Chart-topping original movies have gone extinct. People have a lot of explanations for this, but they're all incomplete because they don't realize the same thing is happening everywhere. An oligopoly has conquered all of popular culture.
My paper with
@DanTGilbert
is out today in Nature: people believe that people are less kind than they used to be, they're probably wrong about that, and we have an idea where this illusion comes from.
Recently I wrote a post suggesting that peer review doesn't work, and then some weird things happened. A tenured professor threatened to get me fired. Strangers sent me unhinged emails. (People said nice things too.) This week I sort through it all.
There are two kinds of problems: strong-link problems and weak-link problems.
Weak-link: quality depends on how good the *worst* things are
Strong-link: quality depends on how good the *best* things are
Most importantly, I don't think we've realized what all these sequels and spinoffs are doing to us. It's not that they're bad––some are great! But movies, TV, music, books, and video games should expand our consciousness, and they can't do that by feeding us reruns forever
Published today in PNAS: people don't know how American public opinion has changed. I've been working on this for a very long time and I'm excited to share it. Here's a thread so you don't have to read the paper.
In television, for example, it used to be pretty rare for two versions of the same show to appear twice at the top of the viewership charts. Now it's common.
As a result of a generation of weak-link thinking, we've got many entries into my least-favorite genre of all time: "Nobel Prize winner explains how they could not do their Nobel Prize-winning work today."
It's a shame that about 30% of Harvard students go into finance and consulting, 'cause they should really go work for the IRS. Wanna see the most thorough audit of your life? Try giving a Harvard undergrad a 3.5/5 on a homework assignment.
To recap:
1) People believe that people are less good than they used to be
2) They're probably wrong about that
3) Biased exposure and biased memory could create this illusion
The other place we predicted less/no decline is in times before people were born, when there are no memories to fade. And indeed, people seem to believe that moral decline began when they arrived on Earth.
People think peer review has been with us for centuries. Not really! Scientific publishing was always a hodgepodge and most outlets were nothing like our current system. Pre-publication peer review only became common in the 1960s.
I know people get yelled at on the internet all the time, I just didn't expect to get yelled at for *this*. Why did questioning peer review strike such a nerve? Here's one hypothesis:
First, people have biased exposure to negative information about the morality of people in general ("if it bleeds, it leads"). Second, they have biased memory––the negativity of negative information fades faster than the positivity of positive information (the Fading Affect Bias)
I wrote an explainer on my blog, but Twitter throttles those links, so you'll have to find it in my bio.
Too bad! Seems like the morality of social media magnates has declined.
Plus, fraud gets published all the time, and it never seems to be reviewers that catch it. Instead, it's good Samaritans and people with inside knowledge. For instance:
One common argument: "peer review is a barrier against misinformation; without it, wackos like creationists could publish their 'findings' unhindered!" Well, here's The Journal of Creation, a peer-reviewed journal all about creationism.
Some anonymous bloggers got a bunch of people to eat nothing but potatoes for a month and they lost an average of ~11 pounds and generally had a good time. The world is bright and full of hope.
What should we do instead? What we do best: experiment! Here's mine, a paper written in plain language where anyone can access it. All data, code, materials, and preregs posted publicly.
Does peer review actually work––do reviewers catch major errors? Not really. In studies, reviewers spot ~25% of major flaws, like "if you look at the graphs, there's no effect"
Scientists never took peer review seriously. For example, if reviews are so important, why don't journals publish them alongside the paper? Maybe because nobody really cares to see what the reviewers said.
This illusion is a powerful tool for aspiring charlatans and despots. It's always helpful to allege decline––"Just put me in charge and I'll make things good again!"
(In 2015, 76% of Americans said "addressing moral breakdown" should be a government priority)
My whole life, I've heard people complain about the demise of human goodness. "Used to be you didn't have to lock your doors at night!" etc. Is this just a vocal minority, or do most people believe this?
Turns out, it's most people. 177 surveys, N = 220,772. Here's a sample:
(By the way, we called this "the illusion of moral decline" because it rolls off the tongue better than "the illusion of the decline of kindness, niceness, honesty, and goodness". Obviously people can use "morality" to mean lots of things. This is what *we* are using it to mean.)
Put those together and you can produce an illusion where, even though every day looks equally bad, you mistakenly think yesterday looked better. This fits with some of our more surprising findings, like the lack of large age effects.
And yet most people treat science like it's a *weak-link* problem. Universal pre-publication peer review, grant committees, scattershot replication attempts––these are great solutions for a weak-link problem, but terrible solutions for a strong-link problem
This isn't just the US. In 2002 and 2006, Pew sampled folks in every country highlighted in red below and asked them whether moral decline was a problem in their country. In *every single country* a majority of folks said it was at least a "moderately big" problem:
Morality probably isn't the only place where we can find this illusion. Lots of people claim that lots of things are declining. Some of them are probably right! But our results suggest that it's easy to see decline that isn't there.
Meanwhile, a meta-analysis that came out last year found that cooperation rates in economic games *increased* from 1956 to 2017 (contrary to the authors' predictions). When we incentivized participants to estimate that change, they got it backwards:
Of course, all of these survey questions are a little weird. Do people actually think that people have become less kind, honest, nice and good?
