TL;DR the study found "that the proportion of negative postings has decreased on the pseudonym-based forum after the law; whereas, the law was not influential on the website in which real names were being revealed regardless of the law." It gives little insight into what happens when a site switches from being pseudyonym-based to real named.
The Real Name Verification Law of 2007 required website owners verify the identities of users but did "not force websites to reveal [their] real names". Thus, it resulted in pseudonym-based forums that can now link to a real identity if needed and sites which always used real names but can now verify them.
On the pseudonym-based forum, after switching from unlinked pseudonyms, i.e. anonymity, to linkable pseudonyms "the proportion of bad postings clearly decreased...in both short-term and long-term compared to the control group, and they are statistically significant". The switch was "not salient" on the real name board that always used real names. Plus one TechCrunch.
The study readily states that the "proportions of bad postings [on the real name forum] are smaller than those [on the pseudonym-based forum]" across the board. This suggests a real name board will contain less vulgarity.
If one is worried about switching from unlinked to linked pseudonyms, note that the "composition of user groups did not change over the period regardless of the law".
We should also qualify articulating these findings to the United States given that "South Korea’s household broadband penetration reached 95%, which was the highest rate among those of all 57 surveyed countries" and that their political culture tends to be a mark more boisterous than we have here.
[1] http://www.computer.org/csdl/proceedings/hicss/2012/4525/00/...