Impostor Syndrome vs. the Dunning-Kruger Effect
raptori.dev
raptori.dev
When I did a coding bootcamp long ago in a career switch I couldn't help but look around the room and think "Hey ... some of these people really shouldn't be arbitrarily encouraged anymore."
It's a rough thing to think / do, but also maybe would have been for the best for many of them.
To their credit I talked to some folks who were in the camp a few years after me and they reported that the class sizes naturally shrank as the classes went on. Apparently students were being better evaluated and had ways "out" that encouraged a little more "yeah maybe this isn't for you" type of evaluation.
I have been told I have imposter syndrome by many people. I get all the "You are better than you think you are" and "Trust me, I've seen worse than you."
Regardless, I think I have a pretty accurate assessment of my abilities. I have been professionally programming for 7 years (10 years total) and have a degree in CS.
That being said, I have no fucking idea what I am doing 99% of the time. I just follow guides and documents, and just glue libraries together like a cyber plumber. I probably couldn't even reverse a string by hand if my life depended on it. My job has a pretty low bar for what is considered acceptable, so I have been able to fly under the radar with ease all these years. I'm not being modest, I am being serious. The jaws of many users here would hit the floor if they saw some of our practices.
However, the good news for me is that I do not have to remain like this. Over the past few months or so, I have really tried to increase my knowledge by diving face first in books, lectures, and other resources. I feel like I have improved significantly in this short amount of time. I mean, I am basically starting all over again, but hey, I can maintain the status quo or I can do something about it.
I guess my whole point is that, one may be an "imposter" and/or incompetent, but it's also one's choice to remain as one.
> Failure is not something to be feared—it’s an opportunity to learn and improve.
These pithy quotes can be good for shaking someone out of a perfectionism-induced freeze, but they take the concept too far.
Failure is not a good outcome and should be reasonably avoided to the best of your abilities. You can and should learn from it, but removing all fear of failure isn't realistic.
This doesn't mean you should never start anything that might fail. I think that's what these quotes are trying to get people to avoid. Instead, they set unrealistic expectations about how failure can't/shouldn't/won't have any consequences when, in fact, it often does.
Likewise, making some mistakes is normal. Nobody is perfect. However, mistakes have consequences. It's how you handle and even prepare for those consequences that makes the difference. It's true that you shouldn't be so afraid of mistakes that you can't get yourself to take any risks or do anything, but you shouldn't go so far as to imagine mistakes as having no consequences. If someone is making mistakes at an unreasonably high rate, something needs to change: More training, more supervision, more education, or if all those fail, try a different role.
These little feel-good quotes might shake some people out of being too afraid to try anything, but they also mislead a lot of people into unrealistic ideas about how failure and mistakes shouldn't have any consequences.
Failure is not normal and it is speaks badly of you no matter what's the reason. Most people including me are not perfect and it will happen but that doesn't mean failure is alright, that means that we're an problem.
The only thing one has any right in their life to ever blame are themselves.
Clearly, the word "failure" in this context is insufficient, because we all would agree that some failures are unreasonable and some are an absolutely normal and necessary part of personal growth.
man, I'm so glad you cleared that up for us...
If it will take me 50x tries to do something and someone else of that same background will do it in 25x, that means I was doing something utterly wrong.
Either way, if your 49 failures were unacceptable, then so we their 24 failures. This conversation was never about comparing ourselves to those who are better or worse, but about the meaning and acceptability of "failure".
I disagree that "failure is bad and should be avoided". The degree to which failure is bad and should be avoided depends entirely on the gravity of the consequences. When taking on an ambitious programming project, "failure" is just an unfinished project and a loss of time. It's clearly a different world than something like quitting your job to try to become a professional actor.
Judgement shouldn't be based on whether somebody is failing at something or how many mistakes they are making, but the comparison of the positive vs the negative consequences. When learning something new, you will always make many more mistakes.
The problem with both the pithy quotes and your response (which is, I think, largely reasonable) is that neither one actually defines any criteria for success or failure, or what is an "unreasonable" number of mistakes. It just ends up being a semantic disagreement over what they meant by "failure" and what you understand by "failure", and I think you'd probably end up largely agreeing in the end if you broke down and actually explicitly defined the meanings of the words used.
