So, recently, there’s been a spat amongst some of ye olde freethinkers about sexism, again, and one of the participants — DJ Grothe — suggested that some people at some sites were being told to generate controversial posts attacking people to get more hits. While some people have been asking what his evidence for that actually is, Lousy Canuck, Greg Laden, and Stephanie Zvan have decided to take a more (ahem) controversial approach and try to argue, essentially, that there isn’t a strong link between controversial posts and page views anyway. Which, to start with, wouldn’t prove Goethe wrong about people trying it anyway, but it’s an odd thesis to say the least.
Before getting into their math, I want to outline the case for controversy. First, some empirical numbers. The post that has the most views on my entire site all time is my post on Rebecca Watson and Elevatorgate, which has over twice as many views as any other post on my site. Next is my post on Jerry Coyne’s criticism of a philosophical project funded by the Templeton Foundation that was picked up by both Jerry Coyne and Daniel Fincke, followed by my post about some complaints that seemed to not be about things that were unreasonable, my essay on the morality of psychopaths and autistics and rounding out the top 5 my response to calls asking theists what would convince them that God didn’t exist.
All of these do seem pretty controversial, or tie into existing controversies. But I’m going to go out on a limb here and say that simply being controversial isn’t what gives them their high view point (the Watson post still tends to top my view count week over week, BTW). If I look at where my views come from, most of them come from search engine results, with the next largest category being referrals from other sites. So, if you want to get page views, you need to hook into topics that people are talking about so that they might find your posts while searching for information or so that other people will link to your posts from their blogs, especially if those blogs are more popular than yours. Thus, I will argue, it is not the case that controversy leads to more page views, but the case that controversy can lead to the things that do indeed lead to increased page views.
Again, look at my numbers. Why is the Rebecca Watson post so much higher than everything else? People searched a lot for that term, and I wrote the post when the iron was hot and people were talking about it. That gave it an initial big boost of reads that it never relinquished. The “Fearlessy Amoral” essay comes up a lot in searches as it combines psychopaths, autism, and empathy, which are it seems to me fairly popular topics. On the flip side, the “Coyne doesn’t know philosophy” post, as far as I can tell, didn’t get a ton of hits from search engines, but got a ton of them from references from both “Why Evolution is True” and “Camels With Hammers”. People read those sites — which are far more read than mine — and clicked to see what I was talking about.
Thus, to get page hits on the Internet, you have to be findable. And, of course, with so many sites out there that goes without saying. One way to get that is to be controversial, even in a bad way. Even if many people post to say that you’re ideas are stupid on their blogs, a certain percentage of their readers will click to see what you actually said, and so your hits will increase. Make it a topic that someone is hearing about through vague E-mails or vague posts and people will also search for that topic and potentially read your posts.
So, what my first blush analysis would suggest is that while controversy — under the right conditions — tends to produce page hits, it won’t always do so nor will it be the only thing to do so. So, posting pictures of kittens might as well. Or a post — any post — about Skyrim (watch this one soar [grin]). If true, this would indeed mean that aiming for controversy is a way to increase hits, especially if you do it consistently; you might miss out on some that don’t take off, but you may have some others take off. Additionally, it’s also easier to hide trying to go for page hits by “picking fights” than by some other methods.
However, there is an issue here, which is how we can tell if a post is “controversial” or not. And that’s where we get into the analysis of our three erstwhile bloggers already referenced. They decided that the best way to judge whether a post is controversial or not is to see how many comments it has, and they all did various charts showing that there doesn’t seem to be a strong correlation between them all.
Now, I find myself highly skeptical of this definition of “controversial”. First, it seems to me that in general there will be a weak correlation between page views and the number of comments, simply because the more views you have of a page the more likely it is that someone will comment, even if only with a “Oh, that’s so true!” or “Oh, that’s so cute!”. But I’m not convinced that that correlation will be strong enough to generate any useful data, and so wonder if tying any of this to comments will simply result in misleading information (ie some high view posts will have a lot of comments, and some won’t, regardless of content). Second, I would expect the correlation between number of comments and controversy to be the same sort of correlation: present but weak. Looking at my own data, of my top five most commented posts only one of them matches the most hits: the one on Coyne, and only that because I got into a debate with someone on that topic. The highest hit post on my site and the one that seems to be the most unequivocably controversial — the Watson one — comes in at 7, with about half the comments of my most commented post. So, more controversial posts will have more comments — most of mine have no comments at all — but there doesn’t seem to be a strong link between controversy and comments. If people are going off to write their own blog posts on your controversial post, your comments will be low while your counts will be higher.
Thus, I don’t think their charts and graphs provide any data against the idea that controversy can and does generate page hits, because they use the number of comments to indicate controversy and the relation is not tight enough to do that. Because of that, their numbers come out wrong and so it isn’t surprising that they don’t find a correlation there. Unless, of course, I’m missing something; mathematics has always been my weakest academic subject.