Are Christians Happier Than Atheists on Twitter?

University of Illinois psychologist Ryan Ritter just published a study in Social Psychological and Personality Science in which his team analyzed tweets from Christians and atheists — known as such because of the people they followed:

Ritter and his colleagues analyzed more than 877,000 tweets from 7,557 Christians, and more than one million tweets from 8,716 atheists. The believers were followers of one or more of five major Christian public figures (Pope Benedict, Dinesh D’Souza, Joyce Meyer, Joel Osteen, or Rick Warren); the non-believers followed one of more of five well-known atheists (Richard Dawkins, Sam Harris, Christopher Hitchens, Monica Salcedo, and Michael Shermer).

(For those who don’t do Twitter, Salcedo — a.k.a. @Monicks — is known for tweeting atheist aphorisms/quotations/etc.)

By looking at the kinds of words the followers used, Ritter confirmed what we might have expected:

The results: In their tweets, Christians expressed more positive emotions, and fewer negative ones, than their atheist counterparts. In contrast, the non-believers tended to use “a more analytical thinking style,” which, the researchers write, is “associated with less happiness.”

“Christian followers were more likely to use insight words characterized by certainty and emotion, whereas atheist followers were more likely to use insight words characterized by skepticism and analysis,” they report. “The percentage of words expressing certainty was higher among Christian tweets than atheist tweets.”

It is pretty fascinating to see the words in that visual. Christians used the word “know” 211 times per 100,000 words, compared to 198 times per 100,000 words for atheists. Meanwhile, atheists used “thought” more than Christians did (59 times per 100,000 words compared to 44 times per 100,000 words, respectively). But without context, we really can’t jump to any conclusions about why those words were used.

I don’t doubt the conclusions. But just for the sake of it, let’s ask some important questions:

Were the researchers right to use the people they did as representatives of who Christians and atheists would follow?

Calculating whether someone is Christian or atheist based on who they follow on Twitter isn’t even close to a perfect science. I follow the Pope. I followed Rick Warren until he blocked me (yep). And many, many Christians would tell you that Meyer and Osteen and D’Souza practice a version of Christianity that perverts the Gospel in some way or another. Ritter acknowledges all of this, but still believes the method works because of the sheer number of tweets and followers analyzed:

people can follow these public figures for reasons wholly unrelated to their religion. Despite the imperfect nature of this sampling method, the large-scale nature of Twitter data appears robust.

The atheists are more representative, I think. But another small issue: the @ChrisHitchens account that was analyzed is neither verified nor very productive — posting 10 tweets total, most of which were links to videos on a now-defunct website. Make of that what you will.

Also, given what Dawkins and Harris tweet about, they’re more likely to draw people who are generally more pissed off about religion (i.e. not “happy”), as opposed to casual atheists.

For what it’s worth, the researchers said that, of the thousands of people they analyzed, “[t]hirteen followers… were following both a Christian and an Atheist public figure in our sample, and were excluded from the final analysis.” So people like me wouldn’t have been included in their sample.

What should we make of the fact that Christians seem happier on Twitter?

Well, I suspect that, much like you’d find in the blogosphere, atheists on Twitter tend to be defensive, reacting to events that make us upset/angry, while Christians tend to write about their relationship with God and other things that make them happy. These are huge generalizations, of course, but it’s what I’ve experienced.

We’ve seen studies that show religious people have a higher level of wellbeing than atheists — but we also know that wellbeing improves when you’re part of a close-knit community and are frequently given reasons to be optimistic. Churches provide both of those things and atheists have to essentially fend for themselves or create alternatives that are nowhere near the scope or size of what religions provide.

Do the words we use accurately reflect our happiness levels?

This is tough to say. Atheists are (in theory) skeptical and more interested in what’s demonstrably true than religious people are, even if those answers don’t always make us happy. Our word choices reflect that. Of course we’re not certain about everything. Of course we’re less likely to be swayed by emotion. But atheism isn’t synonymous with depression.

