Hacker News new | past | comments | ask | show | jobs | submit login
Hedonometer (hedonometer.org)
206 points by mhb on Aug 15, 2019 | hide | past | favorite | 70 comments

I’ve found myself really wanting this, but for the impossible-to-gather complement of the data — how are people doing when they aren’t consuming social media and news media?

I find myself often wondering: is the whole world anxious and confused and sad right now? Or maybe just my country? Or maybe just highly-online folks? Or maybe just my bubble?

Because as far as I can tell, conditions in the world are better than ever before but I have all these signals telling me that I’m totally wrong about it: I’m surrounded by fear and anger and depression. But I don’t know how to calibrate at all, I have infinite information available to me and no way to sift through it, and the global marketplace of information seems to mean that anything that attracts attention is either highly biased or a statistical outlier or is otherwise just taking advantage of some human bias that makes it sensational enough to attract eyeballs and spread.

> conditions in the world are better than ever before but I have all these signals telling me that I’m totally wrong about it: I’m surrounded by fear and anger and depression

These are mostly orthogonal, no matter the scale. "Fear and anger and depression" are much more personality traits determined by genetics and raising, than they are reflections of how pleasant or unpleasant one's life is lately.

Due to hedonic adaptation (a.k.a. the https://en.wikipedia.org/wiki/Hedonic_treadmill#Happiness_se...), people mostly act how they act no matter what their environment is currently like. Short, sharp changes (like Christmas, as seen on the Hedonometer site) can make large numbers of people temporarily feel something strongly enough to show up statistically; but long-term environmental states (e.g. recessions, famines, wars) do not cause long-term statistical shifts in reporting on emotional scales. People get used to things, and revert to their emotional set-points.

In other words, a great world and a horrible world would be indistinguishable when viewed through the lens of "how happy people are." You can't get global utility from global happiness.

The evidence surrounding hedonic adaptation you're citing is misleading. The twin studies, for example, cite a 50% heritability, which, ignoring their problems, still suggests far from a preponderance of genetic influence. And those studies are difficult to interpret vis-a-vis their environmental assumptions, as are many other studies. E.g., it's not that impressive to me to show that people return to baseline from some anomalous event, when so much of the environment is structured in a very long-term stable way. The person who gets married still has to return to their same SES background, the same vocational history, the same environment, etc and so forth; conversely, the person who gets divorced still has the same job background, same workplace environment, and so forth and so on.

I also feel like the hedonic adaptation hypothesis leads to this kind of absurd conclusion that people are unaffected by anything regardless of severity. That is, it doesn't matter what we do to others because it's their fault how happy they are.

In any event, I think these hedonic adaptation studies are really misleading in some ways.

This project (the hedonometer project) is somewhat similar. I've done research on similar topics, and one problem you run into fairly quickly is that there's nothing really "attached" to the data you're modeling. The assumption is that the verbal content on Twitter will reflect people's mood on average, which is fairly reasonable, but the content on Twitter is fairly disembodied. That is, someone might tweet about how horrible Alan Rickman's death is, but they themselves are in a fairly good mood. So you see recurring time series patterns with, e.g., Christmas because the positive language is part of what is socially entrained, even though it's disconnected from changes in your mood. You tweeting about having fun with your family this year, when everything is going well, is not the same as tweeting about having fun with you family last year, when everything was going to hell.

It's admittedly a complicated topic, but I think this stuff gets radically oversimplified.

> I also feel like the hedonic adaptation hypothesis leads to this kind of absurd conclusion that people are unaffected by anything regardless of severity. That is, it doesn't matter what we do to others because it's their fault how happy they are.

I wouldn't say that this is an absurd conclusion, as long as you divorce "is happy" (as in, "frequently experiences happiness-qualia") from "has utility" or "is functional" or basically any other measure of QALYs.

Pretend that instead of talking about "happiness", we were talking about "sneezing." It'd make sense that how often you sneeze isn't a reflection of your long-term emotional state, right?

Well, why would you expect "how often you smile" or "how often you experience qualia of contentment" to be a reflection of your long-term emotional state? What privileges those physiological facts as telling you something about yourself, over the fact of how often you sneeze?

Consider: people cry while watching sad movies. This is an short-term emotional state evoked by the movie. It doesn't outlast the movie. Would you expect someone who experiences more movie-cry-sessions to have a lower long-term emotional state/utility?