You should definitely try things that might fail.
However, failure within those projects should be actively avoided.
These pithy quotes that try to remove any downside from failure are misleading. In business, failure has consequences. You need to understand what's at risk and make a best effort to steer the project toward success.
These over simplified quotes that try to strip all of the downsides away from failure are a cheap trick. Yeah, they convince some people to get past their fears, but they also mislead other people into thinking that there are no consequences.
You need to understand the risks and consequences, not just wave them away with pithy quotes from someone's blog.
I agree with that. I agree with all of what you're saying. I don't think the pithy sayings are telling people to actively fail, though. The point is to not become overly avoidant of taking on ventures because of a fear of failure. And I think the avoidance of failure is a bigger thing than you're making it out to be. Most people who I know who don't achieve their goals and dreams fail to do so because they simply never start, or they start and see how hard it is going to be and then give up. I know very few people who jump headlong into things and fail often, and they are also the people I know with the most successes in their lives. Anecdotal, but the fear of failure preventing people from attempting anything that they want to achieve is a real and very common thing.
> To find the sweet spot between impostor syndrome and overconfidence, you first need to understand that you can make mistakes without it reflecting on your competence.
is incomplete without describing the environment and people you are around when you make these mistakes. The teams I have worked with until now have all been extremely kind in dealing with other engineers making mistakes and focused on how we could reduce such mistakes as part of a team. It helped me build a mental checklist on dealing with such a situation and when/how to take the risk and when to bring in additional help. I never really noticed the impact of this until I worked with a colleague who seemed to have trouble releasing their work to production and always kept delaying things out.
- Your partner says they're ok with trying a threesome.
- Your partner says they're ok with trying that new restaurant.
to bring it more inline with tech.
software developers need to be able to safely break things, that's why they don't work directly in production. a hotfix bypasses the normal processes and goes straight to production. It's not that you don't do it, you don't do it without an acknowledgement that the reward outweighs the risk and because you respect the risk you take care to minimize mistakes.
Works well for impostor syndrome folks. But I think what the people in the back (the DK folks), need to consider is that you can also do well without it reflecting on your overall competence.
It's more about putting ego aside altogether. You're a mushy piece of meat that thinks. Feel good about your accomplishments, feel bad when you fuck up, learn and move on.
There is an actual Imposter Syndrome that afflicts people who are highly qualified/credentialed, people who have qualified for entrance into elite programs, and put in the time and effort to progress to graduation; and yet for psychological reasons, they have great self doubts about their "worth". This is akin to depression. (a third party would look at you and say "wow, you are super competent and qualified, it's terrible you feel you aren't," and no pep talk they give you is going to help you.)
Then there are bouts of insecurity and guilt whereby any of us can feel like "I fibbed on my resume and interview to get this job, and I now I'm surrounded by people who know what they are doing, I hope nobody figures out I snuck in." This is akin to sadness. (a third party would look at you and say "ha ha serves you right, but it's always fake-it-till-you-make-it, everybody feels this way to start out, just grit your teeth, you'll be fine," and that pep talk will help you.)
Everybody experiences sadness; sadness clears up and goes away. Not everybody suffers clinical depression, and it does not go away on its own.
Dunning Kruger? it's not really an Effect, it's an observation/tendency that is explained by simple logic along the lines of, "you don't know what you don't know".
In many cases, depression does in fact go away on its own.
If it were simply the idea that you don't know what you don't know, then you would expect to observe over confidence in people who know almost nothing about a topic. I believe D & K observed a sort of bell curve. (actual knowledge level on the x axis and confidence of knowledge on the y)
I wrote a simple phrase to refer to the "tip of an iceberg of associated ideas". To illustrate what I mean, you could rephrase your more detailed description of the gaussian to use "my" phrase to say "these people don't know anything, these people know more but don't know what they don't know" and these people know even more but do know what they don't know" etc.
The former is a problem because it often comes from a lack of self-awareness (and peers willing to speak up when you faulter). The latter can be a problem because it prevents you from fully offering what you have to offer.
Of course there's a sweet spot, but if you had to favor one or the other, lean towards Imposter. When you're in DK-mode you're far more likely to make "bad yeses" and those can be painful.
