Why you can't dislike something on Facebook | Hacker News Reader

Why you can't dislike something on Facebook

benfrederickson.com

I don't think this is it. The point of allowing only Likes and not Dislikes seems to me a social thing rather than a usage thing.

Netflix lets you "dislike" something (implicitly using star ratings) because no one sees those ratings except you.

Amazon and Yelp let you dislike things because there is, for the most part, no interactivity around the ratings. A Yelp review or Amazon review generates no flame wars or ill will towards the user base.

So, at least from the examples in this blog post, that leaves Reddit, where downvotes frequently embolden trolls, causes large amounts of moaning about receiving said downvotes, and escalates flamewars as the downvoted have karmic reason to fight more and with increasing aggressiveness.

Disliking things has a negative social impact. It escalates disagreements and makes everyone simply observing the content feel just a bit dirtier for having to witness the ensuing melee.

Sure, it reduces the breadth of possible communication, but people go on Facebook and G+ to feel good and catch up, so a system where things can only be liked and it's hard to hate on something is preferable, even if it is unrealistic.

I don't think 'like only' reduces the breadth of communication that much more than simple like/dislike does already. Especially if only the net difference is shown (1000 upvotes - 999 downvotes says a lot more than 1 vote.)

When I see articles upvoted on HN I'm never quite sure if the number of votes is telling me 'good article', 'raise your awareness of this issue' or 'good discussion about this in the comments'.

    > [...]but people go on Facebook and G+ to feel good[...]

Very true, but to add to that another factor that shouldn't be underestimated is that it's companies advertising on these platforms that are footing the bill for their operations.

Fewer companies will want to advertise with you if their company/product page can be disliked. To name an example Monsanto has ~22k likes (https://www.facebook.com/MonsantoCo) they'd probably have at least two orders of magnitude as many dislikes if people could "dislike".

I'm not sure that's necessarily true. Sites like Amazon, NewEgg, EBay, and anyplace else with a ratings system for products and product sellers seem to do fine, despite the potential for getting negative reviews and ratings. These are sales channels rather than advertising channels, so maybe that makes the difference.

Contrary to the reddit example, I think downvotes here on HN work out decently. I've mostly only seen spam, off-topic, rude etc. comments downvoted here.

Downvotes on HN?

Yeah, when you get to some level in karma points (currently 500, I think), you're allowed to downvote comments. Comments with negative score will appear lighter/harder to read.

Yes I think this is definitely a more likely reason. Having "dislikes" on Facebook could also make people feel bad about things they've posted, which would make them less likely to associate negative feelings with the site, likely leading to decreased usage for some people. So it's in FB's direct interest to not have "dislike."

You nailed it - wherever people can dislike, downvote etc. there will eventually be a lot of flaming, mobbing and other bad behaviour that discourages people from using that website. We have a forum where you can like and dislike posts and some people actually spend a lot of time disliking posts by people they don't like systematically (up to 50/day are allowed). We're now considering a very low limit for dislikes/day ...

While the data is interesting, my understanding is that pales to: if your posts get Liked, you keep using Facebook, and if your posts get disliked, you stop using Facebook.

Some people think Facebook is a drama creation engine. That's a bit unfair: it's a usage creation engine, which does allow usage by people who are prone to creating drama. (Talk to e.g. teens sometime: for some people, implicit competition within real-life friend groups for who gets the most likes on their photos is A Thing.) Facebook is not unaware of social patterns like this, so they make product choices which make them less likely to negatively affect Facebook's interests.

> for some people, implicit competition within real-life friend groups for who gets the most likes on their photos is A Thing.

I didn't realise this until my sister told me recently that I was "getting pretty good at Facebook".

Upvoted, you're getting pretty good at Hacker News.

Here's an interesting social networking concept: the mixture of real names with anonymity. Your name would appear if you "liked" something, but dislikes would be anonymous. You could also have the option to post comments anonymously. Of course, bullying and drama would certainly be more prevalent, but nonetheless, it'd be an interesting experiment. (I also think people would be less afraid to speak their mind when a friend posts something incredibly stupid.)

To continue your thought experiment, I imagine that if it allowed plugins or extensions of some sort (like facebook apps), somebody would inevitably write a machine learning algorithm that attempts to deduce who (out of a given set of people) wrote an anonymous comment.

