Project Activities
Like many states, school districts in Texas approve capital investments through voter referendum. Using a regression discontinuity approach, the research team will compare districts in which a school bond measure narrowly passed to specific districts in which voters narrowly rejected a measure. This approach will allow the researchers to identify the causal effect of bond passage (and subsequent investment) on student performance. To conduct the primary analysis, data on Texas school bond elections will be merged with student-level and campus-level administrative data drawn from multiple state-level data systems.
Structured Abstract
Setting
Participants include students enrolled in public elementary and secondary schools in the State of Texas from 1994 to 2011. Researchers will focus on those enrolled in the diverse set of districts that held school bond elections (roughly three-quarters of all districts in the State).
Sample
The main analytic sample consists of the 5.5 million K-12 Texas public school students from the more than 820 school districts that held school bond elections between 1997 and 2009.
Intervention
The intervention to be examined is school facility expenditure (e.g., modernizing classrooms, renovating and repairing buildings, enhancing technology, and building new schools).
Research design and methods
A regression discontinuity design will use variation in school facility (i.e. capital) expenditures generated by the outcomes of close school bond elections to mimic random allocation of capital spending amounts to different districts. Using this design, differences in outcomes between districts where bond measures barely pass and barely fail can be attributed to the causal effect of bond passage and the subsequent increase in capital spending this entails.
Control condition
The "control condition" is the baseline state of school facilities in districts where bond measures narrowly failed.
Key measures
Student outcomes include standardized achievement test scores, school attendance, disciplinary actions, and high school graduation. Student-level data will also allow for the construction of district- and school-level outcomes that are not typically available, including means and distributions of achievement scores and attendance, overall and separately for subgroups (e.g., grade or economic status). Other school-level variables (e.g., teacher retention, peer composition) will be constructed from administrative data.
Data analytic strategy
Both static and dynamic regression discontinuity approaches will be used to examine the effects of bond passage on district-level expenditures, intermediate outcomes and student outcomes.
People and institutions involved
IES program contact(s)
Products and publications
Products: The products of this project will be estimates of the causal effects of school facilities spending on student outcomes in a diverse set of school districts in Texas. Peer reviewed publications will also be produced.
ERIC Citations: Find available citations in ERIC for this award here.
Journal articles
Martorell, P., Stange, K., & McFarlin, I. (2016). Investing in schools: Capital spending, facility conditions, and student achievement. Journal of Public Economics, 140 : 13-29. Full text
Supplemental information
Co-Principal Investigators: McFarlin Jr., Isaac; Martorell, Francisco
Student-level data will be obtained from the Texas Schools Project, housed at the Education Research Center at the University of Texas at Dallas. District- and campus-level variables will be constructed using the Texas Education Agency's (TEA) Academic Excellence Indicator System. Financial expenditure data will also be obtained from TEA. In addition, data from a census on school facilities conditions will be digitized to characterize the proximate effects of the intervention and as baseline conditions (prior to intervention) of individual school campuses. Campus-level data will be used to assess which campuses are most likely to be impacted by successful bond passage, how capital investment alters the physical infrastructure students that are exposed to, and how outcomes differ within districts. Student, teacher, school, and district characteristics in both primary and secondary data sources will be examined as moderators and mediators of the relationship between investment in school infrastructure and student outcomes.
Questions about this project?
To answer additional questions about this project or provide feedback, please contact the program officer.