I believe that impostor syndrome is a lack of overconfidence in one's abilities- that implies that you have the ability to assess competence and to assess the skills needed for a job or task, and you don't default to "I know I can do it well". When you encounter a novel job or task, you don't immediately assume competence, and you will remain uncertain until you fully understand the requirements of the job or task. A lot of impostor syndrome comes from "don't fully understand what's required, but it looked complicated from the outside", and will go away with greater understanding provided you actually have the skills and the competence to self-assess.
If you lacked the ability to assess competence in yourself, then you likely would not have impostor syndrome.
I don't see why it's a balance. Impostor syndrome typically lessens over time as you become familiar with the job/task and the requirements necessary to perform well, and are better able to assess your own skills against those requirements.
If impostor syndrome doesn't go away with time, maybe you are correctly assessing that you don't (yet) have the requisite skills.
If you never get impostor syndrome, then you either aren't challenging yourself or you might not have the subject area competence to assess your own skills level vs. the jobs/tasks you take on, e.g. Dunning-Kruger effect.
It's one thing to know that DNS travels over port 53, but leveraging that in troubleshooting is harder, since it may involve a firewall, switches, trunked ports, servers, workstations, etc.
It just takes time to learn the practical side and that's rarely taught in schools from what I've seen.
YMMV.
I'm not disagreeing and I can allow that IS a natural and expected safety valve.
That said, I suggest that some individuals are better off without IS operating. They're less-than-qualified but they commit because they believe they'll figure it out in time.
For low-resource individuals, this can be the least-unlikely path from destitution.
Life is a frontier if you keep moving. As you progress in life you have fewer yardsticks to judge yourself against. At kindergarten we all chant in unison. By middle school you follow what the teacher says, but some people learn differently. It's probably the last "Standardised" (SAT) test you ever take.
At high school you are expected to read books by yourself and you choose different books than others. By university it's mostly hands-off, and only the lectures are common amidst diverse self-study. At PhD you know everything about a vanishingly small field, and by definition the doctoral graduate is the singular world expert on their tiniest corner of knowledge.
Each of us on a journey from a common place to a unique viewpoint in life, that can be quite lonely.
Regardless of whether you follow an academic path I think this applies to everyone. At the start it's easy to see how you measure up to others. By 40 you need to work hard to find smarter people to challenge you and you're not sure where you stand. You see that being smart or confident isn't all that, and it ruins some people. You see people you thought were smart do really dumb things, and people you considered below you race ahead by making wise life-choices.
I like the article because it hints at an important truth, that you can simultaneously be too self-assured and not enough. Our value systems for judging ourselves are weak in this society.
But as the objective clarity of that sweet spot wanes, something else kicks in as you age, which is a kind of contentment/wisdom. You care less about where you stand and are comfortable with mistakes. A few regrets adds character. Financial or institutional success are no longer good indicators as you start to turn to more spiritual ways of knowing where you stand.
> I believe that impostor syndrome is a lack of overconfidence in one's abilities - that implies that you have the the ability to assess competence --
I think that competence in an ability is not perfectly synchronized with ability-assessment-competence.
I think ability-assessment-incompetence can result in both under- and overconfidence, before accounting for a self-esteem bias. Being underconfident is also an incorrect self-assessment, which doesn't imply ability-assessment-competence.
> If you lacked the ability to assess competence in yourself, then you likely would not have impostor syndrome.
Here I disagree the most. I think impostor syndrome can be the result of ability-assessment-incompetence, somewhat detached from their true competence in the actual abilities involved.
Meanwhile, at home, I would practically be shaking with excitement waiting for the show to start, I so appreciated it.
This tale somewhat suggests that people who perceive themselves as incompetent and inauthentic might spend more of their time striving at work, which could raise the bar, maintain a high standard and eventually breed something resembling confidence.
Or it could just continuously undermine their natural confidence and sense of self-worth and debase them such that they are easy to overwork and manipulate. It can also just feed into fears that invalidate the satisfaction of any jobs well done, leading to burn out and feelings of futility.
