TE
TechEcho
Home24h TopNewestBestAskShowJobs
GitHubTwitter
Home

TechEcho

A tech news platform built with Next.js, providing global tech news and discussions.

GitHubTwitter

Home

HomeNewestBestAskShowJobs

Resources

HackerNews APIOriginal HackerNewsNext.js

© 2025 TechEcho. All rights reserved.

The Dunning-Kruger effect is autocorrelation

627 pointsby ljosifovover 1 year ago

71 comments

tempestnover 1 year ago
I don&#x27;t buy this take, and this rebuttal does a better job than I could of explaining why: <a href="https:&#x2F;&#x2F;andersource.dev&#x2F;2022&#x2F;04&#x2F;19&#x2F;dk-autocorrelation.html" rel="nofollow noreferrer">https:&#x2F;&#x2F;andersource.dev&#x2F;2022&#x2F;04&#x2F;19&#x2F;dk-autocorrelation.html</a><p>Basically, this autocorrelation take shows that if performance and evaluation of performance were random and independent, you would get a graph like the D-K one, and therefore it states that the effect is just autocorrelation. But in reality, it would be very surprising if performance and evaluation of performance were independent. We expect people to be able to accurately rate their own ability. And D-K did indeed show a correlation between the two, just not as strong of one as we would expect. Rather, they showed a consistent bias. That&#x27;s the interesting result. They then posit reasons for this. One could certainly debate those reasons. But to say the whole effect is just a statistical artifact because random, independent variables would act in a similar way ignores the fact that these variables aren&#x27;t expected to be independent.
评论 #38416964 未加载
评论 #38417571 未加载
评论 #38421875 未加载
评论 #38416264 未加载
评论 #38420608 未加载
评论 #38416567 未加载
评论 #38418980 未加载
评论 #38417530 未加载
评论 #38420526 未加载
评论 #38418986 未加载
评论 #38419149 未加载
评论 #38419793 未加载
评论 #38417811 未加载
评论 #38429665 未加载
评论 #38417883 未加载
评论 #38417237 未加载
评论 #38416187 未加载
评论 #38418258 未加载
bitshiftfacedover 1 year ago
The authors did &quot;X - Y vs X,&quot; but that&#x27;s not even the biggest problem. The authors subtracted two measures that had been transformed and bounded from 0 to 1 (think percentiles). What happens at the extremes of those bounds? How much can your top performers overestimate their performance? They&#x27;re almost at 1 already, so not much. If they were to overestimate and underestimate at the same rate and by the same magnitude in terms of raw values, the ceiling effect on the transformed values means that the graph will make it look like they underestimate more often. The opposite problem happens for the worst performers.<p>See &quot;Random Number Simulations Reveal How Random Noise Affects the Measurements and Graphical Portrayals of Self-Assessed Competency.&quot; Numeracy 9, Iss. 1 (2016), particularly figures 7, 8, and 9.
评论 #38417205 未加载
评论 #38416895 未加载
评论 #38416672 未加载
评论 #38420701 未加载
评论 #38419163 未加载
评论 #38416673 未加载
评论 #38423645 未加载
r0uv3nover 1 year ago
The discussion between Nicolas Boneel and the author in the comments of the article is interesting and Nicolas expresses the doubts I had when reading this. The whole point of the DK effect is that people are bad at estimating their skill, so if you assume that they randomly guess their skill level then of course you will replicate the results.<p>The correct model for a world without DK should be something like (estimated test scores)=(actual test scores)+noise, and then the only form of spurious DK you&#x27;d expect is caused by the fact that there&#x27;s a minimum and maximum test score. But this effect would be proportional to the variance of the noise, and I assume the variance on the additional dataset is too low to fully understand the effect seen there.<p>Also, in this model on average everyone should still guess correctly in which half of the distribution they are, but even the bottom quartile seemed to estimate their abilities as above the 50th percentile
评论 #38417691 未加载
评论 #38416369 未加载
评论 #38419031 未加载
评论 #38416391 未加载
snarkconjectureover 1 year ago
Nonstandard terminology warning: the author is using &quot;autocorrelation&quot; in a way I&#x27;ve never seen before. There is a much more common usage of &quot;autocorrelation&quot; to refer to the correlation of a timeseries with itself (shifted by some amount).<p>If you use autocorrelation to refer to the thing in OP, you&#x27;ll probably confuse people who know statistics, and vice versa.
评论 #38418896 未加载
评论 #38418295 未加载
评论 #38419359 未加载
评论 #38418807 未加载
randomizedalgsover 1 year ago
Consider the imaginary world that the author describes, in which people&#x27;s estimate of their score is independent of their actual score. Wouldn&#x27;t it be fair to say that, in this imaginary world, the DK effect is real?<p>The point of the effect is that people who score low tend to overestimate their score and people who score high tend to underestimate. Of course there are lots of rational reasons why this could occur (including the toy example the author gave, where nobody has any good sense of what their score will be), but the phenomenon appears to me to be correct.
