"Let’s be honest: that last sentence was far too generous. Many of us engaged in practices that, in hindsight, were borderline dishonest. We abused experimenter degrees of freedom, engaged in questionable research practices, p-hacked, massaged our data—you pick the euphemism."
One of the things that comments like this always make me think of is the PhD students and postdocs who *didn't* engage in practices that were "borderline dishonest". They got driven out of psychology because they couldn't complete with those who got publications in the "top" journals. Senior psychologists need to think long and hard about why they have high paid and prestigious jobs (typically at the taxpayers' expense) while those who weren't "borderline dishonest" left psychology as "failures".
Nope. Just some random postdoc who is still having trouble comprehending just how little psychology research from "the good old days" has any scientific value. To be clear: I'm genuinely impressed whenever I see a senior psychologist describing their earlier research practices as "borderline dishonest". We need more of this! And it sure beats alternative approaches to interpreting the remarkable run of lucky strikes from the "good old days" (e.g. https://roybaumeister.com/2023/11/13/ego-depletion-is-the-best-replicated-finding-in-all-of-social-psychology/ ).
I would say that I was in a mild depression for 5 to 7 years because of all of this. It’s not easy admitting you are wrong and it’s definitely not easy when you are at the twilight of your career thinking about legacy. So while I do not agree with the choices that others have made here, I also understand it.
Sorry to hear about the years of mild depression :( To my mind, you're an inspiring (and all too rare) example of a senior psychologist who has opened up about problems with their research from "the good old days" and is now pushing things in the right direction! Keep up the good work!!
I am not an anxious person, but I definitely went through some related anxiety at the dawn of my career finishing grad school as an ecologist. I started learning about the reproducibility crisis via podcasts, and 2P4B with you and Yoel was one of them. I got really concerned about power tied to my sample size. I was taught to do something close to p-hacking but at least I provided a list of models tested and their p-values, but tried not to focus on their magical "significance".
Nonetheless I found it to be a hard time. It felt rough leaving academia and not wanting to stand on publications of my own work, and it didn't feel great because even though I left academia, peer reviewed papers are an important aspect of my career area even outside of academia.
While I found that practices in ecology were better than what I was hearing from social psychologists, I still had worries about some practices. While I loved research, I chose to leave academia and not follow a publish or perish career. Now, regrets from that choice...
It is interesting that the implication is that such a critique can only be from an insider such as Mr Schimmack!!!!
Can you imagine if the general public knew the truth!! (e.g The more often they see it on a TED talk, the less likely it is to be true)
How much is Social Psych relying on fooling MOST of the people most of the time? And how long can it last? I worry about the backlash, because we are not even fully there with Trump and his recent antics IMHO.
But the deeper reason for all this is that liberals believe -- axiomatically and dogmatically, not for scientific reasons -- that in a just society, everyone would be equal, and if people are not equal, that is presumptively an indicator of some sort of injustice. In this sense liberalism is more like a religion than a science. But it's a religion that controls most of what goes by the name of social science. Sociologist Christian Smith called this The Sacred Project of American Sociology, but it really applies to all of social science.
For those who don't subscribe to the religion of equality, i.e. nonliberals (which description encompasses most human beings), the religious nature of much social science is clearly evident. Fortunately there are a few nonreligious social scientists, like Lee Jussim, who have an alternative perspective, despite the best efforts of academic institutions in general to make sure that no one hears that alternative perspective.
A fun question for liberals is: what's one piece of quantitative scientific evidence that shows all people are equal? I mean, there isn't one. People are different in every possible way that science can measure. Human equality is a theological belief.
Nice post. There really is no reason for anyone to take stereotype threat seriously. I posted something to this effect on the SPSP listserve about 8 years ago or so ... and you can imagine what happened next.
Hey, I thought there was a similarly largescale RRR for race and stereotype threat. Forscher among the crew. You know whatever happened to that? I heard it got shut down because of Covid, but that is now a while ago...?
I heard about a big study looking at the stereotype threat effect over time looking specifically at the black-white difference, to be led by Neil Lewis Jr. However, that was many years now and not sure what happened. Maybe COVID casulty? Even this RRR was almost knocked out by COVID...
