I Researched How To Do Research, Here's What I Learned

ELI5/TLDR

A guy who has made videos for 19 years went out and researched how to research. He came back with five steps: know what question you’re actually asking, check your biases before you start, cast a wide net and then follow counterarguments, vet every source like a paranoid detective, and be honest about what you don’t know at the end. The running theme is that certainty is a red flag and “do your own research” usually means “I went shopping for the conclusion I wanted.”

The Full Story

Pick a question before you pick a side

The first move is deciding what you’re actually trying to find out. This sounds insultingly obvious, which is why most people skip it. They start with a topic and drift. Topics are infinite. Questions have edges.

His example is nuclear power. You could spend a year on it and never come up for air — history, waste, Chernobyl, weapons, Oppenheimer’s lunch order. Too wide. The usable question is narrower: are we underusing nuclear out of fear? Now there’s a shape to argue with. Now you know which rabbit holes to ignore.

He also notes the question mutates as you go. You learn something, realize the original question was slightly wrong, refine it, keep moving. Research isn’t a straight line from confusion to answer. It’s a spiral.

You are already biased, and so is everything you read

Before touching any source, acknowledge the shape of your own head. Your feed, your friends, your neighborhood, your childhood — all of it pre-loads your opinions. Anything that arrives feeling like clean truth probably isn’t.

“If I come across an article that lambasts a group that I disagree with and really strongly aligns with what I agree with, that’s actually a red flag for me. It feels like I’m not getting all the information. I’m drinking in the sweet sweet tribal milkshake and it tastes like truth.”

The warning that follows is worth sitting with. You can cherry-pick sources to defend almost any belief. The internet is big enough that wanting to be right is sufficient to feel right. That’s not research — that’s shopping.

Experts exist for a reason

Related trap: refusing to trust anyone who has studied the thing longer than you. Someone who has spent thirty years on squirrels probably knows more about squirrels than you do, even if some percentage of squirrel experts are wrong or captured or dull. The base rate still favors them.

This is why the phrase “do your own research” sets off alarms for him. Done honestly, it means go learn. Done in the wild, it usually means “I went looking for the alternative to mainstream consensus because the consensus offended me.” The scientific method already assumes consensus is provisional — it’s built to update. Going in determined to overturn it is just bias with extra steps.

The spiderweb

Actual search technique, once you’re ready. Start broad — Google, Wikipedia, Reddit, YouTube. Get the rough terrain. Then refine. Then chase the sources the first layer cited. Then — this is the part most people skip — go search the counterarguments. Read the people who disagree. He calls this the spiderweb method.

A few practical tricks worth stealing:

Boolean operators. peas AND carrots returns both. peas OR carrots returns either. peas NOT carrots strips carrots out. Caps matter.
Quotation marks force an exact-phrase match.
Add PDF to your search and you’ll surface full-text documents — often studies, reports, academic papers.
Append -AI to a Google search to kill the AI overview and get the old featured snippet with an actual source attached.
PubMed and Google Scholar for anything academic.

SIFT: a four-step source check

This is the meat of the video. When you land on a source — article, study, tweet, screenshot — run it through SIFT, a framework from digital literacy researcher Mike Caulfield:

S — Stop. Don’t share yet. Don’t react yet. Read the whole thing, not just the headline.
I — Investigate the source. Open a new tab. Who wrote this? What else do they write? Who funds the publication? What does Wikipedia say about them?
F — Find better coverage. See who else is reporting the same claim. Do they agree? Who disagrees, and why?
T — Trace claims to the original. Follow the links. He points out that links often go to studies that don’t support the claim being made, or support the opposite. The writer is betting you won’t click.

To this he adds a crash course in study literacy: sample size, control group, blinding, meta-analysis, quantitative vs. qualitative. One small flawed study gets laundered through a thousand headlines. Knowing the shape of the evidence keeps you from getting fooled by the summary.

Certainty is suspicious

The quietest and best point in the video. After years of interviewing experts, he noticed a pattern: real experts hedge. They’re more curious than conclusive. They ask more than they answer. They talk about “what we currently think” rather than “the truth.”

“That’s not to say that nothing is certain and nothing can be known for sure, but your sus meter should raise a little bit when you see someone so certain of themselves.”

If a claim feels airtight and forceful and perfectly aligned with the writer’s prior positions, that’s a signal to slow down, not speed up.

End honestly, not neatly

The last step is the one writers cheat on most. When you finish, there’s a pull toward tying it up — a moral, a clean takeaway, a “here’s what we learned.” Stories want bows. Truth rarely has them.

His advice: say what you actually think, say where you might be wrong, and name the factors that could undermine your conclusion. Not only is it more honest, he argues it’s more engaging. Audiences can smell someone hiding the uncertain parts. When he gets stuck mid-video, it’s almost always because he’s avoiding something true that feels embarrassing or boring. Saying it anyway tends to unstick him.

Key Takeaways

Start with a question, not a topic. Topics sprawl. Questions have walls you can bump into.
Revisit the question mid-research. You don’t know what you don’t know until you’ve read a few things.
Agreement is a red flag. If a source perfectly matches your priors, suspect it harder than a source that doesn’t.
“Do your own research” usually means “I went shopping for a conclusion.” Real research includes the counterarguments.
Spiderweb method: broad overview → deeper sources → counterarguments → repeat.
Search operators: AND, OR, NOT in caps; "exact phrase" in quotes; add PDF for full papers; append -AI to kill Google’s AI overview.
SIFT: Stop, Investigate the source, Find better coverage, Trace claims to the original.
Click the links in articles. Writers often link to studies that don’t say what they claim. Most readers never check.
Study literacy matters. Sample size, control group, blinding, meta-analysis vs. single study, quantitative vs. qualitative — knowing these filters out a lot of noise.
Certainty is a sus signal. Real experts hedge. Forceful conclusions deserve extra scrutiny.
End with honest uncertainty. Name your limits, your possible errors, the factors that could flip your conclusion. Neat endings are usually lies.

Claude’s Take

This is a good video disguised as a goofy one. The jokes are load-bearing camouflage for what is actually a solid undergrad-level media literacy lecture. WheezyWaiter is not reinventing epistemology — SIFT is a known framework, boolean search is taught in middle school libraries, and “watch out for certainty” has been said by every honest scientist who ever gave an interview. But that’s not a knock. Most people don’t do any of this, and a 22-minute video that makes them do even one of these things is a net positive.

The strongest passage is the bit on certainty. It’s the kind of observation that only lands if you’ve actually sat with a lot of experts, and his pattern-match rings true — the people who know the most are usually the ones least willing to make sweeping claims. That’s a useful heuristic to carry into almost any information environment, including AI outputs.

The weakest thread is the “do your own research” critique, which is correct but slightly rhetorically stacked. He’s really arguing against motivated reasoning, not against independent research. The phrase has been hijacked by conspiracy culture, but the underlying act — going and learning things for yourself — is exactly what the rest of the video is teaching. A more honest framing would separate the verb from the ideology that captured it.

Score: 7. Nothing groundbreaking, but well-organized, well-paced, and the advice is all correct. The SIFT walkthrough and the certainty warning are worth the watch. If you already teach research methods to anyone — students, your own kids, your Facebook uncle — this is a shareable one.