Return Home


Thursday, November 23, 2017 - 06:00 PM


Back in 1995, Claude Steele published a study that showed that negative stereotypes could have a detrimental effect on students' academic performance. But the big surprise was that he could make that effect disappear with just a few simple changes in language. We were completely enamoured with this research when we first heard about it, but in the current roil of replications and self-examination in the field of social psychology, we have to wonder whether we can still cling to the hopes of our earlier selves, or if we might have to grow up just a little bit.

This piece was produced by Simon Adler and Amanda Aronczyk and reported by Dan Engber and Amanda Aronczyk.

 Support Radiolab today at


Ryan Brown, Eric Day, Michael Inzlicht, Steven Spencer and Claude Steele

Produced by:

Simon Adler and Amanda Aronczyk


Daniel Engber


More in:

Comments [40]

Mary Strong from UK

I agree with Aaron. As a schoolchild in 70s UK, some of our teachers, especially in the maths, physics and chemistry departments, treated girls as second-class in every way. My children's experience of school in the last ten years seemed much better in this respect, so I'm not surprised at the change in results.

Mar. 04 2018 08:50 AM
Joanne Holley from Berlin

Really interesting episode, as usual! but by the end I was just desperate to hear the phrase Publication bias. This is well known in other scientific fields, why not look at how it's been handled elsewhere (though by no means resolved). It doesnt necessarily mean you throw out all old research, but re-evaluation ain't a bad thing.....

Feb. 16 2018 03:55 AM
N.L. from Madison, WI

The work of Claude Steele is extremely valuable, particularly in education research and particularly for minoritized individuals in the US. It provides both a language and a scientific explanation for the achievement gap in classrooms; more specifically in pre-dominantly white classrooms where the dominant culture wrongfully imposes itself onto individual learners in the education process, causing significant stress for students from non-dominant backgrounds (e.g. non-white, non-middle class). Although it is important to note the problem of replicability in social science research, this is nothing new to social science researchers. Further discussion of replication in social science research could have been challenged in innumerable other contexts (i.e. such as in critiquing and interrogating the 'bell curve' study). It frustrates me that this was the study chosen and it confuses me why this was the case. Are there further underlying implications behind this episode altogether?

Furthermore, when conducting research on human subjects, rather than objects in the material world (e.g. in the 'hard' sciences), there will likely be great variation in outcomes across time and space. Unlike objects in the material world, human subjects are are active agents in the real world. In social science research, human subjects carry with them a complex socio cultural history that plays into the way they interact with their surrounding environment. Conducting a study in the US and replicating a similar study in Toronto is like comparing apples to oranges. If a researcher is to replicate a study, then more attention should be given to matching the demographics of the population in the original sample; so that you are indeed comparing apples and apples. The error here is not so much in the part of Steele and colleagues, but rather in the poor attempts to replicate stereotype threat studies.

I am still disappointed in the way Claude Steele was thrown under the bus...
Thank you for your contribution to knowledge production, Steele & colleagues.

Jan. 17 2018 11:39 AM
Giulia from State College, PA

Was I chasing signal? Was I chasing noise?????????????????

Jan. 13 2018 02:29 PM
Erin from Oakland

I wish we could have heard more of the rationale behind Claude Steele's statement: "I don't think there is anything that could change my mind." On the face of it, it seems blatantly anti-science. However, if Radiolab would have dug deeper into the comments that followed this statement, we might have seen a scientist coalescing and summarizing a career of information. Not to allow a scientist to convey that information, taking only the shocking soundbites, is equally, if not more, anti-science. Sorry, that was antagonistic - I really do like Radiolab (with the vehement exception of the Yellow Rain episode).

Let's look at each of the snippets of information Steele provided: "Stereotype threat has been dramatically well replicated". How well? Did Radiolab cherry pick the contrary results from Inzlicht? Did Inzlicht publish these results? Are there not multiple meta-analyses and syntheses of Stereotype threat results? What do they say? "The kinds of stereotypes that are threatening might change?" This seems highly plausible, is there any documentation behind this? "Interventions designed to reduce stereotype threat can have dramatic, long-lasting effects on achievement." Instead of focusing on the one study where results in 2006 could not be reproduced in 2011, can we look at the "dozens of studies" where interventions were successful? Is Jo Boaler's work not directly derivative? Has she not found numerous interventions that facilitate children of color learning math? If assuming that Stereothreat is real encourages and facilitates diversity in STEM, should we not have ample evidence to the contrary before we cast aside a seemingly constructive principle?

