This isn't necessarily true. It's important not to conflate "small precinct" with "suburban/rural area" and "large precinct" with "urban area", since many rural precincts might have more voters than urban ones.

I wish someone would look at the underlying precinct data and redo this analysis while factoring in precinct population density.