Psychology
Spoil your kids, breed a narcissist? Read the study before you believe the headline.
Here is a headline engineered to write itself: spoiling kids breeds narcissists, while praising them grows healthy adults. It slid past your defenses, didn't it — because it lands exactly where you already lean. That smooth fit is not a reason to believe it. It is the precise moment to get suspicious. A new study really did find this; the question is what "find" is allowed to mean.
What the researchers actually measured
The paper, "Praise the light, indulge the dark," ran in Current Psychology in 2026. The team asked 720 university students to look backward: to recall how their parents had treated them as children across three strategies — praise, indulgence, and status-granting — using a scale called PIS. Then they measured the students now, on a model of the "Dark Triad": narcissism, Machiavellianism, and psychopathy, each split into a more malevolent face and a more adaptive one. The reported pattern is tidy. Recalled over-indulgence went with darker, antagonistic traits — narcissistic rivalry, the cold disinhibition of psychopathy. Recalled praise went the other way, tracking with more adaptive adult traits, and early parental praise specifically correlated negatively with antagonism, which the authors read as a protective signal. The result also has a genuinely counterintuitive wrinkle worth keeping: heavily indulged people scored lower on "Machiavellian agency," the healthy, confident ambition you might have guessed coddling would inflate. If anything, the indulgence tracked with weaker self-assurance, not stronger. So this is not a junk study. It is a competent one, doing the ordinary, careful work of measuring traits and reporting how they line up — and that is exactly why the headline is dangerous. Bad studies are easy to dismiss. It is the credible ones, stretched one word too far, that mislead.
Filter one: this is a correlation, and nobody assigned the parenting
Keith Stanovich spent a career teaching one habit that survives almost any psychology headline: separate "predicts" from "causes." No researcher randomly assigned 720 children to be indulged or praised — that experiment would be monstrous, so it was never run. What you have instead is a snapshot of how two things vary together. The headline quietly upgrades that into a manufacturing process: spoiling makes narcissists. But run the third-variable problem. A child's own temperament — partly heritable, present from the start — can shape both how parents treat that child and who the child becomes. A difficult, grandiose kid may pull indulgence out of exhausted parents and grow into a difficult, grandiose adult, with no arrow running from the parenting to the personality at all. Correlation cannot tell those stories apart. The study sees that two things travel together; it cannot see what is driving the car.
Filter two: the data is a memory, and memory is not a recording
Now the part the headline never mentions. Nobody observed these childhoods. The "parenting" in this study is what a twenty-year-old remembers and reports about a childhood that is gone. Retrospective self-report is one of the softer floors a finding can stand on, and here it sits under a personality measure in a way that should make you wince. Picture an adult who scores high on narcissistic antagonism. Ask him to recall how his parents treated him. Is his memory a neutral archive — or is it tinted by the very disposition you are also trying to measure in him? An entitled, antagonistic person may genuinely recall a more indulgent childhood because his present self colors the past, not because the past produced the present. Memory is not a recording you replay; it is reconstructed each time, and the mood and self-image you bring to the act of remembering bend what comes back. The supposed cause and the measured effect are pulled from the same biased witness, on the same afternoon, in the same questionnaire. That does not erase the finding. It means the finding leans harder on a wobbly instrument than the headline lets you see — and that the very direction of the arrow, parent-shaped-child versus child-colored-memory, is left genuinely open.
Filter three: one sample, one slice of humanity
Then the unglamorous filter, the one that ends most overheated claims: who, exactly, was studied? Seven hundred and twenty university students. Useful, but a single sample, and a WEIRD one — Western, Educated, and drawn from the narrow, self-selected world of the university, hardly a cross-section of human parenting across cultures and classes. Stanovich's antidote to single-study fever is converging evidence: a claim earns its confidence when it holds up across different samples, methods, and labs, ideally with longitudinal designs that follow real children forward through time instead of asking adults to glance back. One recalled, correlational survey of students is one data point. It is allowed to be interesting. It is not allowed to be a verdict.
"Predicts" is a finding. "Causes" is a story the headline tells on top of it.
A competent study can support the first while saying nothing certain about the second.
What this means for you
Don't rewrite how you raise your kids over this, and don't turn around and indict your own parents with it. Both moves treat a retrospective correlation as a causal law, which is the one thing it firmly is not. The useful takeaway is smaller and sturdier than the headline, and it travels well beyond parenting. When a study lands that flatters what you already suspect — about your family, your habits, your side of any argument — that comfortable click is your cue to slow down, not speed up. Ask the three questions Stanovich would: is this correlation or cause, where did the data actually come from, and is it one study or many converging? The parenting research is worth knowing; the antagonism it describes is real enough to study. What is not yet earned is the leap from "predicts, in this recalled sample" to "causes, in your life." Hold the finding loosely, and hold the headline at arm's length.
Framework from Keith Stanovich's How to Think Straight About Psychology (斯坦诺维奇《对伪心理学说不》) — correlation is not causation, the third-variable problem, the limits of retrospective self-report, WEIRD-sample caution, and the case for converging evidence over single studies. News peg: "Praise the light, indulge the dark: Parenting strategies and dark personality traits," Current Psychology (2026), 720 university students, recalled childhood parenting (PIS scale) against a Dark Triad model; covered by PsyPost and Simply Psychology. Key caveats: this is correlational and retrospective self-report, not an experiment; "predicts" is not "causes"; a single WEIRD-student sample. Popular-science interpretation, not a clinical or diagnostic judgment.