Finally, two things weren't addressed: (1) What are the p-values of the bulk of the stereotype threat results? I ask this, because it is common statistical practice to adjust your p-value for the number of tests you performed. Would this not counteract the initial, overarching problem identified by the "False Positive Psychology" paper (Simmons 2011, no? Does the podcast state who wrote the paper? It should, no?)? If not, can you explain why? Giving Professor Steele the benefit of the doubt, maybe he has these p-values in mind and performed a mental estimate for "number of tests", "amount of replication", and "average p-value". (2) Haven't there been twin studies that look at how encouragement versus insult affect performance/achievement? In other words, if there is a real-life, first-principles phenomena about how people internalize feedback, does it not bolster an argument that the phenomena also applies it to a complex issue?

Well, in any case, I am not a social scientist. So I am not sure why I am weighing in. Maybe because I believe that "constructive criticism is caring". And, importantly, that we can't overlook the real barrier that implicit biases, internalized and otherwise, cause for our children.

Jan. 09 2018 02:18 PM
Peter white from United Kingdom

Claude Steel: "I don't think there is anything that could change my mind."

Man, this is why a lot of scientists don't view social scientists as true scientists. You can just hear in his voice that he is more or less offended by the idea that the effect he was researching might not be real or as significant as his studies showed. He is not emotionally detached from the outcome which only adds to the doubt I have about the legitimacy of this effect.

What proportion of the people who choose to research the stereotype effect do you think have an emotional attachment to the outcome? It's not hard to imagine that it's a majority and that it would impact their methodology to ensure the 'right' results.

Dec. 26 2017 07:07 PM

You guys should do a podcast on strictly on The Bell would then find that it has solid science behind it and 99.9% of the outrage is totally misplaced. In this episode you continue to propagate the lie about this book.

For anyone interested in this listen to the Sam Harris podcast with Charles Murray.

Dec. 07 2017 07:35 PM
Reuben Stein from Tri-state area

I was curious to hear if a reason for the inability has been researched instead of implying the original research was flawed in some way. For example when the original scientific test was run our minds view the stereotype in question like we would view Schrödinger's cat and once the cat is out of the bag the original hold the stereotype had on us looses it power to control us. I could even go as far to say that near the end of the episode you quote a test done with hundreds of students and they are get a small percent result and then when it is redone again the results are zero.

Dec. 06 2017 06:37 PM
Steve Spencer from Palo Alto, CA and Columbus, OH

But we stress how important it is to use the whole body of evidence in understanding, and communicating what we understand about a phenomenon like ST.

Yes, we, as ST researchers, were given air time. But using these misguided ideas about replication, Steele and Spencer were positioned as the “accused” with questions that presumed ST research was seriously challenged by replicability issues.

We want to be clear about what ST is and is not: ST is not a person characteristic. It is a situational predicament or pressure: being in a situation or doing something in which you are strongly invested, but where you also know you could be negatively stereotyped based on one of your social identities—your age, religion, gender, race, immigrant status, social class, etc. This pressure doesn’t exist when elements of the predicament are not present—for example, in a situation where the likelihood of being stereotyped is low, or in a situation that one doesn’t care much about, or in a situation where the performance involved is easy enough to make being negatively stereotyped unlikely. ST is not a law of nature that is the same across all contexts. It is a pressure that arises from the context. Thus we expect it to vary in strength across groups and situations. For example, we expect it to be a pressure on the academic performance of only those groups whose intellectual abilities are negatively stereotyped in the larger society (e.g., women in math) and then only among those in the group who care strongly about doing well in that area, and who are doing work difficult enough to occasion the relevance of the negative stereotype about their ability—the academic vanguard of the group. No serious scientist ever claimed that reducing stereotype threat was a magic cure for, for example, all of the racial and gender gaps in school performance as Radiolab seemed to suggest toward the end of their piece. These gaps, to some large extent, are rooted in economic and structural realities—as in group differences in access to quality schooling. But as these structural barriers are addressed, the growing body of research on ST shows that lifting psychological barriers can be a big help.
To downplay this depth, breadth, and quality of evidence in support of the ST phenomenon by citing a single study and the reportage of a single disillusioned investigator is, to say the least, not strong journalism. It is certainly not science. There simply is no exceptional systematic replication issue in this vast literature. This is our bottom line.
Had our Radiolab friends not been pre-committed to a story line, this should have been their bottom line too.
Claude Steele, Steve Spencer, and Geoff Cohen