It is still in the works. From the December 2024 PSA update: "Stereotype Threat Study (PSA-005): We are working on the final analyses and robustness checks. "
I wish someone would do something like a betting market on this, or even just a forecasting survey. It'd be really interesting no matter how things come out:
V1. Race/Stereo threat actually replicates, with effect sizes close to Steele&Aronson and/or meta-analyses. (I doubt this, but its possible):
a. Most folks predicting this are right, congratulations, we are not all clueless!
b. Most folks are too skeptical and underestimate the effects. A warning about overgeneralizing from all the Replication (and everything else, Theory, Generalization, Measurement, Statistics, Mediation) Crisis rhetoric. I'd definitely be in this crowd.
V2. Race/Stere threat either does not occur (ns) or occurs at a small to miniscule effect size (small, say, r=.05=.10, miniscule, r<.05 which can still be "significant" at p<.05 with a large enough sample).
a. Most folks predicted a larger effect, and are shown to be clueless about this (see Schaerer et al 2023 for something very much like this regarding workplace sex discrimination). Amazing testament to the worthlessness of "expertise" in social psych or among whoever is sampled.
b. Most folks knew that ST was mostly bogus and get this right. This would be amazing, too, and a tribue to the good judgment -- if not the good research -- of those in the field.
Nice post, Mickey. As always, I appreciate the honest reflection and reckoning. Methodological issues aside, the empirical studies were rarely actually testing the supposed theory. Steele & Aronson, for example, included no assessment of domain identification, a key aspect of the theory! Anyway, it seems clear that a reason why the idea has persisted, and why many even take offense to the idea that stereotype threat is not "real," is that they equate "stereotype threat" with "stereotypes are threatening." That is, claiming that stereotype threat, a specific theoretical account, does not have empirical support is taken as a denial of the impact of stereotypes.
Agree 100%. And I like your framing here, yes stereotypes can be threatening and unnerving even if it doesn't rise to the level of impacting academic performance in high stakes tests. I mean, I feel a variant of stereotype threat when tipping and there are other people around (knowing the stereotypes about my people's, umm, thriftiness). So, yes, stereotypes can in fact affect the people they target. Thanks for the added nuance, Moin!
I'm sure that this is a rather basic question, but in what ways is it taken into account that decades later white women are likely less impacted in general by the inclusion or exclusion of language about intelligence. There's been a considerable shift in public discourse about differences between white men and white women, between white and black people.
This is an area where replicability seems doomed by the potential for variance in results just because of things like how facilitators behave when interacting with people, even unconsciously.
So to pursue replicability over decades of social change that go well beyond the kinds of changes seen in previous sets of 2-3 decades because of the ubiquitousness of information and knowledge spread thanks to the world wide web?
You know, a lotta ins, a lotta outs, a lotta what-have-yous.
I dig the way you do business, Brandt. This is an excellent question and one the field has grappled with. When something fails to replicate 20-30 years later, is it that the original was incorrect or that the culture has changed so much as to make the original no longer applicable? I think both are possible, but I'm skeptical that this failed replication reflects culture change (e.g., that stereotypes about women and math have changed). Apparently, other psychologists have looked at this formally and found that stereotypes about boys being better than girls in math are (a) not that strong and (b) haven't changed much over the past few decades.
Came here to ask the same question. The stereotypes might not have changed much, but it feels plausible to me that a 20 year old undergraduate in 2024 feels the weight of sexist stereotypes differently from a 20 year old undergraduate in 2004.
That said, I don't think we need to fish too hard for explanations of the non-replication - the QRPs you highlight are more than sufficient. Thanks for writing
Question: in what way did this literature engage with the assessment literature and, specifically, examinations of the predictive power of achievement tests across groups? A quick search revealed papers by College Board, which reported (1) group differences in SAT scores and (2) under/over prediction of college GPA based on SAT scores. SAT scores apparently overpredict GPA for some groups that putatively experienced stereotype threat, and they underpredict GPA for other groups that putatively experienced stereotype threat. I'd think that the stereotype threat framework would suggest uniform underprediction across such groups.
Again, just curious if the stereotype threat literature engaged with the assessment literature and, if so, how it integrated psychometric analyses of differential item functioning across groups, over/under prediction of GPA across groups, etc.
My sense is that serious assessment people, like Paul Scakett at U of Minnesota, never bought any of the stereotype threat evidence; and this was well before the replication crisis. The only other thing I am aware of in this space is work by Greg Walton on the topic: https://pubmed.ncbi.nlm.nih.gov/19656335/
I'd like to say I did too but I only noticed the incongruity. I thought it could be a science nerd thing till I saw the Easter Egg postscript. That is genius to gamify reading it properly ahaha.
