I guess if it walks like at duck but smells like a rose, then it must be odd
Very interesting that your theory is that in IE Lesson records a question attempt when the student selects an answer and pushes the button (according to the book, the way it should work). However, Mozilla records a question attempt when they enter (or leave by any method) the question. The result being that max attempts apparently works differently.
Sometimes browser cache and cookie history can mess things up in the testing process for the teacher acting as a student or acting as different students (test1, test2 and etc). This is something I have been rediscovering recently. On the surface it makes the interface appear to be giving different results.
As an aside, I just discovered the Firefox addin for web developers. No instruction manual so I am still groping my way with it as an illiterate web developer trying commands. The "Cookies" pull down menu has a delete domain cookies, with another one for delete session cookies. The View Cookie Information in a Moodle site is interesting and down right scary when visiting other sites
Best Chris