Plant Care Has a Dunning-Kruger Problem

Right now, somewhere in a plant group like Plant Hoarders Anonymous (PHA), someone who has kept a single fiddle leaf fig alive for eight months is confidently explaining root rot to a stranger. They are specific. They are sure. They are also, more often than the comment count suggests, wrong. And here is the part that should bother you: they will almost certainly never realize it.

In plant care, confidence and competence come apart, because the thing that normally keeps them welded together, fast and honest feedback, is missing.

This is not a post about stupid people, and it is not a post about you doing anything wrong. It is an article about a hobby that is almost perfectly engineered to make sincere people sure of things that are not true, and then to protect that certainty from ever being tested. You have been on the receiving end of this advice. You have more than probably also given some of it. I know I have. Neither makes you a bad plant parent or gardener. It makes you a human being working in a domain that punishes humans in slow motion.

The psychological name for the front half of this problem is the Dunning-Kruger effect. The phrase gets thrown around online as a polite way to call someone an idiot, which is both unfair and wrong. The real finding is stranger, narrower, and far more useful than the insult version. So let us start there, because if we are going to talk about overconfidence, we had better be careful not to be overconfident about the science.

Let's Get You Up to Speed

This UG article will help you understand:

• What the Dunning-Kruger effect actually found in 1999, and how the popular version distorts it
• Why slow, ambiguous plant feedback decouples how sure you feel from how right you are
• How confirmation bias and survivorship bias protect bad advice from ever being corrected
• Why the famous "Mount Stupid" chart was never in the original research
• How to calibrate your confidence to the quality of your evidence, not the strength of your hunch
• The single habit that closes the feedback loop faster than anything else

Got Things to Do? This is For You!

The Dunning-Kruger effect comes from a 1999 study by Justin Kruger and David Dunning, who found that people in the bottom quartile of tests on grammar, logic, and humour rated their performance at the 62nd percentile when they actually sat at the 12th. The cause is metacognitive: the skills you need to do something well are the same skills you need to judge whether you did it well, so the least skilled are the least equipped to notice their own gaps. Plant care turns this quirk into a lifestyle, because the feedback is slow and muddy. A plant takes weeks to decline, and by the time it does, the real cause hides behind a dozen plausible alternatives. Confirmation bias and survivorship bias then guard the original belief, so the confident adviser is never corrected. The popular Mount Stupid curve, worth noting, never appeared in the original paper, and a serious statistical critique argues part of the effect is a measurement artefact. The honest version is narrower than the meme, and it still applies to you. The fix is not humility as a mood. It is measurement: close the loop with a number, and start with the light your plant is actually receiving.

What does the Dunning-Kruger effect actually say?

The Dunning-Kruger effect says that people who are bad at something tend to overestimate how good they are, because the same knowledge that would make them competent is the knowledge they would need to see that they are not. It is not a general claim that unintelligent people are arrogant. It is a specific claim about a trap that closes around anyone working at the edge of their own knowledge. In my past marketing career I'd often use the phrase, "They don't know what they don't know." and I would regularly council clients, somewhat facetiously, that they should never underestimating the stupidity of their clients. I also remind them that they are my client and it applies to them too. ~wink ~wink

Justin Kruger and David Dunning ran four studies at Cornell in 1999, testing undergraduates on logical reasoning, English grammar, and the ability to tell what is funny. After each test, participants guessed how they had done relative to everyone else. The people who actually scored worst were the most wildly off. Those in the bottom quartile, sitting near the 12th percentile , on average believed they had landed around the 62nd. The top performers showed the opposite, smaller error: they were genuinely skilled but slightly underrated their standing, partly because they assumed the task was as easy for everyone else as it was for them.

Stat: In Kruger and Dunning's 1999 study, bottom-quartile participants scored at roughly the 12th percentile but rated themselves at the 62nd, a gap of about 50 percentile points.

The authors called the underlying problem a "dual burden." If you lack skill in a domain, you not only get things wrong, you also lack the very skill required to notice that you got them wrong. They described it as a rough analogue to anosognosia, the neurological condition in which a brain injury removes a person's ability to perceive the injury itself. The incompetence and the blindness to it come from the same missing piece.

This is why the effect is domain-specific, and why the insult version misses the point. You know perfectly well you would lose a one-on-one game against LeBron James, because judging basketball is easy even when playing it is not. The trap only springs in domains where judging quality requires the same expertise as producing it. Knowing whether your watering routine is correct requires understanding root-zone oxygen, transpiration, and how light drives a plant's water demand in the first place. Those are exactly the kinds of things a beginner does not have an understanding of yet. So a beginner will not just get the answer wrong. They will feel confident in the wrong answer, and believe in its sincerity.

Here is the analogy that the study hands us directly. Judging whether a sentence is grammatical takes the same grammar knowledge as writing one correctly. Judging whether your plant is getting enough light takes the same understanding of light you would need to set it up right to begin with. Lacking it does not feel like a gap. It feels like an opinion.