Yes people still take the concept seriously, but the goalpost have changed. Now, instead of talking about how stereotypes can hurt performance, we instead might talk about how stereotypes might make people *feel* apprehensive, nervous, and unwelcome. I but this new way of talking, but also note it is a far cry from how it was originally discussed.
Thanks for this, sir. Recently retired from a lifetime in pastoral ministry, I pray for the day that pastors ourselves admit our failures. Afer all, we believe in the Fall and Original Sin, so it should be normal for us to cop to our own dishonesties, let alone mistakes. As you do here.
About the issue of stereotype threat, spending most of my life in communities surrounding institutions of higher education, I'd often warn congregants who were students or profs/administrators that social sciences are influenced by cultural ideologies, whereas Scripture is not—according to the Christian doctrine of special revelation. So when the two are in conflict, we must always submit to the doctrines of Scripture where social science claims to debunk it. Yes, this is an act of faith, but much of the truths taught in higher education are, also.
Cardinal Newman in his "The Idea of the University" was right in pointing out that the removal of theology from the university, what he called "the queen of the sciences," impoverished every other discipline.
Again, thank you for your meekness and humility. I do believe Jesus approves.
It is good to see people taking responsibility for the bad reputation of science in general and psychology in particular. Stereotype Threat has made it's way into workplace and government policy and into our unquestioned assumptions. I feel I've read about it almost as often as a certain prison experiment. I hope you can gain enough momentum for that uphill battle and (yeah I'm gonna do it) hopefully lay Planck to rest at the summit.
Ding Ding Ding! Way to go, Sean-y! You won a 3-month paid subscription. You cool with me using your name in a future edition of Regret Now? I like to name all the Little Urban Achievers.
Agree with Slowlyreading that Jussim has an important perspective (though I might say heroic) in this area.
Not many had the balls to do what he has done. (Assert Stereotype accuracy being highly replicable and point out the QRP's in relation to Stereotype threat)
If you look at the first Steele and Aaronson paper it looks like a joke when viewed in the light of the replication crisis (Splitting subjects into groups of 10 and claiming that the results of 10 people say something meaningful about the world!)
The stupid accusations with Jussim around the mule thing being just one example of the pettyness (relational aggression) and reverse racism he has had to endure.
I am 100% not a "scientist" and find this essay and others I've read recently about the seemingly wide spread lack of reproducibility alarming. "Trust the science" has lost all meaning for me.
It's unfortunate you feel this way because it is scientists (not activists, not politicians, not podcasters) who figured this out and are correcting themselves. To paraphrase Churchill: Science is a terrible way of knowing things--it's slow, error-prone, inaccurate-except it is better than any other way of knowing. Science self-corrects, unlike politicians or religion, and we should applaud those moments of self-correction as evidence that science is working, not failing.
Pure snark, how many more moments of self correction can I expect?
I agree that it is mostly scientists who are exposing charlatans as it is mostly scientists who create the bunkum. My point is that as a member of the general non-science public I am trending from skeptical to cynical.
I hear you, but I urge you to take a few steps from the cliff. Science has built bridges that don't fall, sent astronauts to space, and cured many diseases. Perfect, no; but it builds stuff that works.
Based on the recent papers I've read, there's still quite a lot of charlatanism in science, and many self-corrections due. Outside of economics, it's still routine for social science researchers to make explicitly causal claims when talking to the media about research based on simple regression analysis that omits obvious confounders, even if they have a pro-forma disclaimer about not being able to infer causation in the actual paper. IMO this kind of thing should be strongly frowned upon and heavily sanctioned. It's not okay to misrepresent your research to the public.
"Let’s be honest: that last sentence was far too generous. Many of us engaged in practices that, in hindsight, were borderline dishonest. We abused experimenter degrees of freedom, engaged in questionable research practices, p-hacked, massaged our data—you pick the euphemism."
One of the things that comments like this always make me think of is the PhD students and postdocs who *didn't* engage in practices that were "borderline dishonest". They got driven out of psychology because they couldn't complete with those who got publications in the "top" journals. Senior psychologists need to think long and hard about why they have high paid and prestigious jobs (typically at the taxpayers' expense) while those who weren't "borderline dishonest" left psychology as "failures".
Uli...is that you? But, sigh, yes, you are correct here.