People who are overly confident can behave brashly and do damage, while automatically imposing costs on others, in the form of the time it takes to crack through their false beliefs or the duplication of effort it takes to walk back their mistakes.
So, this is kind of a nothing post, basically a lament. It's not clear whether suffers of Impostor Syndrome or Dunning-Kruger type symptoms have an easier path to a more moderate position, but each one seems likely to be rampant in just about any workplace.
Back to Dunning Kruger, it's largely a statistical trick. Wiki actually has a pretty decent section on this [1], alongside some nice examples. The gist of the trick is that the more random a sample is (in other words the more people are unable to accurately self evaluate) the stronger a Dunning-Kruger effect you'd show. The way Dunning-Kruger achieved the really beautiful perfectly linear graph people always reference is by having people self evaluate their ability to... determine a funny joke. And so broken was their test that they had to eliminate one of their 8 professional comedian judges, because he couldn't determine a funny joke, by the standards of their test. And the remaining 7's scores were judged, by the paper itself, as "moderately reliable."
And I suspect this was intentional. The first thing they would have done is to plot a scatter graph of their results. And they would certainly have quickly realized the transform (by bucketing) and the scatter plot tell radically different stories (again the illustrations on Wiki make this more than clear), but only one of them is true!
[1] - https://en.wikipedia.org/wiki/Dunning%E2%80%93Kruger_effect#...
It is a real thing for humans.
https://www.sciencedirect.com/topics/psychology/overconfiden...
I know plenty of C-suite folks who must be in the minority then.
The CEO babbling about how AI "works" must be knowledgeable and competent, and I, an engineer actually working on our AI product, am the idiot.
Thanks for the help!
Perhaps more of a poor leadership thing we observe.
Like Teddy Roosevelt said - 'Speak loudly, aggressively flail about, and you'll go far.' Or something like that.
Meta: what if this researcher themselves is exhibiting Dunning-Kruger and is overconfident.
“If you are really, really stupid, then it’s impossible for you to know you are really, really stupid.”
Typically it is used to describe when people who don't have complete knowledge about the nuances of a problem or have opinions based on bad information or bad assumptions confidently or aggressively state positions and that confidence is confused with competency.
LLMs are a good example of the colloquial use. They are confident, superficially competent but almost always wrong in specific domains.
It is a problem when people underestimate their blind spots.
Building culture that allows all coworkers to safely raise consent and learning to depersonalize feedback helps with that.
It may not have as popular of a label, but it is well documented in fields like human factors.
Even in doubt you're still doing some mental gymnastics that allow you to place yourself above your peers. It's inherently self centered and arrogant way of thinking about yourself. The opposite of self aware.
A more realistic "impostor syndrome" is the IT idea of 'fake it till you make it'. In other words, its a tacit admission you dont know shit and you're googling for ideas until you get something that sticks. And yeah, most IT is like this. Even Stack Overflow is a meme - search here and copy/paste regardless the quality.
And, well, it's not impostor syndrome if you're a know-nothing idiot.
And the crap about Dunning-Kruger.. It just serves a way to shut experts up and dismiss their experiences and knowledge. Its basically the pseudo-science crap way of saying "My quiet ignorance is better than your direct assurances".
Then again, business types also believe in shit like forcing all employees in Myers–Briggs Type Indicator crap, and then comparing that across everyone. And if you know anything about psychology, you don't do longitudinal comparisons of this sort of crap.
Dunning-Kruger certainly has reached that point, you have to really dig to figure out if they're just throwing the phrase around or they're applying it in a thoughtful manner.
All sorts of phrases end up that way. "virtue signaling", a phrase I've never been a fan of has ended up that way too.
And now the meta-question: did I just…
I get the idea that it usually implies that the person doing so is being insincere but I honestly think the better way to do that is just say they're being insincere and demonstrate why.
Without showing why they're insincere it seems like it is just an empty phrase.
I'm not saying people don't belong, but that the feeling of insecurity, especially among young people, is more often than not justified.
IMV DKE pop-psych is popular because it enables a kind of second-order superiority complex where you can say shit like, "Research shows that everyone who thinks they're smart is actually a huge dumbass. Actually the really smart ones are the anxiety ridden imposter syndrome devs like myself!"
