1. We were discussing mj arrests and use. 2. I showed that there was indeed a positive correlation: more arrests positively correlates with more use.
Now, let's use your numbers from 1996-2004: there is still a positive correlation between more arrests and more mj use.
More arrests? Yes. A higher percentage? No.
With a public policy of marijuana arrests as the lowest priority (and the percentage of arrests did drop), there is a perception of a lower perceived risk. The number of users increased. The number of arrests increased, yes, but did not keep pace.