Scientific Literacy

Science is not a collection of facts. It is a method for producing reliable knowledge, and understanding that method — even roughly — changes your relationship to almost everything you believe.

The distinction matters because the products of science are everywhere, and they can be encountered in two fundamentally different ways. The first way is as authority: science says X, therefore X is true, and to question it is to be ignorant or ideologically motivated. The second way is as process: science produces X through a particular methodology that includes observation, hypothesis, testing, peer review, replication, and revision — and X is reliable to the extent that this process has been applied rigorously and repeatedly. These two relationships to scientific knowledge feel similar but are not. The first is deference. The second is understanding.

Scientific literacy is not the possession of scientific facts. It is the capacity to evaluate how scientific knowledge is produced and what level of confidence a given claim deserves. This requires learning a small number of distinctions that are almost never taught in school despite being essential to thinking about anything empirical.

Key Distinctions That Actually Matter

The difference between a single study and a scientific consensus is perhaps the most important. A single study, even a well-designed one published in a reputable journal, is a provisional finding. It has not been replicated, its methodology has not been widely scrutinized, and its conclusions may not hold under different conditions with different populations. Scientific consensus is something else entirely: it represents the accumulated judgment of a field after many studies, by independent researchers, using different methods, have converged on a similar conclusion. The scientific consensus on climate change, vaccine safety, and evolution is not the same epistemic category as a recent study suggesting that a particular food has a particular effect. Treating them as equivalent — which media coverage systematically encourages — is a form of illiteracy.

Similarly, correlation and causation. This distinction is well-known and still almost universally ignored in practice. The fact that two things vary together does not tell you that one causes the other. The reasons are multiple: a third factor may cause both; the relationship may be coincidental; causation may run in the opposite direction from what seems intuitive. Controlled experiments — where one variable is manipulated while others are held constant — are the methodology developed to isolate causation, and they are much harder to design than an observational study that finds a correlation. Knowing this, you become appropriately more skeptical of claims based on observational data and appropriately more impressed by well-designed randomized controlled trials.

Reading Statistics Honestly

Statistical literacy is a subset of scientific literacy and deserves direct attention. Relative risk versus absolute risk: a treatment that reduces your risk of a disease by fifty percent sounds impressive, but if your baseline risk is one in a thousand, the absolute reduction is small. P-values and statistical significance: a result that meets the conventional threshold for statistical significance is not necessarily large or practically meaningful — it means only that the result is unlikely to be due to chance at a specified level, which is a much more limited claim than it sounds. Effect sizes, confidence intervals, sample sizes — these are the actual content of a scientific result, and they are routinely stripped away in popularization.

Scientific literacy also requires understanding where science is most and least reliable. It is most reliable where controlled experiments are feasible, where outcomes are measurable with precision, and where the phenomena being studied do not change in response to being studied. Physics, chemistry, molecular biology — these are domains where the methodology works extraordinarily well and where the consensus is robust. It is less straightforwardly reliable in domains like nutrition science, psychology, and economics, not because the practitioners are incompetent, but because the subject matter is genuinely hard to isolate, confounding variables are abundant, and — in the case of the social sciences — what is being studied is human behavior, which is complex, context-dependent, and reactive. This is not an argument for dismissing these fields. It is an argument for calibrating your confidence appropriately.

The Civic Stakes

The ethical dimension of scientific literacy is civic. In a democracy, citizens vote on policies whose effectiveness depends on empirical facts — about public health, about climate, about economic mechanisms, about the effects of particular interventions. Citizens who cannot evaluate evidence are vulnerable to manipulation by people who can manufacture the appearance of evidence for predetermined conclusions. This has happened and is happening, on issues ranging from tobacco to lead to climate to pharmaceutical safety. The failure of scientific literacy is not an individual intellectual failing. It is a vulnerability in the political system.

Trusting science and understanding science are different acts. Trust without understanding is faith, and faith can be transferred to fraudulent claims wearing scientific clothing. Understanding does not require becoming a scientist. It requires knowing enough about the method to ask: Has this been replicated? By whom? What is the consensus of the relevant field? What are the known limitations of this kind of study?

These are small questions. They change everything.