Date of Award
Master of Science (MS)
New Mexico has administered and funded State K3+ program to reduce the achievement gap between students in kindergarten through third grade since 2007. StartSmart K3+ project is an experimental research to examine the cost-effectiveness of State K3+. This research attempts to measure the efficiency of the schools participating in StartSmart using valuable information and data collected by StartSmart K3+. A Data Envelopment Analysis (DEA) originally developed to study production efficiency of micro-level organizations, and a regression model are used to analyze the efficiency of the schools participating in the first year of the project in 2011. The DEA is used to measure each school’s inputs and outputs ratio, such as teachers’ qualification and students’ performance, compares them and calculate the efficiency score. Efficiency scores generated by the DEA are biased by construction since the DEA constructs a lower bound on the true efficient frontier. Efficiency scores from the DEA are corrected using the bootstrap procedure as suggested in Simar and Wilson (1998, 2000). After generating DEA scores and correcting the bias, a regression model is used to identify the environmental factors that school may not control and affect schools’ performance. Two-limit Tobit with limits at zero and unity is used to estimate equations. Three performance measurements are identified as outputs: 1) average scores in reading, writing, math and vocabulary (each score is considered as one output and thus there are four outputs in total), 2) minimum scores in four subjects (four outputs), and 3) percentage of students with scores above 90 points in each subject tests based on the Woodcock-Johnson III classification (four outputs). Results suggest that between 50% and 58% of the schools were efficient in 2011, depending on the students’ performance measurements considered. Three Tobit regression models for three different types of outputs are estimated. Dependent variables are bias-corrected DEA scores and explanatory variables are education level of the closest caregiver, poverty rate in the school district and other variables. Results from the regression model tell us that education level of the closest caregiver is an important factor in explaining school efficiency. The time students spend watching television and playing non-education video games has a high impact in changes in school efficiency too.
Schools in areas with high-risk populations will require a greater share of resources to provide the same quality of education enjoyed in more affluent areas. The goal pursue by the Government of New Mexico of reducing the existing achievement gap between students will be limited by these inefficiencies. An efficiency evaluation could be carried out at the end of each summer session to identify inefficient schools and to better allocate resources. Short and long run policies should be implemented to increase schools’ efficiency.
Hedeman, Yamil Vargas, "Assessing Efficiency of Schools Participating in Startsmart K3+" (2014). All Graduate Plan B and other Reports. 371.