Kenneth Chang loooks into some research by Sinan Aral, a professor at MIT, and others, and learns that positive reviews of articles lead to more positive comments, while negative votes are discounted.
They [the researchers] collaborated with an unnamed Web site, the company did not want its involvement disclosed, on which users submit links to news articles. Readers can then comment on the articles, and they can also give up or down votes on individual comments. Each comment receives a rating calculated by subtracting negative votes from positive ones.
The experiment performed a subtle, random change on the ratings of comments submitted on the site over five months: right after each comment was made, it was given an arbitrary up or down vote, or — for a control group — left alone. Reflecting a tendency among the site’s users to provide positive feedback, about twice as many of these arbitrary initial votes were positive: 4,049 to 1,942.
The first person reading the comment was 32 percent more likely to give it an up vote if it had been already given a fake positive score. There was no change in the likelihood of subsequent negative votes. Over time, the comments with the artificial initial up vote ended with scores 25 percent higher than those in the control group.
“That is a significant change,” Dr. Aral said. “We saw how these very small signals of social influence snowballed into behaviors like herding.”
Meanwhile, comments that received an initial negative vote ended up with scores indistinguishable from those in the control group.
The Web site allows users to say whether they like or dislike other users, and the researchers found that a commenter’s friends were likely to correct the negative score while enemies did not find it worth their time to knock down a fake up vote.
The distortion of ratings through herding is not a novel concern. Reddit, a social news site that said it was not the one that participated in the study, similarly allows readers to vote comments up or down, but it also allows its moderators to hide those ratings for a certain amount of time. “Now a comment will more likely be voted on based on its merit and appeal to each user, rather than having its public perception influence its votes,” it explained when it unveiled the feature in April.
Duncan J. Watts, a scientist at Microsoft Research, said the overall findings fit with “cumulative advantage,” the idea that something that starts slightly more popular will build upon that popularity until it is far ahead of its competitors — and conversely, something that does not catch on will usually fade away whether or not it is good.
He cited the new crime novel “The Cuckoo’s Calling,” by Robert Galbraith, which received good reviews but tiny sales when it was released in April. When it was revealed that Galbraith was a pseudonym for J. K. Rowling, the book suddenly had the cumulative advantage conferred by the Harry Potter series and jumped to the top of best-seller lists.
“The biggest obstacle to success is just being noticed,” Dr. Watts said.
"We know we can learn more, and learn faster, by following the pointers made by others to the good stuff. The fact that the general rule can be gamed doesn’t decrease the leverage of following."
The web can be thought of as a network defined by links from one web page to another, or from comments and votes at the foot of an article to the article itself, or from votes and comments on curational sites — like Digg and Reddit — pointing back to the original article or post.
This sort of network has some well-understood properties. It is scale-free, which means that those nodes that already have more links have a higher likelihood of getting more links. Partly because of simple discovery — people follow links because they are looking for things to read — but also because of what the commenter might have said. This is the cumulative advantage that Watts mentioned. Likes beget likes.
We also know that plausible information moves more quickly through social networks: people discount hyperbole and the implausible. As a result, negative commentary simply has less throw weight. This can be stated as a universal: making positive assertions about a topic, individual, or event are more likely to be believed — in two ways — than negative assertions. The two ways: first, the assertion is more likely to be considered a true statement of the viewpoint of the person making the statement, and less likely to be motivated by less noble motivations. And second, the listener is more inclined to accept the positive assertion as a reflection of the truth, that the commented-on article is worth reading.
This is herding, very similar to the phenomenon caused by a person looking up in the street and causing others to stop and look up, as well. After all, there might be something going on up there, and it might lead to something falling on your head. Better to look up than to ignore the signs.
This can be gamed, as innumerable link-baiters have found, but also forms the basis of our social communications. We know we can learn more, and learn faster, by following the pointers made by others to the good stuff. The fact that the general rule can be gamed doesn’t decrease the leverage of following.
Perhaps it’s the end of the year that’s causing so much self reflection, so much concern about work/life balance, about information overload, about obsessive checking of our twitter feeds, about the value of disconnecting. Yesterday it was Daniele Fiandaca and Brad Feld, today, Nilofer Merchant.
In a fragmented world, go deep - Nilofer Merchant
It’s a fragmented world. And it’s only becoming more so. It used to be that when people wrote, they wrote more deeply. In the early days of the web (pre-twitter), I remember hand picking the few voices I would listen to and then putting them into my RSS feeder and checking for their essays. Essays, not tweets, were the way we shared what we were thinking. But as “content” has become more important to maintain a standing online, more and more people are entering into the fray. More and more people who may not even have a point of view to advocate but just want to participate in the conversation.
As content becomes more fragmented, you could try and compete with that by doing more and more, by curating other people’s content, by then running your content through Twylah, by having that “twitter magazine” come out which puts all your tweets and links in one place so that people can catch it if they missed each particular one.
Or you could do the opposite. You could go deep. You could be that voice that everyone listens to because when it speaks, it is so deep and rich that it’s worth slowing down to listen to.
Or, perhaps more importantly, you could chose to follow others who you think have gone deep.
