Industry sponsorship as meta-evidence: a simulation study

Abstract Industry funding is often indispensable to scientific research, but it can also introduce systematic biases that align study outcomes with commercial interests. Recent philosophical work suggests treating funding disclosures as meta-evidence: evidence about the evidential force of first-order study results. This paper asks whether such meta-evidence can be used to improve collective inquiry. Using an agent-based model, we compare two strategies for aggregating evidence in the presence of industry sponsorship: down-weighting industry-funded results and adjusting them relative to publicly funded studies. The simulations show that both strategies are highly context-sensitive. Relative adjustment can correct for bias when public evidence is reliable and sufficiently broad, but it falls short when public data are noisy or sparse. Down-weighting depends on a difficult and partly arbitrary choice of discount weight. The results suggest that industry sponsorship can be epistemically relevant meta-evidence without yielding a straightforward rule for evidence aggregation.