This is in sharp contrast to the rest of the ratings systems around the web...

That's because the like button isn't a rating system. The like button is simply capturing one subjective reaction (the desire to express your overall like of something with zero context) to provide a objectively comparable metric (number of likes). We're talking about apples and oranges here. Rating systems at least try to achieve feedback by providing context, likes do not.

Why people don't usually leave negative reviews - People don't dwell on things they dislike

None of the analysis here takes into considerations around the motivations of people to leave reviews in the first place.

people avoid things they will dislike, causing universally disliked things to remain unpopular. Its important enough that I thought I would sneak it in twice in different forms though.

Or maybe there are loads of hidden variables that are affecting this data. Like, for example, why people chose to write the review in the first place? I'd love for someone to do a scientific experiment to draw conclusions about humans motivation to write reviews.

Nobody sets out to write a terrible novel, open a crappy restaurant or create a horrible movie, and people mostly succeed in not being abysmal.

That chines restaurant down the street has value sets that people can't articulate correctly when describing their feedback (low price, quickly made, etc), and yet will get 1-2 stars on yelp for being "poor quality".

Recommendation systems are complex because we're effectively trying to scale subjective attributes into objective context. It will never perfect and will always be gamed. This debate will last for centuries...

> The truth is that people just don't dislike things online that often

You've got to be kidding me. From what I see, and maybe it's just the places I frequent, that's what the internet is for!

I think constructive criticism is underrated. Criticism often leads to improvement which most everyone benefits from.

As far as his data, sure most people can't be bothered to criticize something they don't care about. But imagine if Picasso hadn't criticized Matisse or Matisse hadn't criticized Picasso. It's not really popular opinion that matters, it's the few people that give a shit that makes all the difference.

Some of the discussion here may be PART the reason...but only a small part. If you really think about what Facebook and Twitter are trying to do, they're trying to construct 'graphs' -- a social graph or an interest graph, depending upon which suits your cup of tea; and graphs are based upon connections. When you "like" or "retweet" or "follow" something/someone, you establish and create a link in this gigantic web of connections. Think of it like developing a system of telephone wires.

Therefore, in this model, it doesn't REALLY make sense to have a dislike or a 'dis-retweet' option because you can't create a 'non-connection' more than leaving a void space between you and said other object. That's why these words aren't in the vocabularies of these tools mentioned.

At least that's the visual mental model that I have, and I would encourage others to consider this.

Hi, since I can't leave comments on you nice post, I'm doing it here.

* Thanks for putting up 'long form' posts which I guess what the kids are calling regular articles which take up more than 130 characters.

* Thanks for writing about your work, which I think more people should do.

* For the five star rating, if the mean is not 3 then the distribution can't be normal. It has to be skewed.

* Thanks for making the nice graphs, but please take pity on the color challenged and use much brighter and color blind friendly colors (or simply use gray scale, which would work just fine for your data).

Again, nice work on doing market research and putting up your findings for us to see.

I especially like that you develop a hypothesis at the end of your data description.

Just an observation: it is possible for the data to be normally distributed with a mean higher than 3, that just means that the distribution is truncated closer to the median on the upper tail than the lower one.

However, just from eyeballing his chart the data in this case do appear to be slightly left-skewed.

Thanks for the feedback!

You're totally right that ratings datasets will skew positive. I was trying to get really get across that they aren't bimodal at all (unlike what some other people have claimed) and tend to look somewhat normal if you squint hard enough.

Interesting for comparison could be Diaspora (http://en.wikipedia.org/wiki/Diaspora_%28social_network%29) github final comment on why they would not implement a dislike button (https://github.com/diaspora/diaspora/issues/1474):

[...]

I can't find the records of the past conversations we've had about this, but the final decision was this:

A like (or whatever) button is a way to show appreciation without leaving a comment. It is pleasant to get likes, it provides positive reinforcement, and it's simple.

Dislikes would do the opposite, it's like someone spitting on you and not knowing why. You can express your dislike in a comment, which is perfectly adequate.

[...]

Moreover being Facebook a commercial system, dislikes would have negative impact on its advertisers. So I think that this togheter with the above reasons made up for a like-only rating system. No need to look for more esoterical reasons.

