In this research The study of Multi-level model (partial pooling model) we consider The partial pooling model which is one Multi-level models and one of the Most important models and extensive use and application in the analysis of the data .This Model characterized by the fact that the treatments take hierarchical or structural Form, in this partial pooling models, Full Maximum likelihood FML was used to estimated parameters of partial pooling models (fixed and random ), comparison between the preference of these Models, The application was on the Suspended Dust data in Iraq, The data were for four and a half years .Eight stations were selected randomly among the stations in Iraq. We use Akaik′s Information Criterion ,deviation statistic and Shwarz's Bayesion information criterion to compare between the partial pooling Models, The results show that the direct affect for the both degrees maximum temperature and the Rising Duston the Suspended Dust, where humidity was on a direct affect ( so increases the average monthly humidity will cause fewer occurrences of Suspended Dustin the same time the results show also the significant affect of geographical are as, and when the comparison between the three estimated models show that the Varying intercept -Varying slope Model is the better model .
This work is licensed under a 
