Intro: I joined Hubski a month and a half ago, coming from Reddit. On Reddit, the separation in to autonomous subreddits mean relative isolation for each community. Sure, practically everyone is subscribed to multiple subreddits, but individual users rarely break the walls between subreddits. Some subreddits will link related ones on the sidebar, but these "webs" are very limited in scope.
Hubski has been described as a mix of reddit and Twitter. The action of following tags and users, within an overall smaller community I've found lends a sense of this just being a forum rather than a social aggregate site. A user may find themselves interacting with some people more than others, but given enough time, it seems one is likely to interact with practically everyone. This makes Hubski obviously unique from reddit or Twitter. A wonderful uniqueness for sure.
Another claim about Hubski I came across in my early experience was the lack of 'group-think.' Of course, as others were quick to point out, everywhere has 'group-think,' but Hubski's unique moderation system does allow a different approach to issue. By not having widespread consequences to being filtered or even muted, posters who may not be widely agreed with are still visible to the majority of users.
So, I ended up wondering how ''popularity" worked on Hubski. Just looking at my feed, it seemed like a lot of posts had either a few shares or a whole lot, but very few had a medium range. I was a (sort of) established user though, and I wanted to look at sort of the average experience.
Experiment: So, I made a second account, to try to replicate the average user. I believe I once saw statistics about the average number of tags, users, and domains followed, probably on the TMI page, but I haven't been able to find them since. What I remember them being though were about 15 tags, 20 users, and a handful of domains.
So, using the tags page, I followed the top 15 tags with the most followers, since I figured those would have the most meaningful activity, compared to the tags with the most posts. I figured there would be little movement in the tags that make up these 15, so I won't bother to list them.
The community page provided some more difficulties in choosing which section to use. There is no list of the most popular posters or commenters, so I decided to try split up the 20 users about evenly balanced. I followed the top five of the most badged, active posters, popular commenters, and active commenters. The most recent badged were left out, because of all the categories, this was the least likely to warrant a follow. Of course, there would be some discrimination with an actual feeling and thinking user, but I ignored this. Many users appear on multiple of these lists, and for those, I simply ignored them in their second lists. I did this around 4:00 Central Daylight Time on 7/27/2015. Looking now, a list has changed, as is to be expected, as these are more ephemeral rankings than the tags.
To preform the experiment, I was interested in looking at how many shares the posts had. Specifically, how the amount of shares was distributed across the posts. I did the first three pages of the second account's feed, by activity, rather than time, which would have skewed results towards the lower numbers of shares. Since this was rather limited, for an interesting comparison, I then did the same with chatter, again the first three pages, and by activity.
Results: For the Feed, formatted as: number of shares:number of posts-- 0:11 1:8 2:16 3:7 4:7 5:6 6:4 7:2 8:29 Total:90
For chatter, same format--
0:32 1:31 2:17 3:5 4:2 5:2 6:2 7:0 8:0
Conclusions:
I am in no way qualified to make any conclusions from these data points, and I realize that the study is very limited in scope, so I'd like to hear any other reactions and thoughts concerning either my conclusions, the set-up, or the data.
My preliminary observations about there not being many posts with a medium number of shares seems to have been pretty much right on. Posts with 5, 6, or 7 shares account for only 13.3%. Somewhat surprising was the jump at two shares. It is twice either 1 share or 3 shares, which seems pretty significant. I'd say that people are more willing to share once they see someone agrees with their opinion.
I was very surprised by the number of posts with 8 or more shares. Almost a third of the posts; more than the next two largest points together. Combined with the low number of posts with a medium number of shares, it becomes obvious that once a post takes off, it really takes off. On reflection, this makes sense given how the user following system works. As each user shares, the post appears in exponentially more feeds. Basically, a popular post is a rumor.
This reveals something deep about the social structure of Hubski. It isn't really a web, it is mostly a pyramid. There are thousands of users with just a handful of followers, if any, at the bottom, many of who follow the prime commandment of 'lurk moar.' There is then a pyramid reaching up, each tier consisting of users with more followers. At the very top is mk, whose posts are practically guaranteed to reach any active user. On the other hand, should the second experiment account ever post anything, it would appear on relatively few feeds, receive relatively few shares, and sputter out relatively quickly.
At the same time, 'web' attributes do exist. The pyramid doesn't work one way, users further up may follow users further down. The global page is a great way around the limits of the feeds. To me, though, the data points I collected indicate that in the broad sense, for better or worse, Hubski is a pyramid. Luckily, social mobility is a simple here. Several of the users the second account followed were less than three months old.
The main thing I gleaned from the comments stats is that it seems in general, votes for comments are more meaningful than shares of posts. On one hand, this makes sense, as a user may share a post they find interesting, even if they strongly disagree with it, while they would be much less likely to vote for a comment they are opposed to. On the other hand, this seems simply weird, because sharing with your followers should be a much more important action than voting for a comment. Maybe this disparity could be explained by the sheer number of comments (which, in hindsight, I should also have collected data on, but oh well, maybe if I do this again at a later date).
So, that's what I have. It leaves some questions and interpretations open, but, personally, I found it an interesting and worthwhile study to do in some free time. I hope you enjoy and find it interesting or useful.
