Contrary to the "crisis of trust" narrative, global surveys show consistently high public confidence in scientists. The real challenge is a crisis of influence, where the scientific voice is drowned out in a media ecosystem dominated by emotionally compelling anecdotes and misinformation, especially on social media.
Evidence is a critical input, but not the sole determinant of a decision. For instance, antibiotics are proven to clear infections, but a terminally ill patient may decline them based on their values. Evidence must always be combined with context, cost, and human values to reach a course of action.
Evidence-based medicine became standard because doctors are taught it in medical school and face legal/professional consequences for ignoring guidelines. Fields like education and conservation struggle with adoption because they lack this powerful combination of mandatory training and strong accountability mechanisms for practitioners.
The emotional pain for professionals who discover their long-held practices are wrong is immense. This visceral reaction is a major barrier to adopting evidence-based methods, as it's a direct assault on their professional identity, expertise, and the realization they may have unintentionally caused harm.
AI is poised to revolutionize evidence synthesis by automating the grueling, multi-year process of systematic reviews. The ultimate goal is to enable anyone to get an accurate, near-instantaneous summary of the entire body of research on a specific question, effectively creating meta-analysis on demand.
When researchers like Ian Chalmers challenged "eminence-based medicine"—deferring to senior doctors' opinions—they faced intense backlash. For systematically reviewing established practices and suggesting they lacked evidence, Chalmers was compared to a terrorist, highlighting the extreme cultural resistance to data-driven change.
The deadly link between front-sleeping and SIDS was detectable after the first two studies in 1970. Had scientists synthesized this evidence—a practice not yet common—they could have reversed dangerous advice decades earlier and prevented an estimated 50,000 infant deaths.
Most doctors don't analyze raw studies. They follow clinical guidelines which function as algorithms. These are the output of a massive, underlying effort by researchers to synthesize thousands of trials into "pre-processed evidence" like systematic reviews, making evidence-based care scalable and efficient.
Intuitively, showing at-risk youth prison horrors should deter crime. However, systematic reviews of the 'Scared Straight' program revealed it backfired. Participants were statistically more likely to commit crimes compared to control groups, demonstrating how well-intentioned policy can cause harm.
An ethical blind spot exists where people find it unacceptable to randomize which unproven school curriculum children receive, yet see no issue with implementing a single, equally unproven curriculum for everyone. This ignores that the randomized trial is the best way to discover what actually helps students.
A routine knee surgery performed on millions, believed to work based on mechanistic reasoning, was found to be ineffective when tested against a placebo (sham surgery) in a randomized controlled trial. This highlights that even visually intuitive interventions can fail in complex biological systems, making rigorous testing essential.
To persuade someone skeptical of evidence (e.g., on vaccines), confronting them with facts often backfires. A more effective, trial-tested approach is motivational interviewing: ask open questions about their concerns, listen without judgment, and gently offer accurate information, rather than trying to "win" an argument.
Intuition excels in areas like chess or boxing where we get immediate, repeated feedback. It fails in complex domains like choosing a charity or making social policy, where feedback is slow, noisy, or nonexistent. We mistakenly trust our intuition in these low-feedback environments where it's unreliable.
While intuitively appealing, small reductions in class size have little effect on student learning because they don't fundamentally change teaching methods. Evidence suggests a significant impact only occurs when class size is reduced drastically (e.g., from 30 to below 15), a very expensive intervention compared to more effective methods.