评论 #38422355 未加载
评论 #38420718 未加载
CalChrisover 1 year ago
The article&#x27;s definition of <i>autocorrelation</i>:<p><pre><code> Autocorrelation occurs when you correlate a variable with itself. </code></pre> Wikipedia&#x27;s definition of <i>autocorrelation</i>:<p><pre><code> Autocorrelation, sometimes known as serial correlation in the discrete time case, is the correlation of a signal with a delayed copy of itself as a function of delay. </code></pre> Of course, 0 delay is the trivial case of time delay but really, the article&#x27;s definition is at best inaccurate. D-K has nothing to do with time delay and calling it autocorrelation seems like a weird pun that doesn&#x27;t quite land.
评论 #38426961 未加载
James_Kover 1 year ago
I think the issue here is a confusion about what &quot;bias&quot; means. If they are self-assessing at random, then the high performers will all underestimate themselves, but this is not a bias towards underestimation as they are choosing randomly.<p>That said, the chart from D-K seems to show a different bias and line up roughly with what you would expect. Someone with no knowledge assumes they are average skill and hence inflates their position, someone who is very good doesn&#x27;t want to rate themselves the best because they assume others know as much as they do. The assumption underlying both groups is that you are normal and others are similar to you.<p>I hypothesise that most people think they&#x27;re average, which is something you could easily test by asking them to rate how well they think the average person would do on a test and comparing it to that individual&#x27;s test score. I&#x27;m almost certain that high performers will overestimate the average, and low performers underestimate it.
abnryover 1 year ago
If there is a linear relationship between test score (X, ability) and test score self-assessment (Y, self-perception), then the random variables are modeled as:<p>$$ Y \sim aX+b+N $$<p>Where N is some statistically independent noise, mean zero.<p>This means the covariance between them is<p>$$ Cov(Y-X,X) = E[ ((a-1)X+b+N -(a-1)E[X]-b) (X - E[X]) ] $$<p>Which is<p>$$ Cov(Y-X,X) = E[(a-1)(X-E[X])(X-E[X])] + E[N(X-E[X])]= (a-1) Var[X] $$<p>To get a &quot;DK effect&quot; we need (a-1) &lt; 0, or a &lt; 1. If a=0, in the case of the blog post, then this is absolutely true. If a=1 (which, along with b=0, is the ideal scenario), then this is barely not true. If a &gt; 1, then we&#x27;d have a whole new effect about arrogant experts.<p>So the only thing that matters from this &quot;auto-correlation perspective&quot; is the rate at which an individual&#x27;s self-assessment increases with their ability. As long as they underestimate the increase, a &quot;DK effect&quot; will occur.<p>However, in the above analysis, we ignored the variable b. If a = 0.8 and b=0, we&#x27;d never have the so-called &quot;DK effect&quot; even though it matches the &quot;auto-correlation perspective&quot; because everyone would underestimate their ability.<p>This tells me that the value of b matters. It is sort of like the prior ability everyone assumes they have. What the DK papers shows is that b &gt; .5, which I think is in line with the spirit of the popular interpretation of the &quot;DK effect&quot;. People should not be assuming they have, at a minimum, a capacity higher than the average.<p>At the same time, the value b isn&#x27;t insanely higher than .5, which also makes me want to cut those unskilled and unaware some slack. It &quot;seems reasonable&quot; to assume your baseline is average. That can&#x27;t be the case, but it feels intuitive.
19f191tyover 1 year ago
That is not an autocorrelation. The OP is equating linear dependence with autocorrelation, which not how we use that term. Autocorrelation is when a random process is correlated with time lagged version of itself.
pmavrodievover 1 year ago
Noone seems to have read OP&#x27;s post in its entirety. A crucial point was made by referencing this paper: <a href="https:&#x2F;&#x2F;digitalcommons.usf.edu&#x2F;cgi&#x2F;viewcontent.cgi?article=1215&amp;context=numeracy" rel="nofollow noreferrer">https:&#x2F;&#x2F;digitalcommons.usf.edu&#x2F;cgi&#x2F;viewcontent.cgi?article=1...</a>.<p>Figure 2 in this paper shows the result of an experiment where skill and perception of one&#x27;s skill are measured independently. To eliminate any statistical artifact of auto-correlation. And lo and behold - on average skill is uncorrelated to the accuracy one&#x27;s own assessment. No DK effect at all. What does show up actually is that more qualified people are more consistent in estimating their skill (i.e. their assessments are less variable), but the mean accuracy is still 0.<p>So indeed, on average actual and perceived skills are uncorrelated. That&#x27;s exactly what the numerical proof with random numbers shows and why in many cases we apply Occam&#x27;s razor.
hn_throwaway_99over 1 year ago
Previous discussion: <a href="https:&#x2F;&#x2F;news.ycombinator.com&#x2F;item?id=31036800">https:&#x2F;&#x2F;news.ycombinator.com&#x2F;item?id=31036800</a>
beinvisibleover 1 year ago
How does the author miss the fact that in his graph the self-estimation hovers around the 50th percentile (as expected with random data), whereas in the DK graph it averages around the 60th to 75th percentile? That <i>is</i> a significant bias.<p>As a control, the author should have plotted Fig.8, but based on DK&#x27;s data (or at least estimate how Fig. 8 would look like based on the result in Fig. 3). Then it would have been obvious that the self-estimation error tends <i>more</i> towards overestimation in the lower quartiles than ist does towards underestimation in the higher quartiles - which is exactly what DK&#x27;s conclusion was.<p>In essence: this article actually <i>confirms</i> that DK is not just a random statistical artifact.