> So you see recurring time series patterns with, e.g., Christmas because the positive language is part of what is socially entrained, even though it's disconnected from changes in your mood.

You suppose here that mood is a fact about our internal experience, rather than a fact about our linguistic rationalizing self-narrative of that experience.

If the "me" that tells you that there's an invisible dragon in my garage, and the "me" that goes into my garage anyway because I don't expect to get eaten, are essentially different people (one living in a land of social narrative; one living in a land of causal prediction and goal-directed action) which one of those selves would you expect to have a "mood"?

(It's not a trick question; I'm not really sure either. But it's worth pondering, I think!)

I agree about the dissociation between verbal awareness/insight/conscious perception about mood and other things, like quality of life or utility. However, what does quality of life matter if it is unrelated to emotional state? E.g., what does it matter what my quality of life is an economic sense (broadly speaking) if my mood is totally different? Putting that aside, there's also this kind of perception of state versus actual state issue which is maybe deeper or trickier to think through too.

I agree with what you're bringing up here, to me I think though it leads to more problems with the hedonic adaptation hypothesis rather than fewer. That is, if all these problems about happiness exist, then what are all those studies measuring?

It's a challenging thing to think through, and there's probably ways of addressing the OP's question about whether or not everyone is more miserable lately. I'm just not convinced the hedonometer approach is the best way (although I'm not sure it isn't).

>Because as far as I can tell, conditions in the world are better than ever before but I have all these signals telling me that I’m totally wrong about it: I’m surrounded by fear and anger and depression.

I have a hypothesis comprised of two ideas.

First, people have been encouraged more than ever to be expressive of their feeling and emotional states. That's the same reason there's a lot of self-deprecating humor with Millennials and Zoomers: it's a way to publicly cope with how you feel while also signalling to others that while everything isn't perfectly alright, let's laugh about it.

Second, people reacting to current events expresses feelings towards a particular thing. It is not necessarily an overall indicator of mood or mental health. I'm happy that it's my birthday. I'm upset that the other team won the championship. I'm sad that my grandmother died. Those things tend to be temporary emotions. I'm personally am upset with a lot of things the current US administration does, but it does not reflect my overall status. As I've seen explained: you can cure being alone by leaving your room and taking a stroll in a park; you can be lonely while surrounded by friends and family at your own birthday party.

I once reasoned that I trust my network of friends, family and colleagues to pass along any information to me that might be relevant to me - as long as I maintain a strong relationship with them - through regular meets, phone calls and texting each other fun memes.

I figured that I don’t need to consume any information from untrusted and unvetted sources (cable news, new york times, reddit, wsj, late night tv shows) - so I disconnected myself from all of the modern news sources and chose to live in ignorance.

Anecdotally, I feel great. I’m at peace, I still find out everything that’s useful - and I don’t waste cpu cycles in my head on things completely outside my sphere of control.

And yes, the whole world is doing really well right now - the way I see it - despite what the TV news anchors yell into the camera every hour.

It's somewhat sad that we choose ignorance to placate our anxiety. This has to be a symptom of feeling oppressed and under a tyranny that suppresses the people asking for change. People are apathetic because there is nothing they can do, so they turn off. The world is flooded with so much noise, no one can be bothered to figure out the signal any more.

It's perfect for government because they can now be elected and govern based on lies and no one cares. Lie your ass off to the remaining people who care to vote, say you'll do everything to fix everything and then just do whatever you want when in power.

No one can compete with that, so other candidates have to do the same.

Isn't this the sort of attitude that enables bad actors to abuse their positions of power?

Bad actors are already very much using the aforementioned media outlets to abuse their positions of power anyway.

I do agree it's a problem - being uninformed of what's going on around you. Not sure that being misinformed is really any better though. I do despair at our general ability to have an accurate sense of the world around us.

It's the sort of attitude that enables a LOT of bad things in the world:

The poster doesn't like "unvetted" news sources like the Wall Street Journal or the New York Times (which get lumped in with Reddit and Stephen Colbert in their opinion). Instead, apparently they consider the information they get from their friends and family as adequately "vetted." Now, that's a very unusual definition of "vetted," plus where are the friends/family getting THEIR info? The poster is information-gathering via the old game of "telephone" with all the attendant problems and distortion that invites.

Now, the poster also claims they are getting all the info they need. How do they know that? How do they know their trusted friends and family are 100% informed on all of the relevant issues?