Dec. 06 2017 09:36 AM
Steve Spencer from Columbus, OH and Palo Alto, CA

As long-time fans of Radiolab we were disappointed in their recent coverage of research by us and others on stereotype threat (ST), in a show entitled “Stereothreat.” Without evidence, without early consultation, without openness about the show’s intended message during our interviews, the show’s narrative implied that there was significant concern about the replicability of research in this area. We want to be clear: any special concern about the replicability of stereotype threat research is not warranted by any developments in the literature which now includes over 350 studies involving thousands of research participants on essentially all continents of the world. Perhaps they wanted to shore up an image of scientific balance—they’d earlier done a flattering show on the concept--but the recent episode seemed designed to weave together innuendo to call into question a well-established body of research.

To establish this narrative, they made two striking selections. First, from hundreds of investigators doing this research, they found one who, having easily replicated stereotype threat effects while living in New York City, found it difficult to replicate ST effects after moving to the University of Toronto. He didn’t explain this inconsistency. He offered no evidence that ST was generally unreliable. He merely said he worried that it was. Worries are not evidence. This is one “worried” researcher against a vast and established body of empirical, refereed research. Yet he is positioned in the show as the “truth-to-power” figure against an image-protecting orthodoxy. The second selection is of a self-described “contrarian” journalist to make sense of ST research. He describes other research he has questioned, and then one ST study that “failed to replicate” two earlier studies showing that even modest efforts to reduce ST can improve minority student success in K-12 schooling—while downplaying over 30 studies replicating similar positive intervention effects in different academic settings over thousands of students including women in STEM fields. This is clearly cherry picking to make a point—stressing one study against dozens of others with different findings.

A point about science is also relevant here. If you test a given causal relationship enough times—say the effect of reducing ST on academic performance--some replication failures are inevitable. Behavioral science is probabilistic. This is true for all behavioral theories. Thus, if you use on one, or a small number of replication failures to disqualify a theory, while downplaying a vast majority of replication successes, you set a standard that no theory can achieve or sustain—especially as research on the theory proliferates. Replication failures or non-effects can be informative. They can reveal the boundaries of a behavioral effect—for example, the conditions under which ST occurs and those under which it does not occur. Pay close attention is our byword. (to be continued).

Dec. 06 2017 09:34 AM
Michael from Australia

I've got to agree with J. Dollarhyde on this one.

When I heard Claude Steele saying that there wasn't anything that could make him think it wasn't true, I was stunned. Anyone who says something like that cannot be considered a serious researcher in anything.

Dec. 06 2017 02:08 AM

I found myself very interested by this episode on a personal level. I was diagnosed with dyslexia and ADD in 2nd grade and put into a program that allowed me extra time on tests and extra help from my teachers. I was also put on Ritalin and later Adderall. Neither of my learning disabilities are severe.

Being the type of strong headed person I am I tried my hardest to not use any of the extra resources and do as well in school as my peers. I stayed on medication until I graduated high school but by the end of middle school I was doing so well that I was taken off the program that offers extra help. I decided I would no longer take medication and went off to college at University of Delaware entering with a 3.7 gpa and plenty of AP college credits under my belt.

