Another reason came to mind. (And I realize this is all academic discussion at this point, that attributing is a thing which should be done, but still had a thought to share)
I am trying to decide what stories first published in 2013 that I want to nominate for the Hugo. 95% of my short fiction consumption comes from podcasts. Many of those podcasts not only allow reprints, but encourage them. So... How do I decide from the hundreds of podcast stories that I heard in 2013 which ones are eligible for Hugos? I can narrow down what I look at by starting with my Best Of lists for the year, but that just shrinks the overall pool, doesn't help decide which are eligible.
Toasted Cake (for example) attributes first publisher very consistently, with links to the source when available. If there is no attribution, then it was first published by Toasted Cake and therefore is eligible. If there is an attribution with a link or the name of the original publication, then I can follow that trail to find out when that was originally published to see if it was within 2013. Pretty easy, though a little repetitive.
For a podcast that doesn't attribute first publisher very consistently, I don't know if it's a reprint and if it was a reprint when it was first published. I then have to rely on Google-fu with the author and title to find a biblio page for the author or a search hit on one of the previous. It's not at all clear whether I get that right because, for instance, the reprint publication might take up the first 5 pages of the search results if the original publication was very old or not wildly popular. This becomes a potentially much more time consuming task, taking more work than it would really need to take.