Insights
Science is often described as a method for discovering truth. This description is accurate but incomplete. Science is also a social institution — one with its own hierarchies, incentive structures, funding dependencies, and cultural assumptions — and these institutional features shape what gets investigated, what gets published, and what gets ignored in ways that the standard account of scientific method does not fully acknowledge. The tension between science as a truth-seeking method and science as a human institution is one of the most important and least discussed features of modern knowledge production.
The philosopher of science Helen Longino argued that scientific knowledge is not simply the output of individual experiments but the product of a community of inquiry operating under shared norms. These norms include standards of evidence, criteria for what counts as a significant result, and judgements about which questions are worth asking. None of these norms are politically or culturally neutral. The decision to fund research into certain diseases and not others reflects the distribution of political power and economic interest, not purely scientific priority. The decision about which populations to include in clinical trials reflects assumptions about whose body counts as the default human body — assumptions that have historically caused measurable harm when those decisions were wrong.
This is not a claim that science is merely politics by another name. The methods of controlled experimentation, peer review, and replication represent genuine advances in humanity’s ability to distinguish reliable from unreliable knowledge. The argument is more precise: that these methods operate within a social context that they do not themselves determine, and that the results they produce are therefore shaped by that context in ways that are not always visible from inside the practice. A researcher who follows every methodological rule correctly can still produce results that are systematically biased if the question they were asked to answer was itself shaped by assumptions that went unexamined.
The replication crisis that emerged across multiple scientific disciplines in the early twenty-first century made this visible in a new way. Studies that had been published in leading journals, cited widely, and used to justify policy decisions were found, when subjected to rigorous replication attempts, to fail at alarming rates. In psychology, economics, and medicine, results that had been treated as established knowledge turned out to be fragile, context-dependent, or simply wrong. The causes were multiple: pressure to produce positive results, publication bias toward novel findings, small sample sizes, and the subtle influence of researcher expectations on experimental design. Each of these causes is internal to the institutional structure of science, not a violation of it.
What the replication crisis demonstrated is that the social organisation of science — the incentive structure, the publication system, the funding model — can produce systematic error even when individual scientists are acting in good faith and following established procedures. Error is not always the result of fraud or incompetence. It can be the result of a system that rewards certain behaviours and punishes others, and that therefore reliably produces certain kinds of mistakes. Understanding this does not require cynicism about science. It requires the same analytical attention to institutional structure that scientists routinely apply to the phenomena they study.
The practical implication is that consuming scientific knowledge responsibly — as a citizen, as a policymaker, as a reader — requires more than the ability to understand what a study found. It requires the ability to ask who funded it, what question it was designed to answer, what populations it included, whether its results have been replicated, and what the publication incentives of the researchers and the journal were. None of these questions undermine science. They are extensions of scientific thinking — the application of critical scrutiny to the institution of science itself, which is the only thing that can make science more reliable than it currently is.
