Award Date


Degree Type


Degree Name

Doctor of Education (EdD)


Educational Leadership

First Committee Member

Teresa S. Jordan

Number of Pages



The purpose of this study was to determine the major factors that currently influence school transportation costs in the 89 school districts in New Mexico. The study also compared a linear regression methodology with an alternative neural network methodology to identify the most robust predictive model for updating the current transportation funding model for New Mexico Public Schools. Finally, the study looked at the redistributive effect under the two alternatives; The research design was implemented in three phases. In Phase One, literature was reviewed to determine factors that contribute to the cost of pupil transportation, to recognize fiscal models used in funding pupil transportation, and to identify requirements for judging pupil transportation programs. During Phase Two, detailed pupil transportation data were collected from all 89 school districts in New Mexico and analyzed. A statistical analysis consisting of simple and multiple regressions was conducted to determine which factors contribute significantly to pupil transportation costs in New Mexico. Additionally, a neural network was used to determine contributing factors and compared with the linear methodology. The two methods were analyzed and critiqued for robustness. In Phase Three, simulations were used to determine the redistributive effects of the two alternatives as compared to actual transportation allocations; Based on the results of this study the following conclusions were drawn. First, the factors that had the greatest effect on pupil transportation costs in New Mexico were: (1) linear density, (2) area density, (3) daily one-way paved miles per student, and (4) daily one-way unpaved miles per student. Linear density and daily one-way paved miles had the greatest effect on costs as predicted in the literature; Second, the alternative formulae developed in this study were both highly predictive of the daily one-way cost per pupil as indicated by their respective R squared factors. The regression formula was only able to incorporate three of the four independent variables, because of limitations in the statistical computer program. SPSS (SPSS Inc., 1999) automatically removed variables that had no significant change in the R squared value of the regression equation. The neural network formula was able to use all four independent variables identified in this study because of its tolerance of multicollinearity. One of the advantages of neural networks their ability to deal with ambiguous and overlapping data; Finally, the neural network's redistributive effect was less extreme in terms of winners and losers as compared to multiple regression.


Analysis; Funding; New Mexico; Public; Public Education; School; Transportation

Controlled Subject

Education--Finance; School management and organization; Transportation

File Format


File Size

3512.32 KB

Degree Grantor

University of Nevada, Las Vegas




If you are the rightful copyright holder of this dissertation or thesis and wish to have the full text removed from Digital Scholarship@UNLV, please submit a request to and include clear identification of the work, preferably with URL.


IN COPYRIGHT. For more information about this rights statement, please visit