The world is a fast-changing place so to safely not worry about any of those changes, you have to be:

1) near the top of the socio-economic pyramid, AND 2) devoid of empathy

Example: as a white natural-born US citizen, I probably won't have to worry about any upcoming changes to the existing legal permanent resident program personally affecting me. But a recently-approved permanent resident really SHOULD pay attention to the news so that they can learn that many Americans and their representatives want to revoke their residency and kick them out.

Same thing for marriage laws: I'm a straight unmarried man, so I won't personally be affected if the GOP re-bans gay marriage, invalidates all of the existing ones, and jails county clerks and pastors for officiating them in the future.

But as a human being who cares about the people who ARE affected, it would be incredibly selfish of me to ignore not just their plight, but to go out of my way to avoid learning about what is going on.

I feel that being an informed citizen is a critical responsibility, an integral part of citizenship.

> where are the friends/family getting THEIR info?

Are you trying to tell me Alex Jones and The Official Flat Earth Facebook group aren't "vetted" enough for you?

There is zero reason to believe your personal network are better vetted than the New York Times and a lot of reason to suppose that they are probably a lot less accurate. The average person is a moron barely functional beyond the things they have been trained by rote to do. Even their professional betters are often very good at some tasks and worthless at others. Remember our erstwhile presidential candidate who simultaneously was a capable brain surgeon and honestly believed that the Pyramids were built by Joseph to store grain if you are tempted to believe anyone has a universal trustworthiness score rather than a degree of capability on a particular topic.

The average mainstream source is edited multiple times for clarity and correctness, passes through multiple sets of eyes, potentially seeks the opinion of subject matter experts, has the names of people involved which one could follow up with and statements from them. The average casual conversation or share on Facebook has none of these virtues and often a multitude of sins including but not limited to being mangled in transmission and having the ultimate and unknown trustworthiness score of the original primary source replaced with an emotional analysis of the person presently passing it on.

A story from a person whom you wouldn't trust to tell you the time is ultimately judged based on the high esteem you hold for the person who bandaged your knees and baked you cookies and then passed on to others based on their emotional esteem for you or even their legitimate analysis that you seem smart and successful.

Extrapolate and you will began to understand why a pretty sizable minority don't believe we landed on the moon, believe that vaccines cause autism, and or are out there looking for fucking Bigfoot.

The real reason people prefer their in group is that their in group consistently reinforces their preexisting beliefs which is comforting and safe. Unfortunately everyone's in group has a strong bias towards people whom are like self in terms of beliefs, geography, and wealth. In particular if your in group is successful in a wealthy country you are likely to be well insulated from misfortune and will probably receive the collective "everything is OK" social signal long after the society you are riding into the ground passes the point of no return.

If you would pull your head out of the sand and tune into something outside of your bubble you would note some signs of real trouble. I'm wondering as I write this if there were people out there actually watching the Iraqi misinformation minister who crowed about crushing the American military in the streets on Iraqi TV while in fact their nation toppled and their people died. I think you should have a look around.

> The average mainstream source is edited multiple times for clarity and correctness, passes through multiple sets of eyes, potentially seeks the opinion of subject matter experts, has the names of people involved which one could follow up with and statements from them.

None of which says much about how relevant the material is to your life. Do I really need to know about how the republicans did whatever?

It bears on whether I'll be able to afford medical treatment for myself and my wife and whether I can keep breathing and she can keep being alive so ya.

For the vast majority of people, that information isn't really actionable. I maintain that politics is way overemphasized in media relative to its importance to the consumers of said media.

There is more actionable than you are likely to realize.

You can see how people on HN feel about given topics: https://hnprofile.com/compare?search=AWS,GCP%20|%20Google%20...

We're building the same thing for companies: https://insideropinion.com

I’ve found myself really wanting this, but for the impossible-to-gather complement of the data — how are people doing when they aren’t consuming social media and news media?

My study Mappiness gets closer to that. See, for example, https://www.youtube.com/watch?v=_Wo2M8Z-ULM

> Because as far as I can tell, conditions in the world are better than ever before but I have all these signals telling me that I’m totally wrong about it

I personally believe this is influenced by media consumption. I think “Everything is terrible” or “horrible thing happened” headlines are much more likely to get clicks than “world continues to gradually improve” headlines. Combine that with how consistently marketable outrage seems to be to just about everybody, and you end up with a world full of signals that everything is bad.