Nerd Corner: Each of the original charts plotted only four points, one per performance quartile, with perceived ability on one axis and actual ability on the other. The headline pattern is an asymmetry: low performers overestimated by a large margin, high performers underestimated by a small one. There was no smooth confidence curve and no single "peak" of maximum overconfidence.

One detail from the study is quietly hopeful, and I will come back to it. When the researchers actually taught the low performers the skill, the gap between perceived and real ability shrank. Competence and the ability to see your own incompetence arrive together. Learning is the exit.

Why is the famous Dunning-Kruger chart a fabrication?

The famous Dunning-Kruger chart, the one with a towering "Peak of Mount Stupid" that crashes into a "Valley of Despair" before climbing a "Slope of Enlightenment," is a fabrication. It is a later invention, drawn by other people, and it never appeared in Kruger and Dunning's paper. The real graphs were four flat quartile points showing overestimation among the worst performers. There is no mountain.

Myth Check: The viral "Mount Stupid" curve is not data. It is an illustration someone made up after the fact, and it overstates the original finding in exactly the confident, shareable way the effect itself describes.

It gets more interesting, and this is where an evidence-based publication has to hold itself to its own standard. The effect is also genuinely contested. In 2020, Gilles Gignac and Marcin Zajenkowski argued that much of the classic pattern is a statistical artefact, driven by the better-than-average effect and regression to the mean rather than by a special blindness in the unskilled. Related work has shown that a Dunning-Kruger-shaped graph can appear even when you feed the chart pure random noise. The pattern, in other words, is partly baked into how the numbers are plotted.

That critique is itself disputed. Later replication work, including a 2023 reanalysis, still found a real effect, just a smaller and less dramatic one than the meme implies. So the calibrated conclusion looks like this: the cartoon version is wrong, the size and mechanism are still argued over by people who do this for a living, and the modest core observation survives. The less you know about a domain, the harder it is to perceive what you are missing. That much holds up.

An article about confident people being wrong has an obligation not to be confidently wrong itself. The point of saying all this is not to undercut the idea. It is to model the behaviour the rest of the article is asking for: state what the evidence supports, flag what it does not, and keep your certainty proportional to your proof. That habit is the whole cure, and most plant advice never attempts it.

Why does plant care break the link between confidence and competence?

Plant care breaks the link between confidence and competence because it removes the one thing that normally repairs overconfidence: quick, unambiguous feedback. In most skills, reality pushes back fast. You toss a brick of a free throw and the ball misses the hoop. You misjudge the space your car needs to parallel park and you're rewarded with a fender bender or rim rash on your wheel. Overconfidence gets a reality check with immediate, repeated contradiction, whether you enjoy it or not.

Plant care offers the opposite arrangement. The consequences of what you did today typically arrive weeks or months from now. A plant that was overwatered this morning looks perfectly fine for a week or more, then begins to decline in a way that looks a lot like thirst, so the paranoid plant parent waters it more and finishes the job. The signal comes late, comes weak, and comes wearing a disguise.

It is worse than slow. The outcome is overdetermined. A plant sits at the intersection of nine interacting parameters, so any result you observe has a dozen candidate explanations standing behind it. You can credit a recovery, or blame a death, on almost anything you already believed, and nothing in the room will contradict you. This is also why so many plant problems get misdiagnosed in the first place, a pattern I've mapped in detail in the seven houseplant failure patterns.

And then there is the cruelest part, the one that sits right on top of everything. The single most consequential variable in the whole system is light, and light is the variable human eyes are worst at judging. Your eyes adjust to brightness on a roughly logarithmic scale, so a room that reads as "bright enough" to you can deliver a small fraction of what a plant needs to grow. The feedback you most need to learn from is precisely the feedback you are least equipped to perceive without a tool to measure appropriate light intensity. This is why your own sense of how bright a spot is quietly lies to you, and how even with a light meter, things still need a competent understanding to get things right.

FYI: A plant in too little light does not always look like it is failing. It can hold steady for months or even years by drawing down reserves and lowering its own ceiling for future growth. It looks stable right up until it is not, which is the opposite of useful feedback.

Imagine learning to drive if your car only responded to the steering wheel three days after you turned it, and every crash was officially blamed on the weather. You would develop strong, confident, and completely wrong theories about how to drive. That is houseplant care. The structure of the activity all but guarantees that sincere people will arrive at sincere, durable, mistaken beliefs. I've heard estimates from horticulturists that 80-85% of online plant care advice is wrong, misleading, or missing valuable context to be actually beneficial.

Why does the confident adviser never get corrected?

The confident plant care adviser, increasingly occupied by popular influencers never gets corrected because two biases quietly delete the evidence that would prove them wrong, long before it reaches them.

The first is survivorship bias. Every thriving plant on someone's shelf is visible, photographed, and credited to whatever that person happens to do. The plants that died under the same routine were quietly composted and never shown. So the advice is assembled entirely from survivors and is structurally blind to its own casualties. "It works, look at my plants" is a sentence built from the winners only, by someone who genuinely cannot see the losers, because the losers are in the bin.

