Public opinionators don’t seem to like comment threads. A couple of weeks ago, The Washington Post published Anne Applebaum’s “Another reason to avoid reading the comments.” Other pieces with similar headlines include “Just Kill All of the Comments Already,” “It’s time to end anonymous comment sections,” and “RESOLVED: Comments sections need to go.” Some Web sites are not just criticizing comment threads, but limiting commenting in various ways. In some cases, changes have been relatively modest, as in the Huffington Post’s 2013 decision to require that all posters’ identities be identified internally before they can post (under a pseudonym if they desire). But other online publishers, including Reuters and Popular Science, have indeed “killed all of the comments already.”
However, new research suggests that comment threads aren’t as bad as they’re painted to be. Most critics focus on the incivility of commenters, who frequently demean other individuals or groups and/or the post or article to which the comments section is attached. While there are certainly extremely offensive, even sickening, instances of incivility online, there is likely less than critics would have you believe. With several co-authors, I conducted a study of five popular independent political blogs (e.g., Townhall, Daily Kos) as well as blogs associated with two online newspapers (The New York Times, The Wall Street Journal).
Comments were coded as uncivil if they were clearly disrespectful toward another person or group. We found that approximately 10 percent of comments on the independent blogs (on average) were uncivil and only 4 percent of comments on blogs associated with the two newspapers were uncivil. Furthermore, almost all of the incivility classified by our coders was only somewhat, not highly, offensive.
We cannot claim that our estimates represent some type of global average. Even if we could claim such representativeness, levels of incivility will vary by outlet and time period. For example, another recently published study that analyzed comments on the Arizona Daily Star’s Web site found that 22 percent of comments were uncivil. That’s a lot more incivility than we found, but still only a minority of comments.
Of course, some have argued that even a small number of uncivil commenters will influence others in negative ways. For example, Applebaum writes, “Multiple experiments have shown that perceptions of an article, its writer or its subject can be profoundly shaped by anonymous online commentary, especially if it is harsh.” In defense of this statement, Applebaum mentions two studies, only one in specific enough terms that I was able to identify it. That study, titled “The Nasty Effect,” has been cited frequently as a powerful piece of evidence against comments sections. However, its findings have been exaggerated.
While the article is an excellent piece of scholarship, it does not show that “perceptions of an article … can be profoundly shaped by anonymous online commentary” (italics my own). The experimental study found that at most 3 percent of the variance in study participants’ perceptions of the risk of nanotechnology could be attributed to reading uncivil (vs. civil) comments (see Table 1 in the article). The largest experimental effect involved differences in risk perceptions between nanotechnology supporters and opponents, which widened by about 10% of the variable’s range after exposure to uncivil comments (see Figure 1 in the article). Importantly, the main (or direct) effect of incivility on perceptions of risk was statistically insignificant.
In fairness, many critics of comments sections wish to see reforms—such as an end to anonymous posting (see here on why this may not be such a good idea)— rather than a complete end to comments sections. However, in August, Kevin Wallsten and Melinda Tarsi argued at The Monkey Cage that, because an experimental study they conducted showed that trust in media declines after people read comments on media quality, “news outlets that care about their reputations (including The Monkey Cage) should shut down their comments sections.”
Besides the fact that good social scientific practice suggests we should replicate studies before taking such major steps, it’s important to remember that there are also pros as well as cons to comments sections. These pros include:
(1) People often have real conversations in comments sections. In our study, 45% of commenters referred to another commenter’s post, indicating that dialogue was occurring. Some say these conversations can take place elsewhere, like Twitter. However, most Internet users are not on Twitter (and the 140 characters limit makes serious conversation tricky). Furthermore, comments sections allow people to discuss a specific piece of writing they have all read (or at least, one hopes, skimmed!).
(2) Many commenters offer good feedback to authors. When a conference paper of mine was discussed in a Monkey Cage post last year, several commenters pointed out important limitations that I will address in a second, follow-up study. (The study deals with the effects of uncivil comments, so stay tuned.)
(3) Perhaps most important, commenters enjoy participating in comments sections and many report that they learn a lot from them. Applebaum’s op-ed had 1,719 comments at the time of this writing. After reading a sampling of those comments, I think it’s fair to say the vast majority of the commenters disagreed with the author.
This said, comments sections aren’t for everyone. Hosting a comments section isn’t cheap. It requires specialized software, and a human or two (or more) to manage it. Some sites are challenged more than others by trolls and spambots. But let’s at least take seriously the upsides — and not exaggerate the downsides — of one of the main places where anyone with a computer and Internet access can discuss important ideas and current events. Comments sections are admittedly not always pretty, but, then again, neither is democracy.
Elizabeth Suhay is assistant professor of Government in the School of Public Affairs at American University. She specializes in the study of public opinion and political communication.