Spiwuxover 1 year ago
At the risk of sounding like a complete idiot, isn&#x27;t the hypothesis of the original paper still true? Let&#x27;s assume self assessment score is perfectly random between 0% and 100%, so on average every group will always estimate themselves to be 50% correct<p>Then by definition that means people who are unskilled and often incorrect will overestimate themselves, while people who are often correct will underestimate themselves. Take a complete idiot for example. You always get 0% test score. Yet your self-assessment is random between 0% and 100%. Hence you overestimate yourself much more often than people who always get 100% test score.<p>In fact, if the two are uncorrelated, then that still means that<p>1) Idiots don&#x27;t recognize they&#x27;re idiots<p>2) Skilled people don&#x27;t recognize they&#x27;re skilled
dmbcheover 1 year ago
Isn&#x27;t it ironic that they fooled themselves?
评论 #38416569 未加载
gololover 1 year ago
I disagree. Dunning Kruger is not a statement about predicted score correlating with actual score in some way. It states that predicted score does not correlate well with actual score. This can be rephrased as the prediction error having a negative correlation with the actual score. The article then claims that this negative correlation is autocorrelation. That is true but the correlation still exist. The thing is that ideally we EXPECT there to be no correlation of the prediction error with the actual score, but we find autocorrelation. Going back to variables where this autocorrelation is not there, we EXPECTED to find a 1:1 positive correlation between predicted score and actual score but find no correlation, or a weak correlation.<p>So finding autocorrelation when you expected to find no correlation is pretty much the Dunning-Kruger effect here.<p>In fact their example with the random data totally makes sense: Suppose people uniformly randomly estimate their performance. Then the people who are low skilled will consistently over-estimate and the people who are high-skilled will consistently underestimate. Of course there is no causation here, as the people choose randomly, but there is an undeniable correlation. I guess the question is if you view the Dunning-Kruger effect as a claim to low skill CAUSING positive prediction error, or just correlating with it.
xanderlewisover 1 year ago
Naïve take: I’ve always felt like Dunning-Kruger is just the result of the fact that when guessing the value of anything people tend towards some common mean, and so if the true value is low your guess tends to be high, and vice versa. This assumes nothing about what is being guessed, but does assume (perhaps wrongly) that there is a commonly believed mean value and that people tend to imagine they are close to it.
评论 #38416668 未加载
zw123456over 1 year ago
I know I&#x27;m not smart enough on statistics or psychology to evaluate the article but it always struck me that D&amp;K seemed to say something similar to what my grandpa said when I was a wee lad, &quot;The more you know, the more you realize how much you don&#x27;t know&quot;, I know he wasn&#x27;t the first person to say that, but he was the first person to say it to me. I don&#x27;t know if D&amp;K is autocorrelation or not, but I know that an awful lot of people seem to think they know more than maybe they actually do, probably me included. Hmmm, maybe the author of that article as well? I wonder if that occurred to him, seems like a glaring oversight not to at least recognize that possible irony.
评论 #38419070 未加载
chmod600over 1 year ago
A related effect that I&#x27;ve wondered about is: perhaps lower-skilled people compare themselves to the general public, while perhaps skilled people compare themselves to a smaller group of skilled peers.<p>In other words, if you asked me if I&#x27;m good at riding a bicycle, I&#x27;d compare myself to others in the general population and say &quot;yes&quot;. But if you ask a weekend bicyclist, they&#x27;d be better than me but perhaps compare themselves to weekend bicyclists, and rate themselves lower. And the effect might repeat for competitive bicyclists.<p>If true, this could explain why we intuitively believe the DK effect.
mewpmewp2over 1 year ago
My take on Dunning Kruger:<p>1. People really like the idea of smart people being humble and arrogance meaning stupidity, so they like to believe that DK is true, and they like to repeat this.<p>2. Some smart&#x2F;skilled people are humble, some are arrogant.<p>3. Some smart&#x2F;skilled people underestimate their skills, some overestimate.<p>4. Some stupid people are humble, some are arrogant.<p>5. Some stupid people underestimate their skills, some overestimate.<p>Overall, even if there is a correlation, you can&#x27;t tell by just arrogance of a person whether we are dealing with DK or whether it&#x27;s an effect at all. People&#x27;s personalities, skills and everything are a bit more complex than that.<p>Overall bringing DK up seems like some sort of social justice&#x2F;fairness effort rather than something that is actually true given any situation where someone is arrogant.
