In my code, I see this:
if (document.getElementById(\'xx\') !=null) {
    //do stuff
}
if xx element is not defined, will this         
        
getElementById is defined by DOM Level 1 HTML to return null in the case no element is matched.
!==null is the most explicit form of the check, and probably the best, but there is no non-null falsy value that getElementById can return - you can only get null or an always-truthy Element object. So there's no practical difference here between !==null, !=null or the looser if (document.getElementById('xx')).