How we evaluate evidence

shrinkiatry covers a profession, so a lot of what we report is institutional fact: how long residency is, what a board requires, what a regulation says. For those, we go to the body that sets the rule. But we also report on research and trends, and there the discipline of weighing evidence matters. This page explains how we do it.

Institutional facts versus research claims

We treat two kinds of statements differently. An institutional fact, like the length of residency or the schedule of a controlled substance, has a single authoritative source: the organization that sets it. A research claim, like how effective a treatment is or how common burnout is, lives on a spectrum of evidence and needs to be handled with more care. Confusing the two, treating a contested finding as if it were a fixed rule, is a common way that health writing goes wrong.

We prefer primary and authoritative sources

For rules and structures, we cite the source of the rule, such as the ACGME, the ABPN, the DEA, or HRSA. For clinical and scientific claims, we prefer peer-reviewed research, systematic reviews, and guidelines from recognized bodies over secondary coverage. A systematic review or a professional guideline generally carries more weight than a single study, and a single study carries more weight than an expert's unsupported assertion.

We separate what's shown from what's claimed

A single study is not a settled fact. When we describe research, we try to convey the strength and the limits of the evidence, including how large or replicated a finding is, what population it applies to, and what it doesn't prove. Early or single-setting results are labeled as such. An association is not a cause, and we try not to let careful research get reported as more certain than it is.

How we handle conflicting evidence

When good studies disagree, we don't pick the one we like. We say the evidence is mixed, describe the shape of the disagreement, and lean on the weight of the better-designed and better-replicated work. Where a question is genuinely unresolved, the honest answer on the page is that it's unresolved.

We state uncertainty plainly

When evidence is mixed, thin, or evolving, we say so. That's especially true for fast-moving areas like artificial intelligence and telemedicine regulation, where today's accurate statement may need updating tomorrow. We'd rather a reader leave with an accurate sense of uncertainty than a false sense of resolution.

We attribute and date numbers

Statistics are tied to their source and dated where it matters. If a figure is an estimate or a projection, we describe it that way rather than presenting it as a fixed count, and we prefer ranges to false precision when the underlying data is itself a range.

We update

When better evidence or newer data appears, the page changes, and significant changes are noted under our corrections policy, consistent with our editorial standards.