The answer is that the Iowa caucuses, to an even greater degree than next week’s New Hampshire primary, have always depended on a significant element of Make Believe. It worked because multiple actors, including or especially the news media, had a shared interest in playing along with the conceit that a small state without much demographic diversity is in better position to view the candidates up close and make judgments without outsized repercussions for the country as a whole.
One reason the gig lasted for decades was because of political reporting. There is an old (partially) tongue-in-cheek maxim about how to produce journalism: first simplify, then exaggerate. Generations of reporters applied those principles to writing about early-state results, investing them with significance that, once written or broadcast, was at least partly self-fulfilling.
Everything about Iowa this time around—the questions raised before voting about whether the caucuses had outlasted their utility; the problems after voting in tabulating results; above all, the narrowness of the results—suggests this is the year to stop investing the state with that significance. At a minimum it is appropriate to slow down, not necessarily to a full stop but definitely to half-speed, in projecting the Iowa results into supposed deeper trends.