I also think it's important to note that I grew up as a middle class white female. So I was told as kid that the world was my oyster. So when I think of stereotype threat in relation to my life I think of the moments I doubted myself on tests in college because of my learning disabilities. I had clearly proven in high school that I was more than capable of doing well in the educational system but there were always moments in college during tests where I doubted my ability to do as well as my peers. Those moments of doubt do cause extra stress, especially on multiple choice tests where you have to pick the "most right" answer. (As if those types of questions don't beg people to second guess themselves as is)

Regardless of my doubts I graduated on time after maintaining a place on the Deans list for most of my college career with a 3.4 gpa. So my real curiosity here is: If I hadn't been stressed during tests that my learning disabilities would hurt my performance (because I was told all my life that they would) would I have graduated college with an even better GPA and done markedly better on tests? Does stereotype threat cover a wider area than just being a minority?

As a side note I completely agree that you can not expect the same results on studies of stereotype threat across all generations. Especially considering how intense an impact social media plays on current young generations and the widespread sharing of information. Another interesting question to pose that relates to the newest episode of Radiolab/ More Perfect is: Does stereotype threat apply outside of the classroom and affect the relationship between minorities and police? How do the prevailing stereotypes affect the way an african american man acts when a police officer is pointing a gun at them? (Especially if said african american man was told as a child that they would always be a bigger target for police than their white peers)

Dec. 05 2017 05:02 PM
Yue from New Jersey

As a biostatistician working in the field of drug development, I was surprised to hear that an entire scientific discipline had to rediscover familywise type I error rate and multiplicity control. In the field of clinical biostatistics, the so called "type I error inflation" is an anathema. Clinical endpoints of interest would have to be pre-specified in a trial protocol, and the primary statistical analysis methodology needs to be clearly defined. In addition, public registries of clinical trials, such as the are there to prevent publication bias. For certain therapeutic areas, the invisible hand of economics is also working against the fishing of significant results because with fierce competition and limited financial resource, a company cannot keep repeating the same expensive Phase III trials again and again to catch the rare false positive outcomes. Such topics are routinely taught in graduate level statistics courses, especially in those institutions with a biostatistics focus.

Dec. 05 2017 10:23 AM
Jon from Houston, TX

I had serious issues with the way this episode was summed up. First of all, the analysis of replication attempts for stereotype threat seems far conclusive. They seem to end up with the idea that either you can hold on to the illusion of stereotype threat or believe that the achievement gap is due to something else. This is implicitly condoning, or at least opening the door to the Bell Curve point of view they state at the beginning. This is not to say that its good to have taboo theories for the achievement gap in social psychology, but that they should have disclaimed the complexity surrounding this issue and that stereotype threat is far from settled as myth. Additionally, if they set out to look into the achievement gap and the surrounding debate, they should've explored other theories and discussions, rather than just shoddily poking holes in stereotype threat. Additionally, even if stereotype threat can't be solved with wording on tests, its possible that it has ramifications that can't be cured simply by wording. Maybe discouragement that comes with stereotypes takes place over a longer period of time, and is more powerful and corrosive than thought, isn't that also, if not equally, possible? They mentioned this briefly at the end, saying stereotypes are harmful, but all they really give factual respect to is the assumption that stereotype threat has no explanatory power, even though this is clearly not shown in their investigation, and their earlier reference to Charles Murray is less comprehensive than to be desired. The way they frame Claude Steele at the end is someone defensive of his initial findings, and are confrontational in a dismissive way. They seem to be disdaining complexity for disturbing and simple 'radical truths' that are more grossly oversimplified than the hypotheses that they are investigating. Radiolab is hurting me right now. First Yellow Rain, then Truth Trolls, and now this. How their indulgences have changed. How quickly society can change.

Dec. 04 2017 04:42 PM
Charles Woods

While I enjoyed this episode, I'm really surprised that there wasn't a summary at the end of the episode of the current state of the replication crisis.

This issue seems a lot deeper than just one study in one field.

Dec. 03 2017 10:53 AM
Asad Khan from Arkansas

I thoroughly enjoyed listening to this podcast, specifically Claude Steele's study on gender and race-based performance differences. It was interesting that subsequent studies weren't seeing that big of a difference.

From my perspective as a child psychiatrist, one area that seemed glaringly absent in the whole conservation was the effect of early childhood trauma and its impact on cognition and things list test-taking behaviors/performance under stress.