As I said yesterday: Choosing who to follow is the single most important act in a connected world.
In a post from December 2010, I wrote
I have said for years that I’ve given up on finding a balance in life, I’m going for depth instead. But it’s not really the case. It’s just that I am looking for something larger.
Instead, consider the contour of a well-ordered humanism laid out by Claude Levi-Strauss:
A well-ordered humanism does not begin with itself, but puts things back in their place. It puts the world before life, life before man, and the respect of others before love of self.
So, for me, balance can’t be self-centered, it must be world-centered.
So I seek out people that consider that as a balanced mindset, and who go deep with that as their polestar, as a guide.
ifttt is quietly building an arsenal of powerful small tools that are making them the duct tape of the social web. Just as specific tools like hammers, chisels, saws and APIs are great in the hands of a skilled craftsman/developer, duct tape can fit the bill for connecting anything to anything for the numerous unskilled. The ifttt repository for Tumblr might be where David & the Tumblr Crew mine for clues/ambassadors as they begin to embrace the developer community to create tools for the masses.
I use ifttt to work around the inoperable import-posts-from-rss feature of tumblr, for example. I have a blog called Upstreamed which I follow here at stoweboyd.com, and I have set up a ifttt recipe for new posts of non-tumblr blogs I want to follow are posted to Upstreamed. Then the posts show up in my Tumblr stream, as if Tumblr supported the idea of following non-tumblr blogs.
Twitter is preparing to roll out a fairly significant rethinking of the user experience for the microstreaming service. They are planning to bring the social gestures that users make out in the open. These gestures are the actions of following people, favoriting tweets, retweets, or adding people to lists. Some of that gestural information has been available in Twitter to date, but most of it hasn’t been found in the stream along with the tweets themselves.
The change will come by changing the ‘@mentions’ tab into two:
Specifically, the “@Mentions” tab on twitter.com is being replaced by two new tabs: “@USERNAME” and “Activity”. These two streams will add an additional layer to Twitter and to Tweets themselves, a layer showing the social activity around them.
The @USERNAME (obviously, USERNAME will be replaced by your Twitter name) stream will still show your @replies, but it will also show things like when someone follows you, when someone favorites one of your Tweets, when someone retweets one of your Tweets, or when someone adds you to a list.
The Activity stream will show you all of those things, but related to all of the people you follow on Twitter. In other words, you can see if a connection has retweeted a Tweet, or if they followed someone new, etc.
Siegler doesn’t say that the current Timeline tab — which shows the tweets from you and all that you follow — will remain unchanged, but that is my interpretation at present.
Surfacing social gestures in general — and making favoriting a social preoccupation instead of a not very robust bookmarking tool — is a great way to make Twitter a richer social experience. In fact, this shift feels like Twitter has taken a long hard look at Tumblr, and has decided to capitalize on that social networked blogging platform’s success, which is driven to a great extent by the richness of social gestures, which are presented in stream. Here’s a snippet of my Tumblr stream, showing gestures and a post:
I wrote a piece not too long ago, What Twitter Could Learn From Tumblr, which focused on the efforts that Tumblr has recently put into its support of tags, and curation of tagged topics. (For those still not familiar with Tumblr, you might read Comparing Tumblr To Wordpress.)
But it seems like the social gestures of Tumblr — which are natively presented in the Tumblr stream — will be the first innovation to jump from Tumblr to Twitter.
I wonder if Twitter will take the ‘notes’ idea from Tumblr, as well? In Tumblr, all the social gestures associated with a post can be displayed on that post’s page (depending on the template settings). So If I post something that garners a great deal of interest — getting liked and reposted a great deal — there is a long series of gestures shown on that page. In a sense, the post has it’s own associated stream: all the gestures that it caused.
On Twitter this would mean that the page associated with a tweet — the one reached by clicking on the tweet’s timestamp — might show all the favorites and retweets tied to the tweet. Will have to see if this will be done.
And oh, there is still all that work to be done on tags, which Twitter still doesn’t seem to be very interested in, yet.
I saw MG Siegler’s post this morning about a new ‘Who To Follow’ user search feature on Twitter. I opened the Twitter page and there is was:
Twitter was recommending two people I might want to follow and currently have not been. There is a ‘view all’ link that takes you to a second screen:
and on this screen the rationale (or part of one) as to why I might want to follow, say, Jolie O’Dell is presented in the form of other people that are following her.
This is going to turn out to have results much like the recommended users’ list: those that have lots of followers will be displayed more frequently, which will simply accelerate the power laws.
Now, I am assuming that the ‘recommenders’ — those whose names show up as followers of the suggested users — are people known to me, which makes it a social analysis at least. But I would have to know something about their algoritm to find out if it does more than that.
For example, I might be interested in following more people in the design world, and fewer professional writers. Or, more directly, I might not to see recommendations of people that I used to follow but no longer do. Or I may want to follow people that follow and are followed by people from very different social circles from me.
At present none of this possible twiddling is made accessible to us, but certainly Twitter could wander in that direction over time, making it a much more useful tool for growing your network. But even in this preliminary state, I see that it will lead to a surge in following behavior over the next weeks and months, and an especially big help to newbies.