Ritter even alludes to this in his analysis:

Indeed, non-religious people are equally happy as religious people in non-religious nations (i.e., where they fit in…), and increasing the perceived prevalence of atheism can decrease anti-atheist prejudice… In other words, increases in happiness among non-believers should parallel increases in the availability of secular social support resources and increased feelings of being respected in society, both of which facilitate increased happiness.

Given all of this, if there was a more accurate way to discern who’s atheist and who’s Christian on Twitter, I don’t think you’d see very different results.

Just bear in mind that the link between happiness and religion should not be seen as justification that there’s something “right” about religion. The fact that religious people seem happier doesn’t suggest that their beliefs are true, but rather that their beliefs can be powerful placebos.

(Thanks to @l_chambers for the link!)

About Hemant Mehta

Hemant Mehta is the editor of Friendly Atheist, appears on the Atheist Voice channel on YouTube, and co-hosts the uniquely-named Friendly Atheist Podcast. You can read much more about him here.

  • SushiSquee

    Numbers might be a bit skewed to start with; “believe”, “belief”, and
    “believes” are apparently three separate categories. I get that it’s
    more specific but dividing a word from its plural? Really?

  • C Peterson

    I’ve got to think that anybody with enough tweets to make a statistically useful assessment of their happiness is screwed up, and I wouldn’t trust the results!

  • good_creon

    I was always under the impression that the only thing you could gauge with Twitter was how narcissistic someone was.

  • Art_Vandelay

    In contrast, the non-believers tended to use “a more analytical thinking style,” which, the researchers write, is “associated with less happiness.”

    That seems dubious at best.

  • eric

    Do the words we use accurately reflect our happiness levels?

    That’s the assumption that bothers me (the most). I see no prima facie reason to attribute more happiness to someone that sends more emotional texts and words like “know,” or attribute less happiness to people who text the word “thought.”

    Maybe there’s some earlier psychological work that supports that assumption, but without being aware of it, I’m skeptical that their conclusion follows from their data.

  • JohnJay

    The fact that a believer is happier than a skeptic is
    no more to the point than the fact that a drunken man
    is happier than a sober one. – George Bernard Shaw

  • Sven2547

    Interestingly, the language I see here is not a contrast between theism and atheism, but rather the contrast between gnosticism and agnosticism: “know” and “feel” versus “think” and “believe”.

  • edgar ayala

    Didnt Bertrand Russell say something about christians being happier than atheists is akin to a drunkard being happier than a sober person. Yeah, that.

  • Jansen Waddell

    Someone commented earlier, ’twas George Bernard Shaw

    But yeah, that.

  • smiles

    Ignorance is bliss… People who turn a blind eye to tragedy and wrongdoing, would (presumably) be seen as happier. Also, being part of a minority group that is a de facto “enemy” of the majority, does not lend itself to being all too gleeful.

  • Marie

    For me, my ‘more analytical thinking style’ MAKES me happy. Even when it’s not about necessarily happy things, I much prefer being logical and analytical about what’s really there.

  • jediofpool

    This makes me so mad.

  • Gus Snarp

    I think it far more accurately shows the difference in thought processes between the two groups than the happiness. Sounds like they’re reaching quite a bit beyond the power of the study. The real problem with the sampling is the enormous numbers of both atheists and Christians that are left out. These are very particular subsets of Christians and atheists and there’s no reason to believe it’s really generalizable beyond the sample.

  • Art_Vandelay

    Absolutely. Skepticism is not cynicism and knowledge is emancipating.

    Do any of you folks who used to be religious feel as if though you were happier before you were an atheist?

  • Jasper

    I’m happier when I complain, anyway

  • Blacksheep

    The design of the chart is a bit dramatic. Note the large blue circle that represents use of the owrd “Know” by Christians. The numbers for the word “know” are 211 by Christians and 198 by atheists – pretty close, but not judging from the big blue ball.

    I once read a similar study between liberals and conservatives that found that conservatives reported being happier, and the study concluded (among other things, including faith) that it was because (paraphrasing) “conservatives are realists and see the world as a tough place, so hardships are expected, while liberals are idealists who want a better world therefore are continually dissapointed.”