Early childhood trauma is highly predictive of medical, psychiatric and cognitive "bad outcomes" in adulthood (see ACES study/Nadine Burke, MD's work/Resilience movie). We know that early childhood adversities can create toxic stress that inhibits cognitive performance.

Instead of looking only through the mirror of gender/race bias, it seems that a more comprehensive study might include the perspective of early childhood adversity (or controlling for early childhood adversity) rather than focusing solely on gender or race. I wonder if women or African Americans have experienced more childhood adversities--thus impacting their performance under stress.

Dec. 01 2017 11:12 PM
Ryan from Rice University

Interesting episode. It's good to hear the voices of reasonable skeptics such as Michael Inzlicht, who has himself done research demonstrating the effects of stereotype threat. Having published a number of papers (most with multiple studies in them) on this phenomenon, I can say that this topic proved to have the highest "hit rate" of any topic I have ever examined in the lab. I even have several interesting stereotype threat studies that I never got around to publishing. I think that across almost 10 years of doing research on stereotype threat, I only had 2 (maybe 3?) studies that did not produce statistically significant results. Thus, my confidence in the validity of this phenomenon is almost as strong as that of Prof. Steele. Why, then, the failures of some researchers to find strong evidence of the phenomenon? Several things could be at work. As a journal reviewer, I used to be asked to review papers on stereotype threat occasionally, and about half the time they were poorly done. But I don't think that explains why some researchers, such as Michael Inzlicht, failed to find evidence of stereotype threat. My hypothesis is more along the lines of what one well-done study on stereotype threat showed rather convincingly: that teaching people about stereotype threat actually undercuts the phenomenon! Remember, the reason this phenomenon is believed to occur is that people who are the targets of social stereotypes about their inferiority in a particular domain are distracted by concerns about confirming (or being seen to confirm) this stereotype when they experience challenging problems on a test that is relevant to the stereotype. So, arming them with information about stereotype threat can help to reduce the distraction that worrying about confirming a stereotype can produce. It's also worth noting that Inzlicht starting having trouble replicating this phenomenon, after first successfully demonstrating it, after he began doing work at U. Toronto. Because stereotype threat depends on broad social buy-in to beliefs about the inferiority of certain groups, I would not be surprised if the much more progressive Canadian culture were a terrible place to try to do stereotype threat studies. One might have to look beyond some of the bread-and-butter topics common to American culture to find evidence of stereotype threat in Canada -- perhaps examining Native Canadians, for instance, rather than women in math. Operating in the southwestern U.S., I never had much trouble tapping into test-takers' concerns about broad social stereotypes about women in math or African Americans in general intelligence domains. The social context matters -- that is, after all, what this phenomenon is all about.

I should also note that I love the idea mentioned in this episode of having serious skeptics pair up with "believers" in a phenomenon to do great research. That was, in fact, the nature of my collaboration with Eric Day that was mentioned in the podcast.

Dec. 01 2017 12:52 PM
Simon Adlers Mom's hair dresser from NYC

Good Job guys. This is more like the RL I first fell in love with!

Nov. 30 2017 10:03 AM
jader3rd from Monroe, WA

I for one, do not mind the idea that streo-bias is not a statistically significant factor. I'd rather work with, and collaborate with, someone who is being told that they're being tested, and succeeds on that test; vs someone who needs to be tricked into thinking that there are no consequences for their actions. I'd rather associate with those who succeed under pressure, compared to those who break under pressure.

Nov. 29 2017 08:26 PM
Heidi from CA

One thing I noticed was referenced, but not directly addressed, is the heavy reliance on students- specifically psych students- for research subjects. Even Prof Steele mentions many times his use of his own students in his research. If you've taken (research) university level courses in psychology, you probably noticed two big things: 1) professors and/or assistants either encourage or require that students participate in surveys/experiments run by the professor/university; 2) in the beginning of the semester the professor probably asked that students not use class time as therapy (it's common for people to be drawn to this field, at least initially, because they are looking for answers in their own lives).