The second is confirmation bias. We remember the outcomes that fit our theory and explain away the ones that do not. The plant that survives and perhaps thrives proves the method. The two that died "were just weak plants," or "had a pest," or "never acclimated" The theory itself is never actually on trial, because every verdict that comes back guilty gets quietly appealed. This is one of a long list of mental shortcuts that sabotage plant care, and I've published the full roster separately.

Stat: Most people with a collection of relatively healthy plants over a fairly short time frame (most of today's popular influencers seem to only have started during Covid) position themselves as an above-average grower, which is arithmetically impossible for most people to be. That "better than average" reflex is the same one that statistical critics point to as a large part of the Dunning-Kruger pattern.

Then the social layer takes over and amplifies all of it. Confident delivery reads as competence, so the surest voice on a private channel or in a public thread tends to win regardless of whether it is right. Online communities quite often reward certainty over accuracy, and uninformed opinions travel faster than careful ones, because careful ones are full of caveats that make for worse comments. I've gone deep on this elsewhere, in why so much online plant advice is wrong and in the illusion of plant parent choices. The short version is that the people who most need to be corrected are the least likely to be, and the medium they live in is built to reward them anyway.

How do you move from confidence to competence in plant care?

You move from confidence to competence by replacing belief with measurement wherever you can, and by treating every claim, including your own, as something that could be wrong. The trap is structural, so the way out is a set of habits rather than a personality trait. None of them require you to be smarter. All of them require you to close the loop. I regularly ask myself in various aspects of my life were I am paid to offer my professional advice, "What if I'm wrong?", "Can I prove this?" — if I can't then I should not be relying on it as a foundational element that others should 100% rely upon.

For us as individuals, we need to start by closing that loop with a number. The fastest cure for plant-care overconfidence is an instrument, because an instrument does not care how sure you feel. A light meter turns the invisible master variable into a value you can actually check against what a plant needs, which is the highest-leverage move available to you, since light governs almost everything downstream of it. You stop arguing about whether a spot is "bright" and start reading what it delivers. Another thing I use across my strategic marketing consultancy and something I regularly say to clients, "If you can measure it, you can improve it!"

Pro Tip: Before you trust any advice, including mine, or a routine you have used for years, measure the light the plant is actually receiving. It is the one parameter most people are confidently wrong about, and the one most likely to be the real cause hiding behind whatever you were about to blame.

Next, ask the how-would-I-know question. Before you trust a piece of advice, ask one thing: if this were wrong, what would I actually see, and could I even see it? A lot of the time the honest answer is that nothing would look different for weeks. That means you can't tell yet whether the advice is working, so don't act sure of it. The slower and foggier the feedback, the less sure you should be about what caused what.

Then downgrade "it works for me" from a verdict to a hypothesis. One plant, in one home, with no control group, is where an investigation starts, not where it ends. It is real information, and it is nowhere near proof. The moment you treat your own experience as the end of the inquiry, you have rejoined the bottom quartile, just with more plants.

Finally, rank your evidence and let your tone follow it. Controlled studies and commercial production data outrank a confident comment, which outranks a gut feeling. Calibrate how sure you sound to where your evidence actually sits on that ladder. And keep the hopeful finding from the original Dunning-Kruger study in view: when the unskilled participants were taught the skill, the overconfidence gap closed. Learning is the only exit, and it is open to everyone, including the person from the first paragraph with one pothos and eight months of practice.

The goal is not to be less confident. It is to earn your confidence with evidence, and then to spend it carefully.

The Dunning-Kruger effect is not a story about other people. It names a trap that the structure of plant care lays for every one of us who tries to learn the hobby of plant parenting and gardening in general, by hiding the feedback that would otherwise keep us honest. The people handing out confident, wrong advice are not fools. They are engaging in a social arena that rewards certainty and conceals error, and they were probably given the same bad information you were. You step out of the trap the moment you stop asking how sure you feel and start asking how you would know. Measure the light. Watch what actually happens. Hold your conclusions loosely enough to update them. That is the whole game.

Frequent Questions

Sources and Further Reading

Kruger, J., and Dunning, D. (1999). Unskilled and unaware of it: How difficulties in recognizing one's own incompetence lead to inflated self-assessments. Journal of Personality and Social Psychology, 77(6), 1121 to 1134. https://doi.org/10.1037/0022-3514.77.6.1121

Gignac, G. E., and Zajenkowski, M. (2020). The Dunning-Kruger effect is (mostly) a statistical artefact: Valid approaches to testing the hypothesis with individual differences data. Intelligence, 80, 101449. https://doi.org/10.1016/j.intell.2020.101449

Dunkel, C. S., Nedelec, J., and van der Linden, D. (2023). Reevaluating the Dunning-Kruger effect: A response to and replication of Gignac and Zajenkowski (2020). Intelligence.

Related UG reading: 25 Cognitive Biases That Sabotage Plant Care, Why So Much Online Plant Advice Is Wrong, The Illusion of Plant Parent Choices, Part 1, Why Your Sense of Light Is Lying to You, and The Nine Cardinal Parameters.