But still the article was appreciated and it might spin up interesting conversations.

Anyway, I'd greatly approve a dislike button :)

decentralized social network program pump.io has a dislike button. But features like this are implemented as apps, and so you can implement these kinds of features however you want. In effect a like or a dislike is an activity post on your own account with a backreference to the content, and others who subscribe to your feed see this. You can't aggregate all data as in a facebook "like" system, but I see this as a feature rather than a bug.

Most things don't suck

Nobody sets out to write a terrible novel, open a crappy restaurant or create a horrible movie, and people mostly succeed in not being abysmal.

Not actually true. The abysmal stuff is filtered out for you by gatekeepers in publishing distribution, financing and so forth. If you want an example of how things look without gatekeepers, consider an unfiltered app store (like Android's a year or two back) filled with endless imitations and one-note ideas (Justin Bieber sliding tile puzzle! Miley Cyrus sliding tile puzzzle! Paul Graham sliding ti...you get the idea).

The approach of Facebook and G+ is that if you don't like something, ignore it and it will go away. This doesn't really work that well (see the other popular post about FB on today's front page, and the diminishing quality of Google News as more sources are added to it).

You can't dislike something on Facebook because users will feel discouraged. The whole point of Facebook is to share content with people you know and have content shared with you. By participating in the "negative" this goes counter to that idea.

Facebook is an identity/friend/family driven network, you wouldn't like seeing that John and Mary disliked your marriage. This works perfect in an HN/Reddit environment since you can dissociate yourself and provide pseudonymity.

As for Facebook Pages, maybe there is potential here. Though again unless you are a big brand, I don't know with certainty whether a post with 10 dislikes and faces attached to them is better than no likes at all.

The OP has an interesting way to look at it, though, I'm pretty sure this has been explained more than once by Facebook Engineers.

The problem with a full range rating (positive and negative) is that over time people will abuse it. We've seen this on an older web project (http://www.trendbuero.de/dk/). If you have a negative rating sooner or later little groups in your community will emerge and they will try to dominate everyone with their taste. They will call their peers and ask for downvotes. And when someone get's a lot of downvotes for a post it's unlikely that he will ever post again - he's scared away - forever. In a just positive rating system: If someone does not get any likes, he'll try again - maybe with different content. So traffic wise it's besser to just allow positive ratings.

I think this is shortsighted. The lack of negative rating is an issue. Even though social networks do not allow for them they should at least try to infer them as a 'hidden' or 'latent' variable (from the lack of sufficient amounts of up-votes after a large amount of views).

I'm pretty sure that G+ doesn't try to infer downvotes because of the large amount of sensationalism, alternative medicine, celebrity gossip in the "What's Hot" circle present in your feed by default. These type of posts would not rise very far on Reddit.

Also, downvotes are important when it comes to computing trust metrics in recommendation engines or buyer/seller rating frameworks especially if you want to be able to use transitive trust metrics (A trusts B and B trusts C therefore A can trust C). Negative trust ratings are useful in avoiding bad products or bad actors. The availability of these ratings would be quite valuable in helping with online transactions and give confidence to users the first time they buy from a new source.

Without the two data points, it is more difficult to determine the parameters of multinomial/dirichlet distributions used in, for example, subjective logic which can be used to build networks of trust. see http://en.wikipedia.org/wiki/Subjective_logic

A score is punishment or reward by consensus. I always thought the reason was psychological. There's a (very difficult to subdue) urge for a lot of people to punish those they disagree with; individual merits of the argument be damned.

I've seen it here, and it can't be discounted as just a downvote for tone. I think the wisdom of not being able to downvote someone who has replied to you really shines here (along with Slashdot's more extreme example of killing any mod points on a thread you're participating). It sort of forces you to be more objective before dishing out punishment.

The positive distribution may be skewed due to self-censorship to some degree. People tend to look at their score if it's displayed (one of the reasons, I wish "karma" was not shown and instead showed a Slashdot style label score for community standing). There may genuinely be more likable things than unlikable things since scoring of some sort with a very specific number is involved.

I'm more inclined to believe the negative review assessment. It is true that most folks just browse away to something more uplifting if the conversation or the story itself has become toxic.