Psych students are most likely already familiar with established concepts/research that are taken as assumption in the experiments they participate in, and coupled with the pressure of helping their professor and getting a good grade, they may inadvertently answer in a way that confirms the researcher's hypothesis. This could possibly be why a professor who made a career of researching stereotypes, asking his students to participate in a research project about stereotypes, would time after time find that his own assumptions were confirmed, but others without his background did not. The other problem with using college students is that they aren't a good representation of the general population- usually smart enough to get into college, often are able to pay for it, and usually hold political beliefs in line with the institution they are attending.

I don't want to go too far into the second point, because it can go too far off topic, but mention as another characteristic of the small and specific sample size/demographic.

Sorry about the wall!

Nov. 29 2017 05:23 PM
Jake from State Farm

The replication crisis is well recognized in medicine and life sciences, as well.(

The "softer" sciences (Sociology, Psychology, Biology) largely rely on population statistics to confirm or deny an outcome from an input to a poorly-understood (or black-box) system (such as the brain, or complex cell functions). That it is only now being recognized that experimenter bias can find it's way in to such a method of investigation is absurd. Correlation not implying causation is now a pop-science notion, and yet effects like stereotype threat are still "proven" by experiments with poor controls and multiple possible causative effects.

The "harder" sciences such as math, physics, and chemistry have had a higher bar of proof required for publication for nearly a century, but still suffer from ego, bad data ethics, and poor repeat-ability. It is a hard, but not impossible, problem to solve.

Nov. 29 2017 04:54 PM
Peter from Sydney, Australia

It should be noted that although psychology has had the worst of it, it is far from the only field to suffer from the problems described in this episode. After all, the root causes are human failings.

As an example, look at the work published by C. Glenn Begley, starting from 2012.
Can you imagine how much time, effort and money was wasted chasing dodgy cancer research? It's a scandal.

John Ioannidis is worth reading as well, although he takes a more mathematical perspective.

Nov. 28 2017 05:48 PM
J. Dollarhyde from California

Claude Steele: I don't think there's anything that could make me go, "Uh oh, this whole thing is not true."

In one sentence, all of his scientific authority is gone, all of his conclusions now suspect. If a distinguished researcher openly admits they would not consider data that contradicts their own findings to be valid, what hope is there to educate climate-change deniers and creationists?

Richard Feynman: The first principle is that you must not fool yourself and you are the easiest person to fool.

Nov. 28 2017 04:10 PM
Jose Velasquez

A fair summary of stereotype threat and the replication crisis.

Professor Steele has posited a reason for ongoing achievement gaps in higher education. Whether or not this holds we will see. Professor Steele himself has said the mechanism to create the stereotype threat outcome or cure said outcome is a form of priming. And priming does not replicate. So their needs to be a new mechanism if we are to considered if stereotype threat exists.

But Professor Steele is far from the worst offender on the abuse of stereotype threat in the achievement gap. His experiments held the SAT scores of the different races constant, ie they all had about the same test scores. However, you would not expect people with different test scores to excel at the same rates, yet when affirmative action throws people with quite divergent test scores into the same learning environment many people blame stereotype threat for the achievement gap.

Nov. 27 2017 04:33 PM
becca from Burien

Yes, Aaron (commenter above). If the goal of social science is to advance our awareness and bring about change then perhaps not being able to replicate a study (especially a famous one) years later would be considered a GOOD thing. Could be that educators (and even parents) who were aware of the original study have changed their approach to, say, testing and test prep and female students are no longer swayed by old stereotypes about math and are welcomed into STEM studies.

Nov. 27 2017 02:24 PM
Jim from Omaha

Couldn't this simply be a self esteem issue?

Nov. 27 2017 10:52 AM
Jonathan Allen from Tacoma

Not through more than the start of the podcast, but all you folks had to do was to cite The Bell Curve, in the context of (Oops, Contemporary) American education, and all the buttons got pushed.
I remember when that book was a best-seller, bought and summarily (and wryly appropriately) turned into a door stop by that many Republicans.
If memory continues to serve, the primary, operative thesis was that, while Africans were acknowleged not to be genetically inferior to Europeans, specifically regarding intelligence, African-Americans were.
In other words, regarding ostensible communal intelligence, an interval of less than half a millenium was somehow of less relevance in the context of (systemmatically traumatic) history than of ...the jaw drops... genetics.
Magical thinking.
As such, a remarkably concise articulation of the essence of racism in its modern, ideological forms. In order to Get pseudo-science, you need scientific vocabulary.
...To abuse. An ominous precedent for what we're seeing now, on that many fronts at once. Right, there are others, only more disturbing the further you look.

