Modeling Cost Overrun of Public Schools Projects

Poor cost performance in construction project is a common problem worldwide resulting in significant amount of cost overrun at account closure. This study aims at to develop model relationship between the amount of the cost overrun (CSTOVR) in schools construction projects and accepted bid price (LCO), estimated cost (ECO), contractor rank (GRA), experience of the supervisor engineer (ENXP), location of project (LOCA), year of contracting (YEAR), and contractual project duration (CDR) as input parameters before work starts. The study covers two story (12 classes) school projects awarded by the lowest bid system completed during (2007-2012) in Karbala province-Iraq. Probability distribution is established for each identified parameter and subsequently a simulation is developed to produces artificial data. The simulated parameters are used to develop the regression models to predict the cost overrun. It is found that the two developed regression models have the ability to predict the cost overrun (CSTOVR) for school projects, as an output, with good accuracy having correlation coefficient (R) of (85.9%) and (85.2%), determination coefficient (R2) of (73.8%) and (72.6%) respectively.