Psycasm is the exploration of the world psychological. Every day phenomenon explained and manipulated to one's own advantage. Written by a slightly overambitious undergrad, Psycasm aims at exploring a whole range of social and cognitive processes in order to best understand how our minds, and those mechanisms that drive them, work.
My posts are presented as opinion and commentary and do not represent the views of LabSpaces Productions, LLC, my employer, or my educational institution.
Please wait while my tweets load
Last week I dropped the comment that I thought it was normal and healthy for a blogger to check their stats. I was quickly corrected by more experienced bloggers than myself, who made the comment that the discussion generated in their comment boards was a more important metric.
With this, I agree. If I didn't want dialogue I'd write a book. But it seems intuitive that the degree of discussion is necessarily going to be linked views.
So, on my third day of holidays, I've decided to engage in some Recreational Statistical Analysis. There's nothing too fancy here, because there's not much fancy stats could have told me about my blog - with one exception (to my mind), which I will cover at the end.
But first, some descriptives:
Since joining LabSpaces I have made 34 posts, 22 of which were ResearchBlogging posts. I have an average word count of 889.26 (SD=338.78). Across all posts my mean views (as of 21st of December) 394.62 (SD = 253.84), but if you exclude the three outliers (Music Testing Athletes and Psi I, and II) then it drops to 340.97 (SD = 189.60).
Across all posts, on average, I get 8.30 comments (SD = 11.70); but if you exclude the outliers that drops to 5.65 (SD = 5.24) - as you can see it's entirely normal to receive no comments on any given post. The two means for average number of total unique commenters (whole pop vs excluded sample) is 4.06 and 3.68, respectively (SD = 2.92 and 2.76, respectively)
The ratio of comments I write to comments others write is 1 : 3.84.
I collected data on a number variables.
I assigned the Criterion Variables as: Views, Comments, and Unique Commenters (these values get subdivided later).
I assigned the Variable Factors as: Research Blogging posts, Number of Outgoing links, Media (Type and Presence), and Word Count.
Data I would like to have had: Number of incoming links, Number of Unique Hits per Post, and Time on Page
[The following correlation include the outlier values]
Number of Views is significantly correlated with Comments (r = .55, p>.001) and number of Unique Commenters (r = .53, p>.001). Comments and Unique Commenters is obviously highly interralated (r =.69, p>.001). Personally, I didn't expect the values to be quite so high, at an r value of .55, number of views accounts for 30.25% of the variance in number of comments.
Wordcount correlates with number of views, too: r = .38, p=.026. This was surprising, I would have thought a negative correlation would have existed where shorter posts were viewed more often, but apparently not. I would have been particularly interested in Wordcount correlations to unique hits and time on page. It's possible that longer posts are read in more than one viewing, and so may be skewing the results. Wordcount, however, is correlated with number of comments at r = .51, p=.002. These seems to suggest that the above possibility is not likely, and that longer posts just seem to get more attention.
I had to dummy-code the dichotomous Research Blogging variable, where +1.0 = Research Blogging Post, and -1.0 = non-research blogging post. Now this variable was the one that spawned the previous post (link). There exists a significant positive correlation between Research Blogging and Views, r = .49, p=.003; this data supports my assertion that personal posts get less views than non-personal posts, however, Research Blogging is not correlated with number of comments, r = -.13, p = .461, which supports Drugmonkey's assertion (one that I agree with) that number of comments may be a better metric of a posts impact. I do get more linkthroughs when it's a ResearchBlogging post - from their main site, from their twitter, and from a number of other aggregators, such as PsyDir. Yet, there is no correlation between ResearchBlogging and number of Unique Commenters (r = -.13, p = .461), and so if it difficult to know how well my ResearchBlogging posts are being received, how long people stay on the site, or if they decide to share the article.
With regard to the two other Factors, Media and Outgoing Links, there exist no correlations with any of the Criterion Variables. That's a little disappointing, but then I've always suspected that pictures and media never added that much to a science-themed article anyway.
