1.032 (2.56) Intercept 869.72 ¿ ∗ ¿¿ (179.79) 1540.087 (569.81) R-Squared 0.0381 0.0553 Observations 144 144 There are several factors that also affect enrollment that are not accounted for in this model. Student performance has a large effect on whether parents choose to enroll their child. Information on student performance for the specific schools chosen was not available and therefore is not shown. Student performance, extracurricular spending, and instructional spending could also suffer from multicollinearity bias. On average students who participate in more extracurriculars have better academic performance which could lead to the two being correlated (cite). It's also possible that charter schools with worse student performance may hire

more skilled teachers to raise performance and spend more on instruction. Another factor that ideally would be included is the quality of local public schools and how much they spend on extracurriculars. In areas where there is more competition from public schools charter schools may spend more on extracurriculars to remain competitive and not have as many students enrolled when compared to areas that have less competitive schools. Conclusion This paper intended to shed light on the association between extracurricular spending and how it impacts school choice. It is important to understand what parents understand and value in their choice of schools. Using a linear OLS regression the effects of increasing expenditure were shown. The loss of statistical significance between the first and second regression can be explained due to multicollinearity bias. When investigated further the correlation between the chosen betas may have impacted results. While the initial findings of this paper are interesting, future research on this topic should consider controls that are not correlated with the X of interest in order to get a better picture of the issue. Works Cited United States Census Bureau. B01001 SEX BY AGE, 2021 American Community Survey 5- Year Estimates. U.S. Census Bureau, American Community Survey Office. Web. 8 December 2022. http://www.census.gov/. Abruzzo, K. J., Lenis, C., Romero, Y. V., Maser, K. J., & Morote, E.-S. (2015, November 30). Does participation in extracurricular activities impact student achievement?. Journal for Leadership and Instruction. https://eric.ed.gov/?id=EJ1097547