Published on 2026-07-25
Is 'do your own research' actually good advice?
The advice sounds unbeatable: don't just believe, go check. But how you search quietly decides what you'll find, often before you've typed a word.
Nobody can really argue with 'do your own research'. It sounds like the opposite of gullibility, the responsible thing, the grown-up move. Don't take anyone's word for it, go look for yourself. The trouble isn't the instruction. The trouble is what happens the moment you sit down and actually start typing.
Your search box already knows what you think
In 2025, researchers Eugina Leung and Oleg Urminsky published a paper in the Proceedings of the National Academy of Sciences describing something they call the narrow search effect. Across 21 studies, some run on Google, some on Bing, some on AI chat tools, they found the same pattern again and again: people's existing beliefs shape the words they choose to search with, and those words then shape everything that comes back. Type 'is nuclear energy dangerous' and you'll get a very different reading list than if you'd typed 'is nuclear energy safe'. Nobody manipulated the algorithm. You handed it your conclusion before you asked your question.
That's the quiet part of 'doing your own research' that the phrase never mentions. You don't walk into a neutral library and stumble onto the truth. You walk in already leaning one way, you choose search terms that reflect that lean without noticing you've done it, and the results politely hand your lean back to you, dressed up as evidence.
The first page of results makes it worse, because most of us barely go past it. If your search terms already skew toward one answer, the top results skew the same way, and the sheer fact of finishing on page one, with several sources appearing to agree, feels like a job well done. It rarely is. It's a room you decorated and then admired for how convincing it looked.
Even careful searchers get stuck
You might assume this only traps people who search lazily, clicking the first link and calling it a day. A 2021 study in the journal Frontiers in Psychology, led by Masaki Suzuki and Yusuke Yamamoto, tested that assumption directly. They gave 275 participants a biased starting belief about a food safety topic, then watched how they searched. The searchers who were more careful, who spent longer reading and even clicked on pages that disagreed with them, still ended up about as convinced of their starting belief as before. Reading the other side isn't the same as being moved by it, and simply looking harder doesn't automatically fix that.
What 'research' would actually mean
None of this means looking things up is pointless, it means the word 'research' is doing more work than the casual version of the advice admits. Watching five videos that all agree with each other isn't research, it's five doors into the same room. Real research means deliberately searching for the version that would prove you wrong, not just the version that explains why you're right, and it means treating independent sources as the ones that didn't copy each other's homework, not the ones that happen to agree with you.
Typing your question into a search bar isn't research. Typing the question your beliefs would rather you not ask, that's closer to it.
- Write down what you already suspect before you search, so you notice if your search terms are quietly built to confirm it.
- Search the exact opposite phrasing of your first instinct, and see if the results actually differ.
- Look for who disagrees with the claim and why, not just who agrees loudest.
- Prefer sources that show their own sourcing over ones that only assert.
There's also a plainer, less flattering way to put all of this: research takes longer than most of us want it to. A real check means reading something you expected to dismiss, sitting with the discomfort of maybe being wrong, and following a claim back to where it actually started instead of stopping at the fifth account that repeated it. None of that happens by accident, and none of it happens in the time it takes to skim three headlines and feel settled.
So keep doing your own research. Just notice, honestly, which question you actually typed, and whether you'd have been just as happy with an answer that went the other way.