At least overconfidence and naked condescension required some courage. The insufferable new meta-game of superiority signalling seems to be feigned humility and hiding ego behind the affectation of "knowing how little you know."
Background: I'm a software developer that has worked for 30 years professionally and has used computers for at least 40. Languages and technologies range across pretty much everything popular (and many that were not).
I've always considered myself intelligent to very intelligent. I own lots of books, love learning, etc. So much so that I considered it a deep part of my self-image. Except while attending a very good engineering school, I generally considered myself the smartest person in the room (yes, insufferable, really).
Now, internally (stored in one of those places in your mind that you know are there but you just pretend isn't), I knew this wasn't the case. I have a terrible memory and have always seemed to have to work at least twice as hard in math and sciences and still only had a vague understanding of what was going on. Even to this day I can tell you a lot of findings about the fundamentals of computer science theory but I couldn't explain them to you except in broad sketches. And if you asked me specifics about a book I just read I would likely draw a blank. That said, I've had very successful career as a developer.
Recently, I began playing Go more seriously after playing casually for years with another (much higher ranked) player. I've joined a local group, play online regularly, study books, etc. While I previously thought I was a pretty good amateur I was quickly disabused of this notion. I was beaten handily by almost everyone I played. Worse, studying made almost no difference. This caused me a bit of a intellectual breakdown. It was as if what I always knew deep down was actually true: that I'm NOT particularly smart or gifted. That I'm at very best "average" if not below. Go shined a big, bright floodlight on all of those insecurities that I hide from myself. My entire self-image was shattered.
I'm still picking up the pieces of this realization but over the past few months it has radically changed how I deal with other people as well as how I perceive myself. I think I'm a little more understanding, a lot more humble, and maybe a bit more "human". I'm still a successful software developer, but now when I'm studying something new I'm not doing it out of some misguided belief that learning it will somehow make me better than other people, but because it will make me a better software developer that can help solve problems better. I don't treat knowledge as something that I can lord over people to show that I'm smarter than them. And I seem to assume better of people than I did before.
And I still play Go, poorly.
Yeah I didn't need to read the rest (did anyways) to know where this was going... I played Go with my father in law once. He said to me:
"What, I thought you were supposed to be smart?"
It was indeed humbling. I do not still play Go.
The first problem is the "effect" can be discovered in purely random data, because it's basically just a statistical artefact. There's a description of what went wrong here [1] and a paper from 2017 here [2] ("Our data show that peoples' self-assessments of competence, in general, reflect a genuine competence that they can demonstrate. That finding contradicts the current consensus about the nature of self-assessment"). In other words it doesn't replicate when the data is handled correctly. By itself that problem means DK should not really be cited for anything.
But even if you ignore that:
• The original study does the usual pop psychology thing of extrapolating from a tiny self-selecting sample of American college students to all of humanity.
• It has a stupid design in which the students are asked to rate how funny jokes are, with some local comedians being the "experts", and this is presented as a measure of social skills i.e. simple disagreement on what's funny is sufficient to get you dinged as incompetent-but-doesn't-know-it in DK's research.
• It also has a problem with circular logic where they found that one of their hand-picked joke experts disagreed with the others on what's funny, meaning that comedian would have been considered one of those idiots who thinks they're an expert, but they were picked by Dunning & Kruger specifically because they were an expert. This shows that the nature of many tasks is subjective and certainly the one they'd chosen is subjective, but instead of fixing their definition of expertise they just tossed the guy out of the expert pool.
(the study does have two other tasks, but there's not enough detail in the descriptions to know if they had similar problems)
Dunning wrote an attempted rebuttal in 2022 [3], but it is unable to directly attack the statistical claims and instead argues that because there have been replications, the effect is real. But this doesn't make sense because if you replicate an invalid methodology you'll get the same results; this doesn't make them correct. He also argues for it by presenting examples of "skills" that are simply brain teasers/well known gotchas, rather than actual skills one might practice.
[1] https://fortune.com/2023/05/08/what-is-dunning-kruger-effect...
[2] https://digitalcommons.usf.edu/numeracy/vol10/iss1/art4/
[3] https://www.bps.org.uk/psychologist/dunning-kruger-effect-an...