This study develops and tests a school disorder and student achievement model based upon the school climate framework. The model was fitted to 212 New York City middle schools using the Structural Equations Modeling Analysis method. The analysis shows that the model fits the data well based upon test statistics and goodness of fit indices. The model accounts for 82% of the variance of student achievement scores on state standardized examinations. The study supports the model hypothesis that poverty and minority status of student population predict school disorder. Lower level of student SES is also associated with lower level of academic achievement directly, and indirectly mediated through school disorder. School disorder affects student academic achievement directly and indirectly mediated by student attendance rate. The effect of school size is in the hypothesized direction, i.e., smaller schools tend to have lower disorder and higher learning. Nevertheless, the effect is small and statistically insignificant, after controlling for student background and school culture variables in the model. The evidence of the study implies that school culture as reflected in school disorder and student attendance hold great potential for improving student learning. The reduction of school size, as currently used by many school districts to improve student learning, may prove to be ineffective if applied alone. In the large context and in the long run, reducing urban poverty will help close the gap between high and low achieving schools. (Contains 3 tables and 2 figures.)