评论 #38416423 未加载
concordDanceover 1 year ago
The author fails to make his point quite badly. Of course if everyone&#x27;s self assessment was random the bottom quartile would overrate themselves! And that would be half of the Dunning-Kruger effect and we could truthfully say &quot;the bottom quartile of people overrate themselves&quot;!<p>The other part where those at the top have a better idea or where they rank noticeably does not come out in his toy example.<p>Honestly, he comes across as not having the slightest understanding of how people interpet those graphs...
hyperthesisover 1 year ago
<i>If</i> unskilled and skilled self-assessed themselves the same on average, then unskilled overestimate, and skilled underestimate.<p>That would be a significant result alone - that no one had any idea. (but as <a href="https:&#x2F;&#x2F;news.ycombinator.com&#x2F;item?id=38416100">https:&#x2F;&#x2F;news.ycombinator.com&#x2F;item?id=38416100</a> notes, there is a correlation).
badrabbitover 1 year ago
In my experience, people abuse flattery too much so it is hard to tell if their positive opinions of me are genuine and with merit. Generally speaking, I try to see the big picture and realize no matter how well I do, in a more global sense at best I am too 50th percentile, slightly above average. It is chance,relationships and supply&#x2F;demand economics that ultimately decide our ability to apply our talents effectively.<p>When it comes to others, I wish more people experienced the D&amp;R effect. It gets frustrating sometimes dealing with smart and talented people who think they are revolutionary rockstars. You know the kind, they see other people&#x27;s work and they are shocked how bad everything is, but never fear, they, our heroes are here to refactor everything until they leave and another hero looks at their work and rescues metropolis from it again. Patience and humility are a rare virtue for all of us.
civilizedover 1 year ago
We discussed this in a previous thread. The author is basically hypothesizing that perhaps people are so universally terrible at predicting their ability, their self-rating is like an unconditional random variable - just a random draw that is not influenced by their actual ability level at all.<p>If this is true, then when your actual ability is high, your self-rating is likely to be lower than your ability simply by random chance. For example, if ability ranges from 0-100, your actual ability is 99, and your self-rating is a uniform random number from 0-100, your self-rating is 99% likely to be lower than your actual ability. Conversely, if your actual ability is low, your self-rating is likely to exceed your actual ability level.<p>When it&#x27;s explained clearly and simply, the criticism raises a lot of questions. Are people <i>actually</i> that bad at rating their own ability? I doubt it.
ezekiel68over 1 year ago
&gt; However, there is a delightful irony to the circumstances of their blunder.<p>Indeed. And I find the tendency of people in this comment section to defend the flawed theory is further confirmation of another scientific finding: that we decide based on emotion and then justify our decision using rationality.
评论 #38418780 未加载
dimaskover 1 year ago
I would call this type of argument a case of regression to the mean rather than &quot;autocorrelation&quot;. That, of course, in principle requires independence between performance and assessment of performance. In many cases, it would make little sense to assume that the performance and assessment of performance are independent. But even then, one can simulate random data with some correlation, and still get a DK effect merely as statistical artifact. An overview of similar critiques, and a similar argument in <a href="https:&#x2F;&#x2F;www.frontiersin.org&#x2F;articles&#x2F;10.3389&#x2F;fpsyg.2022.840180&#x2F;full" rel="nofollow noreferrer">https:&#x2F;&#x2F;www.frontiersin.org&#x2F;articles&#x2F;10.3389&#x2F;fpsyg.2022.8401...</a> .
glitchcover 1 year ago
Geez, this is eye-opening. Thank you for sharing this.
joefourierover 1 year ago
So from my understanding, the Dunning-Kruger Effect paper doesn’t show the distribution of the perceived test scores nor the standard deviation, only an average, which rises with actual test score level.<p>If they showed the spread bar in each bin, you could form very different conclusions. Do low skilled people consistently estimate their score at around 60, or do they give effectively random results centred around 60?<p>Assuming the latter, it could mean that low skilled individuals are completely unable to evaluate their performance while higher skilled people are slightly better at it but still not very good, giving a slightly positive correlation which… is very distinct from what the DK effect implied.
PeterStuerover 1 year ago
You can take out the x from both sides, and the y would still not be a horizontal line.<p>In their eagerness to &#x27;deconstruct&#x27; the narrative, do the authors merely provide another example of Dunning-Kuger by overestimating their own cleverness?
TrackerFFover 1 year ago
The DK effect has gotten WAY more cred than it should. Today, it is just anoter feel-good piece that people use to justify their feeling that they&#x27;re (ironically) surrounded by loud idiots.
vismwasmover 1 year ago
The author measures the Dunning Kruger effect on his random data exactly because he assumes it when generating his random data.<p>By modelling skill and perceived skill as uniform draws between 0 and 100, the unskilled (e.g. skill=0) will over-estimate their skills (estimated skill = 50, the mean on the uniform random variable) and the skilled (e.g. skill=100) will underestimate it (as 50 as well, again the mean of the same random variable). The only ones who will be correct (on average) are the average skilled ones (skill=50).
lencastreover 1 year ago
Wasn’t this DK effect already debunked?
