To answer the third hypotheses a multilevel modeling was used to avoid any independence of residuals assumptions. This statistical procedure was use since the data can be organized in more than one level. This analysis provided an additional tool to evaluate the data. In addition, the multilevel modeling was used to control for school random effect and for the diverse organizational characteristics. Because of the two types of job satisfaction results a multilevel modeling procedure was conducted. Various components were differently associated the two types of satisfaction.
This study utilized the appropriate statistical procedures to find the hypotheses. The data indicated that job satisfaction for teachers come