Nov. 26 2017 10:45 PM
Avi Burstein from Catskills, NY

Overall a pretty good episode, covering a very important topic in the scientific community that more laypeople should be paying attention to. However, the dismissive remarks regarding The Bell Curve and the relationship between IQ and genetics betrays RadioLab's skewed bias that seems to prevent you from approaching certain issues fairly. You guys gave the idea absolutely zero credence, claiming, "there's no scientific backing at all for the idea of genetic differences [in IQ]" when in reality there are many reputable researchers who would disagree with such a claim. For an example of one of those see the journal "Nature Genetics" from this past year, article titled "Genome-wide association meta-analysis of 78,308 individuals identifies new loci and genes influencing human intelligence". (Googling that phrase will turn it up.)

Nov. 26 2017 09:09 PM
Ethan Fischer from Chicago

I started listening to radiolab about 6 years ago when I started my undergrad in psychology. Back then, they focused more on psychology and social sciences and produced some of the most interesting podcasts I've ever heard, echoing and supplementung much of what I was learning.

About the time I graduated, the replication crisis came to light, casting doubt on a lot of my favorite experiments and findings I learned about. Ever since then, I've been feeling pretty disillusioned with psychology and a lot of what I thought I knew.

Anyway, what I really want to say is that I'm very relieved to hear Radiolab voice these concerns in this episode, especially since radiolab played a big role in my education. Seems like you guys have been producing fewer episodes about psychology these days, but it's nice to think that might be due to a concern over validity. Most of all, it's nice to know I don't have to doubt the integrity of your show because you've helped gain some trust by doing this.

Alright I'm just rambling now. Keep up the good work

Nov. 26 2017 07:25 PM
B from Philadelphia, PA

One issue I had was while describe two students, Jad said something like "They got similar SAT scores, so you expect them to do the same in school."

Says who? Who expects that? Not everyone that attends Stanford gets straight A's. Are we expecting that they all do, only because they go to Stanford?

Nov. 26 2017 04:16 PM
Steph from Toronto,Ontario

There was something a little off putting about this episode, that falls in line with other episodes of Radiolab where they try to tackle the topic of race but end up doing so irresponsibly. But that's another topic for another time.

As a black female student at a Toronto university it is fairly obvious to me what the discrepancies in the data are illustrating. The United States is still very much a segregated country (and was even more so in 1995) and university and college represented different things for different people depending on what their background is. Black people from the US are in my experience more likely to see university as end rather than a means to another end. What I mean is that many black students come from rough neighbourhoods where the pressure to escape and succeed is strong and that pressure can drive the students to push themselves to the best of their abilities. Once they enter college or university it can become a place of refuge but it also puts a bit of distance between what drove them to perform the way they did.

While I'm sure many many white students view university in the same way I think that once you pair the background of black students with Claude Steele's work on the stereotype effect you can come up with a better picture of why this is the case. Because even if you're a white student that comes from a rough background it isn't assumed by your appearance the same way it is for black people.

The reason the effect is different in Toronto is because of Toronto's immigrant population. Black people from Toronto overwhelmingly come from Caribbean or African backgrounds meaning countries with either homogeneous racial makeups ones where people of colour are in power. They don't suffer in immediate psychological effects of societal structural racism in same way that black people in the US do.

Also, as was pointed out in the episode, Toronto has a fairly large East and South Asian population which would skew the outcomes seeing as east Asian and south Asian men and women face a set of stereotypes that are completely different if not totally opposite to that of black men and women especially in regards to intelligence.

Nov. 26 2017 01:59 PM
Meegan from Australia

Great podcast as always! Keep up the great work.

Nov. 26 2017 01:47 AM
Samuel A Greenspan from Portland

To add on to the previous comment, stereothreat seems akin to learned helplessness. The tester induces a stereotype meant to shock the test taker and hamper performance. Yet changes in the environment (awareness, media campaigns) and individuals (more diversity in populations) over time may change their expectations and perceptions of a tester trying to induce stereothreat. Perceptions and interpretations of a stimulus vary between individuals, particularly as cultures grow more complex.