Nope. Just some random postdoc who is still having trouble comprehending just how little psychology research from "the good old days" has any scientific value. To be clear: I'm genuinely impressed whenever I see a senior psychologist describing their earlier research practices as "borderline dishonest". We need more of this! And it sure beats alternative approaches to interpreting the remarkable run of lucky strikes from the "good old days" (e.g. https://roybaumeister.com/2023/11/13/ego-depletion-is-the-best-replicated-finding-in-all-of-social-psychology/ ).
I hear you.
I would say that I was in a mild depression for 5 to 7 years because of all of this. It’s not easy admitting you are wrong and it’s definitely not easy when you are at the twilight of your career thinking about legacy. So while I do not agree with the choices that others have made here, I also understand it.
As an aside, have you seen the paper called the natural selection of bad science? It formalize some of your thoughts above. https://royalsocietypublishing.org/doi/10.1098/rsos.160384
Sorry to hear about the years of mild depression :( To my mind, you're an inspiring (and all too rare) example of a senior psychologist who has opened up about problems with their research from "the good old days" and is now pushing things in the right direction! Keep up the good work!!
I am not an anxious person, but I definitely went through some related anxiety at the dawn of my career finishing grad school as an ecologist. I started learning about the reproducibility crisis via podcasts, and 2P4B with you and Yoel was one of them. I got really concerned about power tied to my sample size. I was taught to do something close to p-hacking but at least I provided a list of models tested and their p-values, but tried not to focus on their magical "significance".
Nonetheless I found it to be a hard time. It felt rough leaving academia and not wanting to stand on publications of my own work, and it didn't feel great because even though I left academia, peer reviewed papers are an important aspect of my career area even outside of academia.
While I found that practices in ecology were better than what I was hearing from social psychologists, I still had worries about some practices. While I loved research, I chose to leave academia and not follow a publish or perish career. Now, regrets from that choice...
It is interesting that the implication is that such a critique can only be from an insider such as Mr Schimmack!!!!
Can you imagine if the general public knew the truth!! (e.g The more often they see it on a TED talk, the less likely it is to be true)
How much is Social Psych relying on fooling MOST of the people most of the time? And how long can it last? I worry about the backlash, because we are not even fully there with Trump and his recent antics IMHO.
I went through something similar, but in a different field. I must say, social psych sounds much worse though.
Very important and thank you.
But the deeper reason for all this is that liberals believe -- axiomatically and dogmatically, not for scientific reasons -- that in a just society, everyone would be equal, and if people are not equal, that is presumptively an indicator of some sort of injustice. In this sense liberalism is more like a religion than a science. But it's a religion that controls most of what goes by the name of social science. Sociologist Christian Smith called this The Sacred Project of American Sociology, but it really applies to all of social science.
For those who don't subscribe to the religion of equality, i.e. nonliberals (which description encompasses most human beings), the religious nature of much social science is clearly evident. Fortunately there are a few nonreligious social scientists, like Lee Jussim, who have an alternative perspective, despite the best efforts of academic institutions in general to make sure that no one hears that alternative perspective.
A fun question for liberals is: what's one piece of quantitative scientific evidence that shows all people are equal? I mean, there isn't one. People are different in every possible way that science can measure. Human equality is a theological belief.
Nice post. There really is no reason for anyone to take stereotype threat seriously. I posted something to this effect on the SPSP listserve about 8 years ago or so ... and you can imagine what happened next.
Hey, I thought there was a similarly largescale RRR for race and stereotype threat. Forscher among the crew. You know whatever happened to that? I heard it got shut down because of Covid, but that is now a while ago...?
I heard about a big study looking at the stereotype threat effect over time looking specifically at the black-white difference, to be led by Neil Lewis Jr. However, that was many years now and not sure what happened. Maybe COVID casulty? Even this RRR was almost knocked out by COVID...
What you are referring to is https://psysciacc.org/projects/psa005.html
It is still in the works. From the December 2024 PSA update: "Stereotype Threat Study (PSA-005): We are working on the final analyses and robustness checks. "
I wish someone would do something like a betting market on this, or even just a forecasting survey. It'd be really interesting no matter how things come out:
V1. Race/Stereo threat actually replicates, with effect sizes close to Steele&Aronson and/or meta-analyses. (I doubt this, but its possible):
a. Most folks predicting this are right, congratulations, we are not all clueless!
b. Most folks are too skeptical and underestimate the effects. A warning about overgeneralizing from all the Replication (and everything else, Theory, Generalization, Measurement, Statistics, Mediation) Crisis rhetoric. I'd definitely be in this crowd.