We are experiencing a global pollution event that has never been seen before. We are in the midst of potentially unprecedented mass extinction in modern times.

Can you consider the possibility that some things can improve while other things can get worse? That the cost of improvement in some areas, technological or the economy, will be a trade off in others like our enviroment?

Can you consider that as technology increases, so does the overall impact of humanity, so does the possibility of an extinction event. That there are more nuclear weapons today than ever before and the knowledge of this technology will only increase in the future.

I find your comment incredibly naive in its contrarianism. Oh sure, the media wants to scare you but that doesn't mean problems don't exist. This idea that everything always gets better with time is a ridiculous bias. The Dark Ages happened and it can come again. Progress is not a guarantee.

This kind of catastrophizing is exactly the sort of thing I’m talking about. I also think it’s quite arrogant of you to frame my perspective as naive.

You’ve made a single point about climate change, and chose to only focus on the worst way you could possibly interpret everything, including your own speculations about events that you think will occur. On the same topic you could also make points about renewable technologies constantly improving, and renewable adoption in the market. But the perspective you’ve put forward is that everything is all bad, all the time, and always getting worse. The reality is quite far from what you’ve characterised, and I think that one of the more concerning implications of this point of view is that good things shouldn’t be worthy of attention if bad things also exist.

> The reality is quite far from what you’ve characterised

Post some citations. I can easily support my claims. Can you do the same?

This could be an evidence of my bias, or an evidence of yours. Reality is not so clear.

> conditions in the world are better than ever before

Is it possible that some conditions are improving while some are getting worse? Technological or economic progress at the cost of the enviroment. Things are improving, until they suddenly aren't. Are you paying attention to the weather? This is not normal. We are experiencing a mass extinction event that may be unprecedented in history. Do not tell me that everything is fine.

To me this shows that Twitter is mostly US-and-Euro-centric.

This seems to be measuring the volume of messages with particular keywords on each day. I think that people wishing "happy birthday/thanksgiving/christmas/Valentine's Day" may not actually mean they are happy. (contrast this with a message saying "we're blessed with a new kid" which may denote actual mood.) Most of these generic messages are automatic greetings for particular days. I am not convinced that this actually measures people's moods on particular days.

The same the other way. The day bin Laden died is notably "unhappy" because a lot of people said "dead" or "death" but I do not recall it being seen as a negative event at all.

I saw the point where Paul Manafort received a guilty verdict, and it is also one of the lower points. My first thought was that it's still a negative situation, so even if people are happy about the outcome, overall it's still a drag. It still brings back all the things that had to happen to get to that point.

Also you have to wonder about some of the associations like on the day of the shooting in Dayton, OH [0] one of the few positive words is "white".

[0] https://hedonometer.org/index.html?from=2018-02-15&to=2019-0...

I don't this is necessarily a good measure of how happy people actually are. However, it is a good measure of how people are going to perceive the mood of society.

In other words, if most of what an average twitter user sees on given day is people wishing each other merry Christmas/happy valentines day/etc. then they're going to perceive that things are mostly going OK for people.

I think "happy <holiday>" is usually used as a shorthand for "have a happy <holiday>"

I think the effect you're describing is real, but... I turned off all of the days except for Saturday and Monday and it was interesting to see that with few exceptions (and those exceptions laying mostly around holidays), Saturdays were rated as having a higher "happiness" than Mondays.

Can’t they just measure which emojis people use?

This actually makes a lot of sense to me. Nice idea.

Yeah. I think an NLP solution would be more accurate in actually analyzing the sentiment of tweets.

Even then sentiment analysis is comically broken. I was playing with a tool that analyzed Tweets and plugged in a variety of words and phrases. It thought the term “white male” generally had a pretty strong positive sentiment, but on drilling into the Tweets, every single one that it coded positive was actually very disparaging. However, those Tweets were often overtly sarcastic and/or used terms like “privilege” which S.A. models believe to be an indicator of positivity. Some S.A. models think that the term “white” is positive and “black” is negative (probably not a racist conspiracy, but rather because they are too simplistic to understand the racial connotations).

Yeah I've yet to see any NLP solution that comes close to being able to understand sarcasm. It's not an easy thing to do though it requires understanding the whole culture a statement comes from and the message as a whole because of how much earlier sentences change the actual meaning of words.

I doubt we'll have something that really grasps sarcasm until we get close to full human level AI given how much it can trip up even people when communicated by text.

