A good introduction summarizes and motivates the question asked in the paper, and tells us what has been done about it and what is missing. In the process of this explanation, citations to the relevant literature are naturally intermingled.
But doing this neatly is often difficult. While probably a failure on the writer’s side, sometimes there is a relevant chunk of literature that must be aknowledged in the introduction, but does not naturally fit in the “story” one is writing.
What I do in these cases is to twist the story to include a reference to that chunk of literature. When this works, it’s great. But sometimes it does not quite work, and one ends up with a disorganized introduction.
Reading “A Theory of Credibility”, by Joel Sobel, I find a solution which maybe is sub-optimal, but which I find neat. He first writes a nice and focused introduction, with no references at all. Then he dedicates a single paragraph to explicitely acknowledge previous work:
Recent papers by […] were the first to present models in which […] My paper owes much to this work.
Not as elegant as other introductions. But it allows to keep the rest of the intro focused on what the reader needs to understan the paper, and not on what other researchers want you to acknowledge.