This phrase “do your own research” is usually invoked by people who have nothing to substantiate their claims. In logic, onus probandi places the burden of proof squarely on the claimant. If you assert something, you must provide evidence. Instead, when pressed, many retreat to “do your own research,” which is itself a fallacy: a deflection used to avoid demonstrating that their claim has any merit.
The irony is that most people who rely on this mantra have little understanding of how research actually works. Their “research” consists of superficial search‑engine queries that simply mirror their existing beliefs. Search engines are designed to optimize relevance, not epistemic rigor, so they naturally reinforce confirmation bias rather than challenge it.
There’s a newer twist as well: people now outsource their “research” to AI. I recently searched Google for a post I wrote.
Gemini produced a summary that attributed my views to Tracy Beth Høeg—someone whose pandemic positions are diametrically opposed to mine and have been repeatedly critiqued on Science‑Based Medicine and other sources. The AI hallucinated a connection that doesn’t exist. It’s a perfect example of why relying on AI for answers is dangerous: it can produce confident, authoritative, and completely wrong results.
AI is a tool, not an oracle. I use it to clarify writing, debug code, or help identify methodological flaws I may have missed, but I never use it as a substitute for actual analysis.
Real research requires education and training. Reading a study is easy; critically evaluating its methodology is not. In graduate school, I spent an entire semester in one course doing nothing but dissecting studies for flaws. Basic statistics isn’t enough. You need a full year of biostatistics to competently assess validity. Most people invoking “do your own research” haven’t taken a single statistics course.
A common tell is the appeal to authority such as “I work in healthcare.” This is meaningless. It usually indicates a technical or ancillary role, not expertise in research methodology. Chiropractors frequently misuse the title “doctor” in the same way. Even most physicians are not trained to critically evaluate studies at a methodological level.
When these individuals do provide sources, they are rarely primary. Many don’t even know what a primary source is. Instead, they link to YouTube videos, Rumble posts, or propaganda sites masquerading as journals, such as Science, Public Health Policy and the Law. Sometimes they cite preprints, which have undergone no peer review. If a preprint is more than six months old and still unpublished, that’s usually a sign the methodology is flawed.
Even when they link to actual journals, the sources are often predatory or pay‑to‑publish. Distinguishing legitimate open‑access journals from predatory ones requires training, again, something most of these individuals lack.
When I ask a question that someone with research training could answer in a single sentence, but that cannot be easily found via search, they freeze. It’s the deer‑in‑the‑headlights moment that reveals the truth: they never had evidence to begin with.
I always provide evidence when asked. When someone tells you to “do your own research,” it’s usually because they have none.
