Electoral polarisation or polarising elections?

In last week's Polish presidential election runoff, Andrzej Duda got 10,440 648 votes and Rafal Trzaskowski got 10,018,263 votes. The percentages are 51-49. 

The explanation of this result concentrated on highly plausible factors like rural / urban, culture, party, and of course potential unfairness in terms of stance of state media and the incumbent government. 

But in fact, this knife edge result is strange. It's a bit remarkable that Poland is just 210,000 individual decisions apart on ostensibly fundamental issues (the number of voters that had they switched from Duda to Trzaskowski, it would have changed the result). 

The 2016 USA presidential election was even stranger, because of the thresholds caused by the electoral college. Looked at in terms of impact on electoral college margins, there were around 80,000 critical votes. 

In both countries, lots of people (around one-third in Poland and nearly one-half in USA) did not vote. 

So is it that Poland is really divided almost exactly 50:50 on the issues at play in the election, or that the election itself sets off its own dynamic that sorts the voting public (not the overall public) into these balances of power? 

We tend to think of elections as reflecting societal divisions, which they certainly do. But the messier possibility is that elections aggravate them. The median voter may be a very noise-susceptible voter.