    Story on the topic:

  • Rev. Achron Timeless

    It seems like they were just going with “ignorance is bliss” but needed to make it sound more methodical.

  • CBrachyrhynchos

    Oh, good grief. This has the same flaws as the atheism-autism study that made the rounds last year.

    1. Computer-mediated communication needs to be treated as an unrepresentative data set unless you have a way to independently test the demographic characteristics of the people participating.

    2. Using linguistic features of computer-mediated communication to make inferences about psychological states without dealing with the more immediate variables of medium, mode, pragmatics, and sub-group linguistic norms doesn’t make a lot of sense to me.

    Yeah, I think this is a weak conclusion built on the wrong type of data for that conclusion. But actually trying to identify a representative set of non-theists and recruiting them to participate in a study using an survey instrument about “happiness” that is the product of some work to test its reliability and validity would be too much like work, therefore let’s mine the twitter API! Brilliant!

    (Is a bit biased because shitty datamining studies like this make those of us who have done serious work on CMC analysis look bad.)

  • Reginald Selkirk

    Do people who are really happy waste their time on Twitter?

  • anon

    Very misleading to associate happiness with word choice – saying I’m happy a thousand times may increase the social perception of your happiness which can increase your sense or illusion of what happiness means… but the words are not a meaningful way to assess our different degrees of happiness which is a meaningless word… love and fear – expectations of pain and pleasure .. a foundation created before belief or money that we are safe and life is positive.. these play a much larger role than indoctrination into superstitions which lend themselves to undermining the safety and security of others … which begs the question happy now at what cost later…

  • Rain

    I followed Rick Warren until he blocked me (yep).

    I thought he loves everybody. I guess he’s full of crap. That’s always great when people are full of crap. Actually no, it’s kinda sad.

  • CBrachyrhynchos

    Even if the authors have that research in hand, correlations are not transitive. Even if you show that A ~ B, and another study claims B ~ C, you can’t claim “therefore A ~ C.” Especially not when B is such a complex and multi-factorial phenomenon like human language.

  • linimalD

    Absolutely. I had a close-knit community and family, a ready-made social calendar, a sense of earthly purpose and the ‘knowledge’ that no matter how bad things got my heavenly father loved me, would look after me, and that I’d get a happy ending – an eternity of bliss. There was regular singing in church (singing out loud being something that demonstrably lifts the mood) and an atmosphere of encouragement. On top of it all, I had the satisfaction of ‘knowing more’ than people who were hostile to me personally for whatever reason and those who challenged my beliefs. When personal circumstances forced me to re-examine my beliefs more carefully, the whole house of cards came crashing down and I felt I had been ROBBED of all those things that had contributed to my happiness. I was bereft of the social supports I’d been accustomed to, including family, I had to face a world that seemed ultimately pointless and utterly alien to me. Knowing that I now have a greater faculty for critical thought is poor compensation. Perhaps studies of this kind should be more rigorous in sorting units of analysis from the sample into people who’ve become religious, those who were raised religious and remain so, those who were raised religious but who have renounced their faith and those who were raised non-religious?

  • blasphemous_kansan

    Mine is armchair commentary to be sure, but given that 1,000 words is only %1 of 100,000 words then how can we say that the appearance of a word 44 times out of 100,000 is a statistically significant difference from a word appearing 59 times per 100,000 words? Out of the total 100,000 words the researches are extracting frequency analysis from words that make up less than one tenth of one percent of the total words, so I’m wondering how can this be a useful conclusion when we toss out 4 orders of magnitude from the sample?

    I would wonder what the distribution of some common control words (to which the researchers assigned no ‘religious’ weight) looked like between the two groups. I still doubt we’d have a workable conclusion then, but at least we’d have some idea as to what the noise looks like for word distribution per 100,000 words for the tweets in question, and we’d be better able to reasonably determine whether the distribution of the ‘weighted’ religious words defined by the researchers truly fall outside the realm of random distribution.
    Maybe the researchers accounted for this somehow, but sadly the study seems to be behind a paywall.