Yes. They think this decline has been happening their whole lives and that it's still going on today. From our original studies:
People believe there's been a big decrease in prosociality. That hasn't happened. So why do they believe in it?
Anything worth studying in psychology is probably multiply determined, but we have an explanation that uniquely fits our data
Could we fix it? I'm skeptical. It already can take years to publish a paper, and making it even stricter makes all the other problems with peer review worse.
It also suggests the illusion might decrease or disappear if you turn down one of those effects. And it seems like that happens. People probably don't have biased negative information about people they know, and indeed, they think those people have *improved* (or declined less)
Okay, so people really seem to think that people have become less kind, honest, ethical, etc. Are they right?
One obvious strike against their theory is that violence seems to have declined (ie
@sapinker
's work). But––
We found 140 surveys (N = 12 million) where people were asked multiple times about the current state of morality around them. We found strong evidence of no change over time. For instance:
When people say morality is declining, do they mean that *individuals* have gotten worse (personal change), or that worse folks have replaced better folks (interpersonal replacement)?
This doesn't seem to be a pandemic effect. We found the same results in January 2020 as we did in May 2020, as well as in 2021 and 2022. Plus all the archival data predates the pandemic.
It can be really hard to recognize a strong-link problem, because it requires you to see bad stuff and go "whatever!" It's so tempting to try to stamp out everything bad, not realizing that you're also preventing good stuff from coming into existence.
Surprisingly, we only find small and inconsistent age effects. Older and younger people agree on the amount of decline per year. Older people have just been around longer to see more of it.
In short: people don't know how public opinion has changed. One reason: they think the past was way more conservative than it was. One consequence: misperceiving these changes may lead people to support policies they wouldn't otherwise prefer.
When people complain about moral decline, they mainly mean things like "people don't respect each other anymore." Has *that* changed?
There's no obvious and objective way to answer that question, but there's lots of useful survey data
(I know it's weird for me, Mr. Rise and Fall of Peer Review, to be publishing in a journal. We submitted this last July, and the process of getting it published was one inspiration for that piece.)
The fundamental problem with most attempts to reform science is that they want to turn the Big Ship. Instead, we should be launching lots of Little Ships.
If you made it this far, lemme just say that journals force you to make stuff boring, so I put the interesting stuff on my blog. (Post about this paper coming soon.)
There’s a kind of person who will try to convince you the world is corrupt and the only way to succeed is to be cynical and cunning. You gotta remember these people are trapped inside their own self-fulfilling prophecy
We live in a golden age of people writing words on the internet. Here's some top-tier stuff I've read recently.
1. "The Scientific Virtues" by
@mold_time
. Required reading for anyone who wants to do science.
I’m running a prototype Science House this summer! It’s a great opportunity for anyone looking to do research outside academia. No application, just do some science, post it on the internet, and send me the link.
Finally figured out my teaching job is actually two jobs.
Job 1: teach
Job 2: determine if students' grandmothers have actually died
Job 1 rules, job 2 blows, and these jobs should not go together.
Intelligence tests are supposed to measure problem-solving ability. People who are better at solving problems should live happier lives.
So why aren't people with higher IQs happier?
The best conversationalists talk loudly, intensely, and quickly, smile a lot, and nod or shake their heads when listening. Suggests we could effectively replace humans with grinning, screaming bobbleheads
Advances in algorithmic feature detection and machine learning allowed us to answer the perennial question, “What makes someone a good conversationalist?”
Some things about peer review that make no sense to me:
1) Why don't reviews get published along with the paper? I want to know:
-did anyone run the code and check the stats?
-does my opinion converge with the reviewers’?
-how did the paper evolve in response to feedback?
Instead, it seems like people have a stereotype about the past. In a second study, people drastically underestimated liberal attitudes in the past, but only slightly underestimated liberal attitudes today. (They were only a bit more accurate for today, though!)
On another 10 attitudes, people got the direction of change totally wrong. They thought that support for gun control is increasing, that support for the death penalty is lower today than in 1972, and that opposition to immigration is rising. Wrong on all accounts.
As a psychologist, I'm especially interested in the mood effects: people go from 4.3 to 4.7 (out of 7) over the course of the trial. Not huge, but people are eating mostly potatoes and getting *happier*. Psychologists present effects of this size all the time as "successful"!
I grew up thinking that you can NEVER mix whites and colors in the laundry and one day I did and nothing bad happened and I think all of life might be this way
Polls aren't perfect and people's guesses were noisy, but these biases were big. For instance, people thought that only 32% of Americans said they would vote for a woman for president in 1978, which they thought increased to 70% in 2010. Actual numbers: 74% and 96%.
We then got a nationally representative sample and had them read each question and estimate how Americans had answered it at the earliest and latest time points for which we had data. On 29/51 items, people overestimated attitude change. Graphs:
On another 10 items, people underestimated change. The biggest biases were about race and marriage. I was flabbergasted that ~50% of Americans in 1990 opposed family members marrying someone who is Black, Hispanic, or Asian. Participants were not surprised.
A few years ago I got into reading lots of public opinion polls, and I was constantly surprised. People became *more* in favor of immigration after Trump's election? Support for gun control has *fallen*? White Americans feel as warm toward Black Americans as they did in *1964*?