Finally, I had always suspected that when the author of a post gets involved in their own comments, it's often the death of discussion. There was a significant positive correlation between My Comments in a thread, and the Comments of Others, r = .70, p>.001, which is completely uninformative. I'm not likely to respond to my own post, and will only respond to the questions/comments of others, so it's difficult to discern the direction of that relationship.
In order to determine whether my own commenting was a bad thing with regard to discussion, I conducted a 2-tailed t-test on comments made in threads where I didn't comment compared to a random sample of posts in which I did comment. Total number of comments not made by me were compared, t = .07, which falls well short of significance (2.26). So there is no difference between number of comments made, irrespective of whether I'm involved in the discussion or not.
One 'fancy' test I failed to conduct was a Hierarchical Multiple Regression. This holds other values constant and assess the true contribution of a factor to the criterion as a function of what it contributes above and beyond the other factors inserted. If I were to conduct this, it would have been holding word-length and research blogging constant, and it would have assessed what total views accounts for in terms of comments. I didn't because it's a little bit fiddly, and likely to be non-significant. But the point remains I could have, if I had wanted.
Apparently, not so many blogger look at their stats. Well, I do. Perhaps it's because I'm into Psychology, and without stats Psychology might not really be able to call itself a science, but thanks to stats I know a few (very generalized) things:
1. Research Blogging Good, Personal Blogging Bad (for me, anyway)
2. Longer = Better (views and comments)
3. I can comment without fear of destroying the flow of a thread
4. More views = more comments
This post has been viewed: 471 time(s)
I'm just commenting to say: 1st
Also, I want the minutes I spent reading that mathematical masturbation back. GIVE ME BACK MY MINUTES!!!!!
A very thorough analysis...I think it should be possible to put google analytics code into your blog page and then you would be able to get the data you'd like to have (Number of incoming links, Number of Unique Hits per Post, and Time on Page)
I might be able to help you with that.
An interest in this type of analysis is also needed to be a good marketer. And essentially you are marketing yourself in your blog, so it makes sense that you want to break it down into what works and what doesn't. It's all good information to have...especially if you want to make a living with your writing and podcasts one day.
I liked the summary at the end- thanks for that!
You definitely can add google analytics but it will really only give you data on ~25% of users. Adblock prevents the loading of google analytics data, but it's still beneficial for trends. Don't expect to get accurate raw numbers out of it, though.
I liked you analysis. I don't think it's too surprising that more views = more comments, but the more words correlation is pretty interesting. I would have thought longer posts would end up in fewer comments.
Maybe the problem with your personal posts is that people come to you for scientific information and you just haven't developed a following from people who really care about the person behind the writing. Blog consumers have different wants ;)
Dude, B may be on to something about what people are coming to read you for. Look at me, folks come to me less for scientific content but more about my own trials and tribulations. A lot of your posts are research blogging posts so that may be what your audience is accustomed to, so it may take some time for them to warm to your personal posts. Also, maybe you are writing your personal posts in the tone that you use for your research posts, it could be coming off as too stiff maybe. If you want to bulk up the readership on the personal side maybe write a post about some issue and seek feedback from the readership?
GR- I think what I love about your blog is how you express all the things that happen in the daily life in the lab in such a funny way. You're really funny- even if you aren't trying to be. But its also very real and I can relate to some of the experiences you are going through.
When I think about the blogs I read that are not about science but personal, I look for the ones that make me laugh. Like LabMom- she is so funny too- I loved her Big Brother series of articles- her writing about the episodes was more entertaining than the show itself.
So Rift- maybe when talking about your personal side, try to add some humor or develop writing with wit or sarcasm. I agree with GR that the personal stuff may be too stiff.
However, I do like your research articles a lot because it is completely different than my world and I also like to understand what makes people do the things they do. I'm still waiting for the article on hypnotism and the whole "Manchurian Candidate" thing- is it real or myth?
Most important is that you keep writing and develop a style and your readership will grow. You might see the stats change over time as you grow as a writer and develop a bigger audience.