  • wmdkitty

    Well… they do say that ignorance is bliss…

  • Art_Vandelay

    Knowing that I now have a greater faculty for critical thought is poor compensation.

    Perhaps…but what about the baby flesh?

  • Thom Mills

    You are using skepticism and negative words a lot here. Are you depressed? There is help out there. :- (

  • CBrachyrhynchos

    Granted, I don’t know about the research methods used in this particular case, but a problem I had when I was doing research on word frequencies is that when you have a dataset in the thousands, your statistical power is ridiculous to the point of breaking many tests. So you’ll get statistically significant results on phenomena that have trivial effect size. For my doctoral thesis, I reported that girls (it was an educational system) wrote longer messages than boys, but the mean difference was 0.5 characters.

    I think I liked the admission by Pamela Gay that her doctoral thesis was simultaneously a statistical success and a practical failure. She supported her hypothesis but at an effect size that made it worthless. This is a common failure I see in big-data psychology studies, and CMC studies.

  • Edmond

    It would come as no surprise to me to find that a person is happier when they belong to the 80-90% majority, in a nation where many of the laws are tailored to your religion’s advantage, where your religion is endorsed on money, in the pledge of allegiance, or in the national motto. The people in the minority who must kowtow to the religious social structure around them are bound to be less happy.

  • John Small Berries

    I agree. Is “associated” by whom? The researchers? Based on what evidence?

  • Geoff Boulton

    Is it just the case that American theists are happier because they’re part of the ‘in crowd’ while atheists are shunned and made to feel like outsiders? It would be interesting to see how the figures come out in a secular country like Sweden.

  • Anna

    This is all very interesting, but I’m not sure how much it can tell us about people’s emotional states. Frankly, I would question whether anyone’s Internet activity is an accurate reflection of that. I certainly wouldn’t describe myself as unhappy, for example, even though the majority of my comments here on FA are in reaction to stories that strike me as negative. I might be complaining on a regular basis, but I don’t dwell on those negative stories once I shut off the computer.

  • Feminerd

    But but but they did maths! And it looks all sciency!

    I’d call the results interesting enough to call for more rigorous study, because this sort of datamining is rife with all sorts of confounding variable issues. However, it’s certainly not powerful enough to call it more than a fairly unimportant data point in a sea of sociological studies.

  • Sar

    I used to be very religious, in short, the answer is not only no but hell no. Before I realized I was an atheist I felt like I was in a massive hoard of people that were all indulging in a happy thought. It wasn’t that I didn’t see the appeal of God, hell I persuaded myself God must be real because then things were ‘good.’ But, they weren’t. I felt more ostracized and unaccepted and UNHAPPY when I was in church more than I did anywhere else. I felt unfulfilled, I felt like none of the answers I wanted about anything were available to me and if I asked too many questions I was probably a sinner or thought wrong, it made me feel like I was somehow inferior to the other believers. That’s not to say I was happier when I finally left the church. I still felt lost and like the questions I wanted answered were evading me. So, I think that the reason some might believe Christians are happier is because they believe they are having existential questions answered. For me, I found answers, or more to the point, the way to think about the questions, in philosophy. Philosophy became my church, science and reason became my bible. And now, now I am much happier. Happiness, for most people, comes from having a purpose. Many of us find that God or the belief in some higher power isn’t at all the only way to find purpose.

  • CBrachyrhynchos

    To clarify, Pamela Gay is an astronomer. I did a bit of CMC research. The two are not the same.

  • Acleron

    The large data set also means that small systemic errors become significant. Their claim that the large size overcomes these problems is wrong, it exacerbates them.

  • Richard Wade

    Several commenters here have raised these points well, but here is the comment I left over at the Pacific Standard article:

    This is a useless and misleading study. The researchers acknowledge that the social stigma against atheists could be the source of dissatisfaction, but their study simply adds to that stigma. The take away for most lay people is “Atheists are unhappy people.” There are far too many fallacious assumptions in this sloppy research. “Thinking” words and “feeling” words do not translate to unhappy and happy. That’s an Olympic leap to a conclusion.