评论 #38416631 未加载
评论 #38416313 未加载
评论 #38416308 未加载
评论 #38416334 未加载
评论 #38416265 未加载
toasted-subsover 1 year ago
Idk I genuinely feel like after having to deal with 10+ doctors who all had different opinions. The last doctor finally made the same conclusion as me and he was the last person I had to see.<p>There&#x27;s always exceptions. And sometimes reading publications pertaining to a very specific thing should give you more say on a subject.<p>I just feel bad American tax payer money and the best years of my life was spent on telling medical professionals they don&#x27;t know what they are talking about.
im3w1lover 1 year ago
It&#x27;s fascinating how great Elo and similar ranking systems are at curbing DK. You just get a number, and that&#x27;s how good (bad) you are. It&#x27;s incredibly precise too, there&#x27;s just no arguing with it.<p>Also since the topic is D-K I&#x27;m a bit scared that I&#x27;m the fool here, but isn&#x27;t he misusing the term autocorrelation? What he describes sounds like just normal correlation?
dclowd9901over 1 year ago
I think what this article is missing is “the chart DK should have used.”<p>Instead we get a spurious explanation that doesn’t make a lot of sense based on completely fabricated data. It’s entirely natural for something that looks like DK to emerge from randomized data, especially when the Y axis is represented by some number of the mean (actually 50ish in this case).
zephrx1111over 1 year ago
A more generalizable explanation is regression towards the mean: everybody thinks they are an average person.
nitwit005over 1 year ago
If self evaluations are random, and you group a bunch of them together, then you&#x27;ll see values around the 50th percentile. That&#x27;s why their self evaluation line is nearly flat.<p>In the actual data though, the line clearly trends upward. The people who did well appear to be scoring themselves non-randomly.
mattbitover 1 year ago
This is not ‘autocorrelation’, it is regression to the mean. I find the article unclear and imprecise. For those interested in a better overview of the Dunning–Kruger effect, I recommend this short article by McIntosh &amp; Della Sala instead:<p><a href="https:&#x2F;&#x2F;www.bps.org.uk&#x2F;psychologist&#x2F;persistent-irony-dunning-kruger-effect" rel="nofollow noreferrer">https:&#x2F;&#x2F;www.bps.org.uk&#x2F;psychologist&#x2F;persistent-irony-dunning...</a>
评论 #38420269 未加载
austin-cheneyover 1 year ago
The best way to differentiate DK from autocorrection is motive. Low performance people will focus on motives that reinforce the perception of their competence, for example preferring code style over code delivery because while both may be arguably important one requires less effort and risk to attain.<p>There is research to qualify this out of Stanford. People will shift motives to attain complements and the types of compliments received will dictate the challenges they are willing to accept. When a compliment is specific to an action and measurable people will strive for continuously more challenging tasks to continually receive specific compliments. When compliments are generic and directed to the person they will tend to preference progressively less challenging tasks so that they continue to shine relative to the attempted effort. The differences in behavior produces a natural Dunning-Kruger effect wherein people seeking less qualified activities are more likely to over estimate their potential and degree of success.<p>This also statistically verified in research that correlates predictions to confidence. The more confidence a person is in their predictions, such as political talk radio hosts, the less accurate their predictions tend to be.
beltsazarover 1 year ago
I don&#x27;t know if I agree that it&#x27;s an autocorrelation, but one way to explain The Dunning-Krugger Effect is by acknowledging this simple fact:<p>Most people think that they are an average person, but they can&#x27;t be all average—there must be some people substantially below the median. Therefore, those people must overestimate their abilities.<p>This also applies to other aspects, such as attractiveness. Less attractive people would overestimate their attractiveness.
评论 #38416711 未加载
评论 #38419896 未加载
anonymouskimmerover 1 year ago
&gt; If the Dunning-Kruger effect were present, it would show up in Figure 11 as a downward trend in the data (similar to the trend in Figure 7). Such a trend would indicate that unskilled people overestimate their ability, and that this overestimate decreases with skill. Looking at Figure 11, there is no hint of a trend.<p>There certainly <i>is</i> a hint of a trend. Why do people, when visualizing data with a distinct trend, say that because the &quot;error bars&quot; from a particular statistical test overlap zero that no trend exists!?<p>Freshman <i>trend</i> to over-confidence. Grad students <i>trend</i> to under-confidence. Undergrads in general <i>trend</i> to over-confidence (though this trend decreases as year in school increases), and post-graduates, whether grad students or professors, trend to under-confidence.<p>These &quot;trends&quot; are not statistically significant, but they certainly are a trend!<p>Also, the random data distribution in figure 9 doesn&#x27;t show the same trends as Dunning-Kruger&#x27;s curve in figure 2. Perhaps there is at least one psycho-social mechanism here worth investigating?
评论 #38416882 未加载
评论 #38416678 未加载
Jenssonover 1 year ago
Psychologists using their pet theories to explain results and then people taking that explanation as the truth when they should really just look at the data is probably an as large problem as the replication crisis.
salty_biscuitsover 1 year ago
It&#x27;s just correlation, why do they keep calling it autocorrelation.
