I figured it was probably sub 20%. 10% is still quite high, in my eyes.

There's some overlap, as one can be friends with coworkers and go to bars with them.

The paper aims to categorize how people met rather than how their relationship developed. This is how they describe their criteria for categorizing relationships (the paper is focused on internet dating, but the same criteria would be used for coworkers):

>By meeting online, or meeting through the Internet, we mean that a couple’s relationship began with an online interaction and then developed into a personal and physical relationship. We coded couples as having met online only if the online interaction was crucial to their having met, regardless of how the couple communicated once they had met. … If the couple first met decades earlier, fell out of touch, and then rediscovered each other through Facebook, that would be “meeting online” for our purposes.

Based on my reading of their criteria, the relationship would be categorized as 'coworker' if they either knew the person at work prior to visiting the bar, they were introduced at the bar through a coworker, or a work event took place at a bar.