    “the non-believers tended to use ‘a more analytical thinking style,’ which, the researchers write, is ‘associated with less happiness.’” That’s a ridiculous assumption. Associated by who? By people who assume all analytical thinkers are like Mr. Spock? And he’s sooo unhappy. That’s a fictional TV show, not reality. How could such an association be measured? How do you quantify “happiness”? How do you quantify “analytical thinking”?

    Even in their acknowledgments about the possible biases in their study, the authors still use absurd and meaningless terms like “militant atheist.” What the heck is a militant atheist? In its widespread use, that term seems to mean any atheist who isn’t hiding in a closet denying he’s an atheist. If he simply talks openly and politely about his atheism, he’s labeled “militant.” That extremely low standard would not be applied to label a Christian as “militant.” This study suffers from the same negative stereotypes about atheists that most of society has, and it has simply reinforced that prejudice with more muddled thinking.

  • allein

    I noticed that, too. There’s also “know,” “knows,” “knowing,” and “knew”; and “memories” and “remember,” which seem odd to separate.

  • allein

    Also consider that most people will post the good things more often than the bad. How often do you look at someone’s facebook page and think their life is going so much better than yours, and what’s wrong with you? When really you’re both in the same boat, they’re just not putting it all out there for everyone to see the bad stuff.

  • Monika Jankun-Kelly

    I’m sure whites were happier than blacks in the 1950′s South. Being part of a hated minority tends to make people less chipper, for some reason.

  • M. Elaine

    Thank you. Your comment made me laugh out loud. In the gloominess of my atheism, I had almost forgotten what that sounded like.

  • Ellie

    I’m sorry I don’t understand how this is supposed to determine who’s happier.

    Because we use certain words more than others?

    Seems to me like they are making stuff up.

  • Margaret Whitestone

    Their conclusion is that atheists think more so they’re less happy? What a pantload.

  • Matt D

    Well, when I was a theist, my happiness derived from being selfish. For example, I focused on making my desires and “needs” come to fruition but caring about others wasn’t a priority….in fact they were usually an inconvience!
    Oh, I’d still entertain my family and friends and have fun with others, but I still made those events about me, and when I was alone, it was ALL about me, and to be truthful, that WAS me.
    When I realized this, I stopped being a theist and started taking complete responsiblity for everything I am, past, present, and future. I’m happier now….critical thinking has taught me to maintain better ethics and make better lifestyle choices, and that’s just the tip of the iceberg. Suffice to say, I’m not only happier, but a better person.

  • C.L. Honeycutt

    Those two things are actually correlated. People who are more aware tend to be less happy because they better understand what is wrong in the world.

  • C.L. Honeycutt

    Now see, I know this study is biased to the point of being gibberish, because it associates the word “thought” with Dinesh D’Souza.

  • C.L. Honeycutt

    The laughs echo off of nothing, because we are forever alone in our despair.

    Except for me, I have an icy strawberry-banana smoothie. It understands me.

  • C.L. Honeycutt

    The size of each circle represents its frequency of use per 100,000 words. Whether the word is more frequently used by atheists or Christians, and how much more so in absolute numbers rather than as a percentage, is represented by how far left or right it is. I think.

    I think it’s worth noting that the use of the word “know” in online debates circles (possibly not Twitter because of the shallowness of posts made there) is probably very different. Consider “We know Evolution occurs (because of data)” versus “I know God exists (because I believe)”.

    That linked study could just as easily be interpreted as “Conservatives live in a world based less on reality, and are happier because they are unaware of or ignore the hardships they don’t personally experience.” Of course, you might be making the point that these sorts of studies aren’t terribly useful for that very reason.

  • Kat

    Not at all. But then again, I grew up with a fire and brimstone version of Christianity that focused more on how scary and horrible hell was than any of the more positive aspects. I never even much liked the idea of heaven; I just preferred it to being set on fire forever. When I finally realized I didn’t believe anymore, it was a relief. It wasn’t until later, when I spoke to some people who grew up with less fear-based religious traditions, that I even understood the idea of religion providing comfort. I suppose for some people, realizing there’s no god or hell might be like finding out there’s no Santa Claus; for me, it was more like discovering there are no monsters under the bed. That’s just me, though.