V2. Race/Stere threat either does not occur (ns) or occurs at a small to miniscule effect size (small, say, r=.05=.10, miniscule, r<.05 which can still be "significant" at p<.05 with a large enough sample).
a. Most folks predicted a larger effect, and are shown to be clueless about this (see Schaerer et al 2023 for something very much like this regarding workplace sex discrimination). Amazing testament to the worthlessness of "expertise" in social psych or among whoever is sampled.
b. Most folks knew that ST was mostly bogus and get this right. This would be amazing, too, and a tribue to the good judgment -- if not the good research -- of those in the field.
Nice post, Mickey. As always, I appreciate the honest reflection and reckoning. Methodological issues aside, the empirical studies were rarely actually testing the supposed theory. Steele & Aronson, for example, included no assessment of domain identification, a key aspect of the theory! Anyway, it seems clear that a reason why the idea has persisted, and why many even take offense to the idea that stereotype threat is not "real," is that they equate "stereotype threat" with "stereotypes are threatening." That is, claiming that stereotype threat, a specific theoretical account, does not have empirical support is taken as a denial of the impact of stereotypes.
Agree 100%. And I like your framing here, yes stereotypes can be threatening and unnerving even if it doesn't rise to the level of impacting academic performance in high stakes tests. I mean, I feel a variant of stereotype threat when tipping and there are other people around (knowing the stereotypes about my people's, umm, thriftiness). So, yes, stereotypes can in fact affect the people they target. Thanks for the added nuance, Moin!
I'm sure that this is a rather basic question, but in what ways is it taken into account that decades later white women are likely less impacted in general by the inclusion or exclusion of language about intelligence. There's been a considerable shift in public discourse about differences between white men and white women, between white and black people.
This is an area where replicability seems doomed by the potential for variance in results just because of things like how facilitators behave when interacting with people, even unconsciously.
So to pursue replicability over decades of social change that go well beyond the kinds of changes seen in previous sets of 2-3 decades because of the ubiquitousness of information and knowledge spread thanks to the world wide web?
You know, a lotta ins, a lotta outs, a lotta what-have-yous.
I dig the way you do business, Brandt. This is an excellent question and one the field has grappled with. When something fails to replicate 20-30 years later, is it that the original was incorrect or that the culture has changed so much as to make the original no longer applicable? I think both are possible, but I'm skeptical that this failed replication reflects culture change (e.g., that stereotypes about women and math have changed). Apparently, other psychologists have looked at this formally and found that stereotypes about boys being better than girls in math are (a) not that strong and (b) haven't changed much over the past few decades.
Came here to ask the same question. The stereotypes might not have changed much, but it feels plausible to me that a 20 year old undergraduate in 2024 feels the weight of sexist stereotypes differently from a 20 year old undergraduate in 2004.
That said, I don't think we need to fish too hard for explanations of the non-replication - the QRPs you highlight are more than sufficient. Thanks for writing
Thanks for this, Mickey. Very nice.
Question: in what way did this literature engage with the assessment literature and, specifically, examinations of the predictive power of achievement tests across groups? A quick search revealed papers by College Board, which reported (1) group differences in SAT scores and (2) under/over prediction of college GPA based on SAT scores. SAT scores apparently overpredict GPA for some groups that putatively experienced stereotype threat, and they underpredict GPA for other groups that putatively experienced stereotype threat. I'd think that the stereotype threat framework would suggest uniform underprediction across such groups.
https://eric.ed.gov/?id=ED562614
Again, just curious if the stereotype threat literature engaged with the assessment literature and, if so, how it integrated psychometric analyses of differential item functioning across groups, over/under prediction of GPA across groups, etc.
My sense is that serious assessment people, like Paul Scakett at U of Minnesota, never bought any of the stereotype threat evidence; and this was well before the replication crisis. The only other thing I am aware of in this space is work by Greg Walton on the topic: https://pubmed.ncbi.nlm.nih.gov/19656335/
I recognize the "strands in old Duder's head" line!
I'd like to say I did too but I only noticed the incongruity. I thought it could be a science nerd thing till I saw the Easter Egg postscript. That is genius to gamify reading it properly ahaha.
Ding ding ding! Clearly, you’re no human paraquat. Congrats on the 3-month paid subscription! Can I mention you by name in my next Regret Now feature?