Yet just because sometimes isn't true from now until the end of time doesn't mean it isn't worth studying. Psychology has it's flaws like any other science, ideology or dogma. We can only hope to use this information to improve methodology and better understand study limitations.

Nov. 25 2017 03:29 PM

When I was an undergrad, I worked in a developmental psychology lab. I did well initially and was entrusted to run a small replication study. I was proud to have the opportunity. However, when I was unable to reproduce the initial results, the grad student I was working for as well as the PI of the lab began pressuring me to reanalyze the data. When that didn't change the results, they wanted me to go back and recode the data. When that didn't change the results, they began holding weekly meetings with me to criticize my efforts in what were clearly disciplinary-type meetings. After a couple of those, I got the hint and quit the lab. I work in science now, but not in the field of psychology, which I wouldn't go back to for anything.

Years later, I read the story of Marc Hauser with great interest. Some of the tactics he used to pressure people in his labs (particularly young and vulnerable undergrads) sounded very familiar. I am wondering how many others had an experience like mine and if this could be part of the irreproducibility story.

Nov. 25 2017 12:23 PM
Simon O from Canada

Hi guys,

Great show. The problems highlighted with the social psych literature are very similar to those addressed in medicine a few decades ago, still being grappled with today, namely publication bias and outcomes switching. Medicine has developed specific critical appraisal
Methods for studies and it sounds like many of the sissies described in the episode would not be of such a magnitude to in medicine as a result. Clinical trials have to be registered and prespecify their outcomes before many research ethics boards (mine included) will even consider them for review. Pharma companies and other researchers do brush “negative” trials under the table and select the outcomes that best represent their interests. When researchers and clinicians interpret the evidence or summarize it in a systematic literature review, it is standard practice to review the trial registries and protocols to ensure that all data (whether or not it has been published) can be included. There are statistical tests to look for publication bias as well. In light of this, it boggles my mind that this whole field has such little infrastructure to support transparent and prospective reporting of study procedures and results.

I am thankful I no longer have to do my power poses or restrain myself from eating the marshmallows on the table in front of me!

Nov. 25 2017 03:08 AM
Max from Peekskill, NY

oh good i don't have to drop the first ... lets say opposition, comment.
i Loved the show. spotlighting the lack of scientific rigor Stereotypical to the "soft sciences" by way of a professor criticized recently for enabling sexual predation. confirmation bias, "part of the problem", statistical trends =/= determinism. magnificent.
Jad, Krulwich, all yall. such love. six more years gang, such good work.

and yet, there's four lights.
the supreme step? the next one.
semper disquirunt

Nov. 25 2017 02:18 AM
Nils Bohr

Between this episodecand the last, it should be obvious to the scientifically minded that radiolab and their progressive, mind numb robots are interested in story telling over science.

Psychology and sociology, as the replication crises imply, are jokes. It's no coincidence that leftists have turned science into a relgion yet fail to create a morale imperative. This is what happens when your founder uses women and their uteruses as scapegoats.

Nov. 24 2017 11:45 PM
Jason from Detroit, MI

Ha! Claude Steele sounds kinda like a really smart version of Carl Spackler (aka Bill Murray) from Caddyshack.

Nov. 24 2017 07:48 PM
Sarah from State college

Aaron- I was thinking the same thing! Maybe the manifestation of stereotype threats exist in different ways that haven’t yet been tested.
(e.g. race and computer literacy) Thanks for another great episode!

Nov. 24 2017 06:22 PM
Aaron Pettigrew from vancouver, canada

What a great episode.

It's really interesting to me, though, that there's so little mention of the possibility that the effect of stereotype threat might just be diminishing over time. It seems possible that the effect was strong for university students in the 90s, and it is becoming less-strong (or maybe changing, as Steele hints) as awareness of stereotypes has increased in educational institutions and interventions have perhaps begun to take effect.

Anyway, great stuff!

Nov. 24 2017 04:36 PM

Leave a Comment

Email addresses are required but never displayed.

Supported by