  • Tobias2772

    First, let me say that I am perfectly happy with my life of analytical thought. I gives me great pleasure to be constantly working new stuff out. Secondly, if I had to choose between being certain & happy & wrong or questioning & uncertaining, & growing more correct, I would not hesitate to choose the latter.
    My students and I often discuss the difference between being happy and being fulfilled by life. Again, I choose the latter as being more deeply satisfying and meaningful. I wonder if the study could be redone looking for meaning rather “happy”

  • double-m

    I wonder, has the same kind of study been done for other majority vs. minority cases? I’d like too see if the same phenomenon exists for, say, Black vs. White Americans. Google doesn’t seem to find anything at first glance.

  • Anna

    True, and there’s also extra pressure for members of certain religious groups to be a good “witness” for their faith. They may not always be willing to share their true feelings if it portrays them or their religion in a bad light. I’ve read Mormons are especially susceptible to this.

  • Anna

    Not exactly what you’re looking for, but I thought this was interesting:

  • Wonka

    Meh. I was depressed as a Christian. Now I am depressed as an atheist. The big difference, though, is that it is no longer self-loathing.

  • onamission5

    I don’t get it either. Surely “thinking style” cannot tell us whether someone laughs often, hugs their family, spends time with supportive friends, has a job they find fulfilling, has hobbies that they enjoy, or sees the beauty in little things.

    There is something problematic about the assumption, too, that the way neurotypical extroverts perceive happiness is or should be the norm for all. Some people are perfectly happy sitting around drinking tea and reading, others are happiest when embroiled in problem solving, still others need constant social interaction. None of those three folks would necessarily describe themselves as happy doing the activity of the other, but they would probably describe themselves as happy when doing that which suits them. (ala how do you quantify happiness)

    Not to mention, happy is not a perpetual state of being, nor is it necessarily the most admirable if it comes at the expense of function.

  • Luke

    The analytic thinking – unhappiness link the authors refer to is from cognitive psych work showing that getting people to analyze reasons for things like preferences causes them to become less happy with those things. And even though this particular study is correlational, there is other work showing that making people think more analytically leads to lower religiosity.

  • MadSat

    The sad thing is, this tripe passes for “research” these days. Seriously. Pablum that any high school student could do for a science fair, almost without effort, and it is published in a peer reviewed journal. And the sociologists wonder why they get called “soft” science.

  • WillBell

    Thought seems poorly chosen, but it interests me a bit that we use ‘seems’ more often, that may actually be evidence towards the conclusion they have reached IMHO because ‘seems’ is definitely a word used in the sense of ‘because of this evidence, it seems to me that this conclusion may be correct’.

  • Robin Parker

    I can say that prior psych research has found that analytic thinking (i.e. preferring to think deeply about things) can contribute to depression (Andrews & Thomson, 2009).

    This is consistent with research to research on depressive realism, where people who are more depressed more likely to have accurate views of the world (which could be a form of analytic thinking).

    There’s also reason to believe that the positive illusions people experience to elevate their self-esteem (Taylor & Brown, 1988) may be reduced if greater analytic thinking were invoked.

  • blasphemous_kansan

    Interesting points. Thanks!

  • Laurence Lu

    This image has a slightly misleading message. On the two extreme sides, ‘Thought’ and ‘Know’, the difference seems quite big…but look at the ratios.
    211 to 198 and 44 to 59, respectively.
    For my opinion on the choice of we atheists’ words, I think we’re not as content after finding an ‘answer’ as Christians, and that we ponder and wonder more, and are swayed in our beliefs less by emotions. Whether or not this truly affects our happiness in a negative way, I don’t know. But I’m pretty sure that, all in all, choice of words do define our personalities in some way, beliefs do affect our choice of words.

  • Laurence Lu

    People with higher IQ’s may have less of a pure instinct of contentedness at times, and often the less intelligent don’t have as much anxiety or depressive thoughts. So, in a way, that statement is generally correct.