I think the conclusion here is wrong. There may be fewer negative ratings, but it doesn't necessarily follow that they are less valuable for making recommendations. In fact they may be more valuable. If people go around indiscriminately "liking" random things, but reserve their "dislikes" for things they really hate, then "dislikes" might actually carry more useful information about their preferences than "likes".

On a previous ratings-based site I was working on, we ran into a big problem: distinguishing not caring from not liking.

Because to "like" something, or give it 4 or 5 stars, means you both like it, and it matters to you.

But if you're getting notifications in your newsfeed about a particular band, for example, giving it 1 star could either mean you hate the band's music, or it could mean that you've just never heard of the band and don't want to hear about it.

Depending on the site, you may want to do drastically different things given the two cases, but I've never seen any site that successfully distinguished between the two things in an intuitive way. (And giving 3/5 stars is not really the answer either.)

I must be a very strange/negative person, as I tend to downvote much more often than I upvote, be it on SE[0] or HN.

I also killfile people much more quickly than others, or so I was told. Maybe I’m just angry and hate everyone =)

[0] Not really using it anymore, though.

Part of the purpose of the 'like' button is obviously help form a more complete profile of a given user in order to more effectively customize advertising - would incorporating a 'dislike' button make this task any easier? That is, would an approach that refines advertising content by filtering out subject matter a user actively dislikes have a substantive advantage over an approach that just focuses on homing in on what the user actively likes? My guess is that it doesn't, otherwise it would have been implemented already.

Not necessarily true -- you can report/dislike ads anytime/anywhere on Facebook and Twitter (maybe Google+? I don't use it really)

Consider the burgeoning community of Redditors. Up-vote AND down-vote are integral to Reddit's success.

The article's hypothesis needs MUCH more work/research.

I think it would be interesting for Facebook to allow the user to set/select additional interaction verb-names. So you could have "dislike" button if the poster wanted it there.

My local newspaper often has stories that say "small child dies in horrific house fire" [like this on facebook], which seems.... inappropriate.

The lack of an option for the author to change the verb really does suck. People wanting to show empathy to someone's post on facebook about something shitty which has happened ends up in a long thread of sad faces, which kind of shit up any real conversation which takes place.

Being able to make a thumbs down available for select posts would solve that.

Haha, it's funny because if how accurate it is for me on HN. I see both up and downvote buttons, and it simply slows me down with a risk of pressing the wrong one. I never downvote comments I dislike, I just ignore them and upvote good ones. So I'd prefer to hide my downvote button.

There are some newspaper websites powered by Facebook comments. I wish I could downvote some of the comments I see there. I see some of the most terrible comments get liked a lot and the commenters become star commenters (or whatever FB calls them).

>I think the biggest reason is that even if they did offer a dislike button, people just wouldn't use it often enough to justify the precious screen real estate it would take up.

Ha! Let's test that, shall we.

I see I cannot downvote this story, either.

I guess this disproves the old "People are more likely to go online and say something bad, than something good."

The reason you can't dislike things on Facebook is that what Facebook calls a "like" is not a rating, it is actually a connection.

They are building a network graph, and the opposite of a connection is not-a-connection, which is the default state.

There's nothing "lower" or "worse" than the absence of a connection. This is sort of like the old joke that the opposite of love is not hate, it's indifference.

I wonder:

Whether dislikes, especially from an unknown walk-by, have a stronger emotional effect than likes. (And whether this varies by gender and anonymity of the attacker.)

What subset of posts tend to attract dislikes. (I could easily see questions and things like that attracting dislikes because they don't add general utility and then one dickhead comes by and decides to spit on someone for being ignorant/learning.)

Whether when people dislike something they're really just trying to make it less visible.

Though, since we're really talking about something's size in the search space, if people just ignore it it has more or less the same practical effect as it getting downvoted. Which makes the fact that something can get posted early in the day and be missed, and then be posted later by someone else and be seen kinda ironic.

I don't know, I've long thought that rating systems done on a per-post basis are really bad for promoting positive content. You don't get to customise it to show each user what they want, you don't get to cache any of your work really - and there seems to be something wrong with it, at least based on back of the hand personal eye-balling, that seems to make it wreck communities (i.e. since it's been becoming popular I've found less value in online discussion.)

Reply on news.ycombinator.com