Nice summary article Doc.
Are there any researchers who still currently seriously defend the concept, if so, what's a layman's summary of their arguments?
I also wonder how this debate has evolved within feminist academia circles.
Stereotype threat was always one of the 1st things cited in the discussion of gender/sex differences in STEM outcomes.
Yes people still take the concept seriously, but the goalpost have changed. Now, instead of talking about how stereotypes can hurt performance, we instead might talk about how stereotypes might make people *feel* apprehensive, nervous, and unwelcome. I but this new way of talking, but also note it is a far cry from how it was originally discussed.
Thanks for this, sir. Recently retired from a lifetime in pastoral ministry, I pray for the day that pastors ourselves admit our failures. Afer all, we believe in the Fall and Original Sin, so it should be normal for us to cop to our own dishonesties, let alone mistakes. As you do here.
About the issue of stereotype threat, spending most of my life in communities surrounding institutions of higher education, I'd often warn congregants who were students or profs/administrators that social sciences are influenced by cultural ideologies, whereas Scripture is not—according to the Christian doctrine of special revelation. So when the two are in conflict, we must always submit to the doctrines of Scripture where social science claims to debunk it. Yes, this is an act of faith, but much of the truths taught in higher education are, also.
Cardinal Newman in his "The Idea of the University" was right in pointing out that the removal of theology from the university, what he called "the queen of the sciences," impoverished every other discipline.
Again, thank you for your meekness and humility. I do believe Jesus approves.
It is good to see people taking responsibility for the bad reputation of science in general and psychology in particular. Stereotype Threat has made it's way into workplace and government policy and into our unquestioned assumptions. I feel I've read about it almost as often as a certain prison experiment. I hope you can gain enough momentum for that uphill battle and (yeah I'm gonna do it) hopefully lay Planck to rest at the summit.
roll on Shabbos :)
I mean, yes. But it was also right in the spot where I was talking up the contest. To which I say, Over the Line!
Yeah, well, that’s just, like, your opinion, man
Ringer for a ringer
Ding Ding Ding! Way to go, Sean-y! You won a 3-month paid subscription. You cool with me using your name in a future edition of Regret Now? I like to name all the Little Urban Achievers.
Sure! Great post and looking forward to reading more.
"It’s all very complicated. Lots of strands in old Duder's head."
I thought you won, but Trembling Mad beat you by one-minute. Try again...there is one more reference in there.
Agree with Slowlyreading that Jussim has an important perspective (though I might say heroic) in this area.
Not many had the balls to do what he has done. (Assert Stereotype accuracy being highly replicable and point out the QRP's in relation to Stereotype threat)
If you look at the first Steele and Aaronson paper it looks like a joke when viewed in the light of the replication crisis (Splitting subjects into groups of 10 and claiming that the results of 10 people say something meaningful about the world!)
The stupid accusations with Jussim around the mule thing being just one example of the pettyness (relational aggression) and reverse racism he has had to endure.
I am 100% not a "scientist" and find this essay and others I've read recently about the seemingly wide spread lack of reproducibility alarming. "Trust the science" has lost all meaning for me.
It's unfortunate you feel this way because it is scientists (not activists, not politicians, not podcasters) who figured this out and are correcting themselves. To paraphrase Churchill: Science is a terrible way of knowing things--it's slow, error-prone, inaccurate-except it is better than any other way of knowing. Science self-corrects, unlike politicians or religion, and we should applaud those moments of self-correction as evidence that science is working, not failing.
Pure snark, how many more moments of self correction can I expect?
I agree that it is mostly scientists who are exposing charlatans as it is mostly scientists who create the bunkum. My point is that as a member of the general non-science public I am trending from skeptical to cynical.
I hear you, but I urge you to take a few steps from the cliff. Science has built bridges that don't fall, sent astronauts to space, and cured many diseases. Perfect, no; but it builds stuff that works.
Based on the recent papers I've read, there's still quite a lot of charlatanism in science, and many self-corrections due. Outside of economics, it's still routine for social science researchers to make explicitly causal claims when talking to the media about research based on simple regression analysis that omits obvious confounders, even if they have a pro-forma disclaimer about not being able to infer causation in the actual paper. IMO this kind of thing should be strongly frowned upon and heavily sanctioned. It's not okay to misrepresent your research to the public.
https://federicosotodelalba.substack.com/p/studying-behavior-is-not-going-to?r=4up0lp