This study evaluates the technical efficiency and the changes in total factor productivity of public primary schools Kenya grouped into 72 districts. The approaches used are the data envelopment analysis (DEA) and DEA-based Malmquist productivity index. In addition, a second-stage tobit regression is estimated to determine the possible causes of inefficiency. Mean scores in the Kenya Certificate of Primary Education (KCPE) examination are used as output while gross enrolments, pupil-classes ratio and pupil-teacher ratio are used as inputs. In the second stage, a dummy variable to capture the location of the district is included in the model.The analysis of Tobit regression shows a negative significant relationship between Pupil Teacher Ratio is and efficiency thus high pupil teacher ratio is associated with inefficiency. Districts located in urban areas are found to perform better than their rural counterparts.