It clearly shows a giant sine wave and we're very happy on xmas. So... now we should juxtapose the stock market and see how things correlate. Only then will we be able to answer the age old question if money = happiness.

Edit: Also we're much much much happier with our mother than our father. This has soo much info packed into it.

>we should juxtapose the stock market and see how things correlate.

You shall find this relevant to your interests: https://www.sciencedirect.com/science/article/pii/S187775031... (by a former colleague of mine) they claim to get signals that lead the DJIA by 1+ days.

It's a big mistake to assume the people of twitter are a good sample of the people of the US.

I think the bigger mistake is to think even if it was a good sample it could gauge people’s mood - incentives for posting to social media are either happy or outrage. You will never get a silent majority “just okay” or “neutral-low” by averaging between ecstatic and outraged.

"The squeaky wheel gets the grease"

I think the fact that two of the top ~10 happiest days in 2018 were attributed to a Korean pop star's birthday is excellent evidence that people of twitter are not, in fact, good samples of the people of the US.

Also, the sadness when Trump was elected despite relatively even vote totals tells you that Twitter skews to the left.

80% of the population owns 6.7% of the stock market

but really, what about this sine wave? is this some kind of an error in data?

Looks like a typical seasonality curve. Things tend to repeat at the same period of the year (say, happy in summer, sad in winter). You can see similar patterns at different time scales (weekend vs week, day vs night etc)

but there is this huge ~4 yr cycle

I've often wondered, "Is it just me, or everyone and everything seems to be kinda awful right now?". This points to at least some validity to this idea, or at least as much as the real world reflects onto Twitter.

Is the measurement here flawed? Likely. But it's really interesting to see that some Christmases are indeed happier than others

My team has made a few comments about this lately. We deal with a lot of other teams and keep coming across levels of rudeness that we would probably not tolerate in our own teams.

It feels like something is in the air lately.

Social networking sites usually project perceived happiness vs Real happiness. It is hard to measure real happiness - people often wear diplomatic masks on internet to protect their real mood :) Having said that, I get the idea & this could be a great platform if we extend this to other emotions, political swings etc.,

I would believe that for any site except twitter.

If that site's developer is reading this, the top of the page can't be seen when the width is such that the upper links in the navigation bar wrap. For example, at the typical width of my browser, I can't see the years above the graph. I can't scroll high enough to see them.

One interesting observation: there’s a noticeable gap between average sentiment on Fri/Sat vs. Mon/Tue/Wed (especially when viewing the “Full” trend). Fridays rule—it’s science.

I wonder what accounts for the trough between ~2012-2014?

I imagine that an army of twitter bots sending automatically thanksgiving messages could mess with the results. A sad event in christmas would be balanced by the global cheering and be perceived as "less sad" by the program than in a common day.

The opposite is also possible. That bots can be programmed to create outrage deliberately is not a secret. Outrage means bussiness for newspapers so is a desired goal.

Why is the word "gun" listed as a "bad thing"? That seems biased just to make mass shootings seem worse than other disasters/loss of life.

Doesn't seem accurate from my experience. We should have seen a huge downward spike last night when Amazon Prime went down. ;-)

this is cool, overall kind of amazing to see how a whole ecosystem feels even if it's pretty generalized!

Can we get this data overlayed with the stock market? Correlations?

The stock market isn't driven by sentiment alone, and the population of investors isn't the same as the population in this study

What, no huge downward spike when tumblr banned adult content? :P

How does this correspond to economic data?

So basically humanity loves new year?

Twitter is at best a representative of people in the West. And only a small subset of those people.

Japan has a pretty large Twitter userbase, but this seems to only look at English words anyway.

I assumed new year would be something universal, because there was always a smaller spike right before it that I thought was christmas, so it wasn't big in the rest of the world.

The reasons why are complicated, but what we know so far is that the words "happy" and "new" and used excessively on this day, from which we conclude that people's moods are unusually positive.

Very interesting. The downwards spike of the Trump election is very similar to the upwards spike of xmas day. Nothing more, just an observation.

Aside from them both being spikes how exactly are they similar, and why would that be interesting? The same could be said about pretty much any spike

At a very high level the spikes seem to be of equal and opposite value. You are correct, there is a lot of variance in the diagram, which makes your comment very valid. My post was very lighthearted and by no means data driven.

Guidelines | FAQ | Lists | API | Security | Legal | Apply to YC | Contact