评论 #38418843 未加载
markhahnover 1 year ago
the numeric experiment does not produce a line identical to what DK report. if DK&#x27;s line where horizontal at 50%, it would indeed be nothing but autocorrelation.
chiefalchemistover 1 year ago
DK for me is simply: &quot;You don&#x27;t know what you don&#x27;t know.&quot; When that happens, it&#x27;s easy - surprise, surprise! - to misjudge your skill level. In a way, it almost feels cruel to ask someone with too few points of reference to say how much they know. The fact is whether high, low, or in the middle...they are guessing.<p>On the other hand, with enough experience the depth and breadth of your context improves, as it should. At that point, mis-self-assessment is the result of arrogance, bravado, etc. That&#x27;s a different problem than simply not knowing.<p>If nothing else, DK has a case of apple v oranges.
dilawarover 1 year ago
David Dunning response (2022): <a href="https:&#x2F;&#x2F;www.bps.org.uk&#x2F;psychologist&#x2F;dunning-kruger-effect-and-its-discontents" rel="nofollow noreferrer">https:&#x2F;&#x2F;www.bps.org.uk&#x2F;psychologist&#x2F;dunning-kruger-effect-an...</a>
rom1vover 1 year ago
If Y = X + estimation_error, then substracting X (in Y-X) removes the correlation rather than adding it.
lopatinover 1 year ago
Oh I read about the about the DK effect a while ago. I&#x27;m pretty much an expert in Psychology now, AMA.
poweraover 1 year ago
Nope.<p>I must object to this paragraph: &quot;To be honest, I’m not particularly convinced by the analytic arguments above. It’s only by using real data that I can understand the problem with the Dunning-Kruger effect. So let’s have a look at some real numbers.&quot;<p>He then goes on to use synthetic data.<p>Beyond that dishonest slight of hand, this is in the category of &quot;one thought experiment didn&#x27;t prove the phenomenon exists, therefore it must not exist&quot; logical errors.
a-dubover 1 year ago
i think of acf as a measure of repeating temporal structure and how &quot;strong&quot; and &quot;long&quot; it is, if it exists.<p>that is, it gives you a notion of if and what order of an ar model should fit any repeating structure in the data.
dahartover 1 year ago
Most people, even here on HN, do not know what the DK effect actually claimed to show. It does not show that confident people are more likely to be incompetent. Their primary result shows a positive correlation between confidence and supposed skill. (What skill, you ask?*)<p>This article suggests DK is even simpler than autocorrelation, that it’s just regression toward the mean. <a href="https:&#x2F;&#x2F;www.talyarkoni.org&#x2F;blog&#x2F;2010&#x2F;07&#x2F;07&#x2F;what-the-dunning-kruger-effect-is-and-isnt&#x2F;" rel="nofollow noreferrer">https:&#x2F;&#x2F;www.talyarkoni.org&#x2F;blog&#x2F;2010&#x2F;07&#x2F;07&#x2F;what-the-dunning-...</a><p>I don’t know which statistical artifact it is, but I am quite convinced that the so-called DK effect is not demonstrating something interesting about human psychology, I don’t buy that this is a real cognitive bias. I’ve read the paper several times, and the methodology seems to be lacking rigor. They tested a small handful of Cornell undergrads volunteering for extra credit, not a large sample, not the general population, and tested <i>nobody</i> who actually fits the description of ‘incompetent’ in a meaningful way. They primarily measured how people rank each other, not what their absolute skill was - and ranking each other requires speculating on the skills of others. There are obvious bias problems with asking a group of pampered Ivy League kids how well they think they rank.<p>* One of the four “skills” they measured was ability to get a joke - “appreciation of humor” - Huh? This is subjective! The jokes used aren’t given in the paper, either. Another was ‘grammar’ tests.
thewanderer1983over 1 year ago
The Dunning-Kruger effect isn&#x27;t as the article first quotes. It&#x27;s an effect that everyone experiences. We as humans tend to over simplify things we don&#x27;t understand well or at all. Therefore we over estimate our expertise on these subjects. We also tend to under estimate how much an expert on subjects we do know well. Everyone does this. It&#x27;s not just dumb people.
评论 #38416505 未加载
psychoslaveover 1 year ago
I went through the whole article, and I am not only very skeptical about the claimed debunk but wonder what kind of psychological trope you might label as corelative to such an article.<p>I mean &quot;bad science built only on rhetoric&quot; is a double edged sword, you know.<p>To start with, the graph presented at the end does not look like the one from the original article, where the self assessment does grow significantly, though it starts higher than average and grows less quickly than external assessment.<p>Also the article focus on &quot;random&quot; data set which, but we know that there are different classes of apparent noisy plots. Noisy distribution of self assessment would actually be an informative figure too.<p>So the biggest issue here is its kind of pretending that whatever the way the ordinate value is coupled to, if it includes the abscissa in its definition you&#x27;ll get the same kind of plot as a result, which is obviously false. You could easily come with arbitrary values coupled to &quot;x&quot; that would look radically different.
