Cognitive modeling asks a different question than most learning analytics: not 'did the learner finish,' but 'what does the pattern of their responses reveal about what they actually understand.'
A learner who answers quickly and confidently is telling you something different than one who answers correctly but hesitates, or one who gets the right answer through a wrong reason. Traditional quiz scoring collapses all three into the same number. A cognitive model keeps them separate.
That distinction is what lets a system intervene at the right moment — reteaching a concept with a different analogy before a misunderstanding calcifies, rather than after it shows up as a mistake three modules later.