Thought the article was great overall. I would nitpick the jumping off point a bit though.
The polls appear to have been wrong by around 2-2.5 points. That’s actually closer than they were in the last election, and is a very typical amount of error. To ask polls to be any more accurate than this is a very tall order, and polls accurate to within 2-3 points are still extremely useful.
If someone thought this election was a lock then the error was looking at the wrong sources. It was not a failure of the system to make a reasonable prediction. 538 and the betting markets had Trump’s chances in the 20-30% range. Elections don’t have to get very lopsided at all to end up 90%-100% in favor of one candidate. Both sources deemed Hillary the favorite, but also clearly categorized the election as too close to call.
It’s a common hypothesis, but I think there are some questions to answer.
1) If social desirability effects were a factor, why didn’t they show up in the Republican primaries? Trump slightly underperformed his polling there.
2) If they were a factor, why were they such a small one? And how can we claim to see the effect in such a small polling error? Polls were more accurate this time around than 4 years ago, and I don’t think anyone thought there was a social desirability effect in 2012.