This paper has estimated the school level technical efficiency of government aided lower primary schools in Hailakandi district of Assam based on primary data of a statistically robustly chosen sample of 95 schools. Selected non-input factors are modelled to explain school level technical inefficiency. A Cobb-Douglas stochastic production frontier with inefficiency effects is estimated. The estimated results indicate that both school infrastructure quality and teachers ‘contact hours influence the composite output index positively and significantly. Among non-input factors, frequency of mid-day meals during the survey week, and average BMI of interviewed children, have positive impacts on technical efficiency, while remoteness and lack of accessibility have a negative influence on the same. Reachability and accessibility are two vital non-input factors that influence school performance and efficiency. The study concludes that school performance has an enormous scope of improvement, especially if school infrastructure is enhanced by providing independent classrooms along with appointment of teachers up to the full capacity of five per school.