6510over 1 year ago
I was curious if the self assessment is done before or after the test.<p>Bing chat gave me this wild answer:<p>&gt; The effect is usually measured by comparing self-assessment with objective performance. For example, participants may take a quiz and estimate their performance afterward, which is then compared to their actual results 1. Therefore, people estimate their ability before the test by Dunning-Kruger.<p>In the case estimation <i>is</i> done before: If you&#x27;ve had training, like a soup of ingredients, that matches the priorities and biases of the test it would be strange if no measurable effect remained.<p>If it&#x27;s done after: You can create trick questions specifically designed to test if someone learned a specific thing. A good test would test for that. If someone didn&#x27;t learn the specific thing they could give&#x2F;guess the wrong answer with some confidence.<p>The design of the test has great influence on how poorly you&#x27;ll think you&#x27;ve done. I would argue that the superior test is the one designed to fool you. Hans Rosling famously created a multiple choice test with 4 answers per question with average results below 25%.<p>On a more fascinating note, unskilled means all areas of expertise outside your own.<p>People who are universally unskilled in all areas are of course more likely to think they are unskilled. In reality these people know little bits about many things.<p>This in contrast with people who spend all day, every day, for their entire lives pondering topics inside their area of expertise. If you are doing one thing you aren&#x27;t doing all of the other things.<p>Wikipedia had hilarious instances of experts contributing to countless articles accidentally ending up on the wrong page. Suddenly they have no patience, think they know everything and act like children. It&#x27;s funny because you cant just ban valuable contributors.<p>I would love to see this DK test done with professors furthest removed from the area of expertise.
resource0xover 1 year ago
Can someone explain the difference between Dunning-Kruger effect and &quot;illusory superiority&quot; effect (<a href="https:&#x2F;&#x2F;en.wikipedia.org&#x2F;wiki&#x2F;Illusory_superiority" rel="nofollow noreferrer">https:&#x2F;&#x2F;en.wikipedia.org&#x2F;wiki&#x2F;Illusory_superiority</a>)?
评论 #38418977 未加载
falserumover 1 year ago
Article feels like a personal attack towards D and K.
BrenBarnover 1 year ago
Yeah I don&#x27;t buy this either.<p>I do think the original Dunning-Kruger plot is a bit of an odd presentation. The way I look at it is just to say that people&#x27;s self-estimates of their ability fall into a relatively narrow range (e.g., 55-75th percentile on the graph), whereas their actual abilities of course cover the whole range from 0-100th percentile. You don&#x27;t really need the plot of &quot;x versus x&quot; (average score in each quartile). You just need to say &quot;people&#x27;s self-assessments seem to start unrealistically high and only go up a little, even as their ability goes up a lot&quot;.
eagerpaceover 1 year ago
Is this the opposite of imposter syndrome?
fnord77over 1 year ago
wikipedia&#x27;s article intro on this doesn&#x27;t state it is invalid :&#x2F;
epigramxover 1 year ago
&quot;Autocorrelation is the statistical equivalent of stating that 5 = 5.&quot; no sure if the author has some ..dunning-kruger there.
jongjongover 1 year ago
This makes sense. IMO, the reason why Dunning-Kruger effect is so popular among the upper classes (along with Impostor Syndrome) is that it helps to provide justification for social inequalities as it corrects inner monologues.<p>&quot;How come I have so much given that I&#x27;m not as skilled as these other people? I must suffer from impostor syndrome.&quot;<p>&quot;Look at all these people complaining instead of taking responsibility for their own failures, they probably suffer from Dunning-Kruger effect. Their work must not be good enough.&quot;<p>But of course this requires a certain detachment from reality (hence why many upper class people have blind spots). If they actually took a look at the evidence, they may find that some of these &#x27;Dunning-Kruger people&#x27; are actually far more skilled than they imagine. I think it explains why people like Jürgen Schmidhuber who made significant contributions to AI tend to be ignored. Then because people are ignoring them, they are compelled to promote themselves harder to try to get their fair share of attention but they are then put in the &#x27;Dunning-Kruger basket&#x27; until someone with a very good reputation like Elon Musk comes along and gives them credit. I think the same could be said about the mathematician Srinivasa Ramanujan; many mathematicians ignored his work or assumed he was a fraud because he seemed too sure of himself for someone who was completely unknown at the time. If such gross injustice can happen in a perfectly-quantifiable field like math, you can be sure it can happen in any field.
bszaover 1 year ago
Article claims Dunning-Kruger is present in a population where everyone estimates their own skills based on dice rolls. Someone who estimates their own skills based on a dice roll is objectively crap at estimating their own skills. Dunning-Kruger claims people are objectively crap at estimating their own skills.<p>Where is the contradiction?
riazrizviover 1 year ago
A general problem with Dunning Kruger is the assumption that if you score low on a test then you are bad at the subject it is evaluating. I’ve taken enough bad quizzes that purportedly evaluate skills that I am an expert in, to know that that is a leap.
eterevskyover 1 year ago
I think this article would&#x27;ve made more sense if it had a title &quot;The Dunning-Kruger effect is regression toward the mean&quot;, because that&#x27;s what the author is actually showing.