Hmm. I haven't gotten into the numbers or anything, but I think there are other factors involved as well. From my personal experience, frequency of comments do matter. I'm not good at the whole data analysis thing, so I'll just put what I notice . . . My posts either do very well, or bomb horribly. There's not a lot of middle ground, though I do see that subject matter does influence their success rate. I have a decent number of followers (HI GUYS!) for such a young account. I've seen accounts much older than me don't have as many followers. I'm not saying this to brag though. I think the reason may be is that I comment more than some people. I've noticed that on days where I spend half my time in front of the computer, commenting away on Hubski, I'll gain 2 or 3 new followers. Sometimes I might lose one though (was it something I said)? The number of followers I have has done jack all for my post shares. It's still hit or miss. I don't post for shares though. I post what I love and what I would love talking about. Same thing for sharing. If I see something worth talking about, I'll share it. If I see something I'm dying to talk about, I'll share and comment on it. Lastly, ever since I've been doing social sites, so about 15 years now, I've always scrubbed my accounts and started a new one about every 5-8 months, sometimes ditching a site altogether for newer pastures (I don't miss you one bit FARK!). I'm torn on Hubski though. For once, I think my account is worth something. I don't know if I'll be able to bring myself to scrub it once it's time.
I agree a lot about the commenting. You make good intelligent comments, and just so you know, you were in the 20. And you bring up a good point, I should have discriminated between visibility (number of feeds it's in) and popularity (number of shares). And yes, you should try out alt accounts. For this one though, I unfollowed everyone, left a quick bio to say to ignore it, and logged out forever.
I could go through each post, look at the tags and who shared it, then sum all the followers for both. So technically, yes, but it would take so long, no significant number of data could ve taken in a time frame small enough that it wouldn't change pretty significantly while I did the sum. Nor would I know how to account for overlap. I'm sure there is some sort of code that could do it, basically a web crawler, which are used on lots of social networks, but I couldn't make one.
I like experiments. Thank you. This reminds me of something I once did on OkCupid. My real profile, I was honest. My profile indicated I was intelligent, creative, adventurous and literate, and also non-pretentious. Little iinterest was shown by other okcupiders. So as an experiment, I created a separate profile. Female. Straight. Found a photo online of a young asian woman, stock photo. Filled out my profile with empty, hollow phrases, no real personality, vague, superficial. This second profile got bombarded immediately with IM requests from hundreds of guys. (sadness)
Oh bleh, moderating OKCupid is the worst. I made an account to see how the site worked, then left it alone for six months. Then I was invited to be a moderator! Then I spent a few hours deleting pictures of genitals and... well, just like you said, arguing with other moderators about whether or not some creepy message from a creepy guy was creepy enough to be too creepy. I gotta say, though, they have some well-put-together moderation tools. I've been a volunteer moderator for several other sites and none have had such a streamlined moderation UX as OKCupid. It's really a power multiplier for moderators and a great example of how important UX can be to get the job done.
I'm with you. Never created a fucking alt. I have made an alternate account for other purposes, however.
A fucking alt. Interesting. Would this alt be one you created for you to fuck (a'la Ryan Gosling) , or do you give it independent will, and its entire design is to do the fucking all on its own, with whoever it chooses? I won't judge you either way, am merely curious.
Next time I would like to know whether I'm in a fucking experiment or not. IRB's exist for a reason ... and taking it off the reservation to use our collective behavior in an experiment without telling the subject is: 1. unethical
2. amoral
3. and go fuck yourself. But facebook did it ... I guess so can you!
I'm highly confused here. This kind of 'experiment' isn't something that requires a review board. This is simple statistics gathering from an already established set of data. It's amateur data gathering as well (no offence to jleopold.) This is like people watching at the mall and telling your friend that you noticed that teens tend to 'shop' in larger groups than the elderly. This is neither amoral nor unethical. I could understand the animosity if he had tried to manipulate content, 'game' the number of shares, or otherwise cause something to happen which wouldn't normally, but in this case he didn't do any of those things. Maybe I just don't understand, though. I'd like to hear your thoughts, though.
Woah. Hey. Wait a minute now. I'm curious, what's with the strong, negative reaction? What he did, at least from what I understand, doesn't seem that outrageous. All he did was really an observational experiment, seeing how following certain people or certain tags would affect what information he's exposed to. It doesn't sound like he actually did anything that would actually affect either A) the behavior of the people he was observing or B) the dynamics of the website. Should he maybe of said "Hey, I have an idea that I want to try out! Let me know what you think."? Yeah. Probably. To be fair though, this is probably the most benign experiment I've seen in a while.
It wasn't unethical and in fact informing his subjects they were being observed would negatively impact the quality of the study (from what I understand). Informing someone that you will observe them cannot do anything except impact their behavior. Even if they choose to act as they already were, the fact that it becomes a conscious choice instead of an unchallenged course of action would alter what the study 'results' would 'say.' In fact it's possibly a bit rash of backtoyoujim to assume he was even part of the experiment, other than by dint of the fact that he was using the same site as someone else who was doing a thing on that website. btyj is actually none of these, at this current time. I don't think observational studies (this is far more accurately a study than an experiment) can actually harm or impact a person's experience or behavior. If they do, then they clearly fail at being observational studies. I am all for people trying to observe behavior and patterns through publicly available and publicly taken information/actions, especially in such a benign scenario as "what is the behavior and pattern en masse"? I don't see how these studies can hurt the groups they observe really, so long as the observation is nonobtrusive. Now, watch-watch-watch, collect my data, and sell it? That of course is a problem and I'd be super pissed and so on. However... ...not what's happening here. Carry on with the navel-gazing, less so of the outbursts that non-private information was collected by a co-user/co-inhabitant/co-person to look at for group behavior/information out of sheer curiosity and theory-testing.I followed the top five of the most badged, active posters, popular commenters, and active commenters.
In this case, informal meant a high school student using public data to literally count dots while watching TV, which will have no ramifications, lacks all academic significance, released, nor used, any private data, and which made largely eclectic conclusions. Perhaps I should have asked first, but this is the equivalent of a student looking around the lunchroom and then sharing his observations with a friend. You do something similar everyday, whether you share your observations or not. If it makes you feel any better, based on your profile history (am I allowed to look at that?), it doesn't seem that you affected the study in any way.