I write about two things - career and work-life balance. I notice that I have two overlapping but distinct audiences for those posts. Of course, I'm using the comments as a basis for this observation, not analytics (seriously, I love math but those statistics made me want to cry for mercy). So it might just be that the same people read all my posts, but only those readers that identify with the topic comment. Hard to know if this is true, even with the analytics.
My point is, write about what you want often enough, even if it's not research blogging, and eventually it will gain it's own unique audience. It just takes time for a new reading/commenting base to build.
Woah, thanks all for responding. Unexpected. It's about me, and it's short - what an outlier...
...and 'mental masturbation' - haha, that's exactly what it is.
But more seriously, thanks for responding. I use Google Analytics a couple of other sites I run, and I've essentially found them less than useless. I do like how they can track what search terms were used, but I'm not generating enough traffic on any of the other sites to find them useful. Perhaps here they have a place - does anyone else use them here?
Secondly, I agree with the general sentiment that most people come here for what I can offer by way of psychological science, and not because I'm a personality. That's alright, it's that way by design, I guess the insight is that I don't have a big base of a readers. Props to becca though, for reading all the way to the end despite not enjoying it.
It's funny, I've come to love stats thanks to Uni. It can be pretty exciting pushing in data and watching the correlation matrix come back out. No, anybody else?
I'm commenting to improve your averages! ;)
For me, comments are nice, but I still check my stats. (I use feedburner and analytics on blogger) I also like to see where people have come from and how they are finding me (I definitely don't focus too much on the numbers.. I figure I'm a "niche" blogger!). Do my stats change how I write? eh, not really. Maybe it would encourage me to do a quick shout out on twitter when a new post is up but not really change the subject matter.
Honestly, I would rather have one or two heartfelt comments on an important post that meant something to me, than dozens of comments on a post that was merely comment bait. (The way lots of bloggers have giveaways, or ask a question to the readers etc.)
Stop on by, meander thru my thoughts.. say something if the urge strikes you, but I'm not waiting up at night worrying about if you agree with my latest post.
Did I forget to use the sarcastic font when I made that comment about nobody caring about stats? my bad....
I love stats too, but dude, numbers but no graphs or even tables? Want scatterplot of # page views vs. # comments, with linear regression and R^2! Would have taken many fewer of my precious, precious minutes (minutes I could be spending procrastinating on writing my thesis elsewhere!).
I am very annoyed I can't find the study right now, but it's not exactly something I believe because I've seen statistics on it- controversy breeds comments. I think longer posts lead to comments because it gives you more space to say something wrong.
Also, it should be obvious given that background, but I was *trying* to be helpful, in my own meta and obnoxious way. If controversy breeds comments, the best thing a commenter can do is pick an on-topic and intellectually honest fight. It's a little like trolling, but not evil. Like the shoemaker and the elves... elving?
*sigh*Alas, my controversy failed.
Props to you for being so gracious though.
It might be too late for you to make plans now (and I think registration is closed) however you might be interested in attending this conference for bloggers and people communicating science online next year. http://scienceonline2011.com/
Some of the topics might be of interest to you- here is the program:
Brian- I saw that you will be there so I look forward to hearing all about it.
Rift- back to your last question- I also love stats and finding trends. When it comes to my personal posts, I am not following the stats so much, and like LabMom, it wouldn't change what I write about anyway. At the other sites I write for, they follow the stats much more closely and it will influence what we publish. I prefer to write as I wish and not feel the pressure of having to get stats up by writing all articles a certain way. But I consider it good practice- freestyle writing vs. structured and "in the box" writing.
On thing to consider is the blog site itself. Blogs belonging to companies or publishers do not get nearly the comments that independent pages do, and the articles are far less controversial. They usually do not have new content every day- more like once a week. They are also usually just one person sites and not a collaborative network. Being in a collaborative site is an advantage if you want views and comments and you are trying to build a following. However, you may want to advertise your blog to students of psychology if you want more more controversial discussion from people in the field who know the research.
I am going to scio11, it should be a lot of fun! I'm doing a discussion with Mary Canady about biotech marketing too. It sucks that the conference is in North Carolina and Rift lives in Australia. It'd certainly be good for anyone in this space to check out conferences like Scio.