评论 #38420213 未加载
lifeisstillgoodover 1 year ago
The Dunning Kruger effect is simply the same reason expensive projects are undertaken and never hit budget - not because we cannot estimate costs but because if we did we would never do anything.
notShabuover 1 year ago
every domain of expertise has two &quot;elo&quot; systems, the niche one and the broader one.<p>e.g. you can learn basic juggling in 30 minutes that you are top 10% of your friends&#x2F;colleagues etc...<p>however within the juggling community itself this is known as the &quot;3 ball cascade&quot; a really simple trick relative to the ones that requires years to master. an outsider may not be able to tell the difference between the 1 year expert and the 10 year master.<p>a lot dunning-kruger can be explained by people in one or the other not understanding the other system
gsuuonover 1 year ago
I think the take-away from Dunning-Kruger is that most people think they&#x27;re closer to average than they are, on both sides.
dudeinjapanover 1 year ago
So you&#x27;re saying that the Dunning-Kruger effect applies to Dunning &amp; Kruger.
pie_flavorover 1 year ago
This take is a perfect example of Dunning-Kruger itself, ironically. <a href="https:&#x2F;&#x2F;andersource.dev&#x2F;2022&#x2F;04&#x2F;19&#x2F;dk-autocorrelation.html" rel="nofollow noreferrer">https:&#x2F;&#x2F;andersource.dev&#x2F;2022&#x2F;04&#x2F;19&#x2F;dk-autocorrelation.html</a>
评论 #38417465 未加载
RevEngover 1 year ago
What Blair Fix&#x27;s article gets wrong is that there are two stark differences between what Fix generated with random data and what Dunning and Kruger observed in theirs.<p>Fix has each person guess randomly between 0 and 99 where they will lie in the percentiles. They simulate every person having no idea and giving equal probability to being the best or the worst. If we then sort them by how well they really did into quartiles and then evaluate the average of how well they thought they would do, we get what we would expect: each quartile has an equal chance of predicting that they will do well or do poorly, with an average expected percentile of 50, which is what you would expect by a random guess.<p>Note two key things about this: - All quartiles guessed the same - there was no correlation between what they guessed and how well they actually did - All quartiles guessed the expected average percentile - 50%. This means they were unbiased in how well they thought they would do.<p>If people were unbiased but also unaware, this is the null hypothesis we would expect: on average people predict themselves to be average and there&#x27;s no correlation between how well they predicted they would do and how well they actually did.<p>Now compare that to what Dunning and Kruger observed: - The quartiles did NOT guess the same. There was a bit of an upwards trend, which suggests that people at least somewhat were able to determine their actual percentiles, even if only weakly on average. - The predictions were biased. All groups estimated they would do better than the expected average. That is to say, on average, they thought they were above average. This is an important bias. - The differentials between quartiles are not equal. The first and second quartile typically predicted the same, over-estimated value, implying that neither group had any idea they were better or worse than each other. However, the upper quartile consistently estimates a higher average. That is to say, people who perform well, on average, believe they are performing even better than those who don&#x27;t perform well. And perhaps most surprisingly, there was often a statistically significant dip at the third quantile. Comparing their beliefs, people who did well believed they had done worse than the people who actually did worse.<p>Fix also fails to go beyond the first figure of the paper. After seeing this inconsistent behaviour between the quartiles, Dunning and Kruger then test what happens if the respondents are given an opportunity to grade each other - therefore getting an idea of what the percentiles actually look like - and to have their skills improved - thereby possibly making them better able to judge their own and each other&#x27;s abilities. Again, if Fix&#x27;s premise that this is all just a result of manipulating the autocorrelation of an otherwise unbiased random sequence, then these interventions should have no discernable effect. Yet, Dunning and Kruger find markedly significant changes after these interventions, and those changes are different within the different quantiles.<p>It is precisely this difference between quantiles which is the Dunning-Kruger effect. Fix effectively makes their point for them by building a null model and showing what would happen if there were no Dunning-Kruger effect - if people were fully unaware and unbiased. Instead, it is the way in which Dunning and Kruger&#x27;s observations deviate from this model that is the very effect that bears their name.<p>Instead, all that Fix manages to do is point out how confusing the plot is that Dunning and Kruger produced. The plot can easily be misinterpreted to suggest that it&#x27;s the difference between y and y-x that is important. Instead, in their writing, Dunning and Kruger actually focus on the differences in how y-x changes when the situation changes, demonstrating that it&#x27;s actually dependent on knowledge and how different people respond to that knowledge. What they actually show is that delta(y-x) vs x has a nonzero relationship and this is particularly interesting.<p>Perhaps if Dunning and Kruger had not included the example of perfect knowledge as a comparison, but instead included the example of unbiased and unknowledgeable that Fix produced as the thing to compare against, the Dunning-Kruger effect would be much better understood.<p>Further, both could benefit greatly from plotting and tabulating not just an average, but the overall distribution within each group. Fix should know that variance is just as important as bias. Even if all groups are biased in their prediction, differences in variance between each group indicates their confidence in their belief. Knowledge should help to reduce both bias and variance. A guess with high variance tells us little, while a guess with low variance tells us quite a bit. Even if all quartiles predicted the same average, we wouldn&#x27;t fault those with little ability for guessing a high number if they did so with low confidence. On the contrary, we would expect people with high ability to be more confident (and correct) in the assessment of their ability.
评论 #38420713 未加载
hyperthesisover 1 year ago
It&#x27;s Dunning-Krugers all the way down - including this self-referential smugness.
greenthrowover 1 year ago
Lmao this article is an example of Dunning-Kruger at work. The author thinks they have found and are revealing something but they are just failing to fully understand the subject. Amazing.
评论 #38416354 未加载