Quantitative methods are widely regarded as guarantors of objectivity, comparability, and epistemic reliability across scientific disciplines. Over recent decades, however, their role has subtly shifted: quantification increasingly functions not merely as a methodological tool for reducing complexity, but as an implicit standard for what is considered real, relevant, and scientifically legitimate. This paper offers a systematic analysis of this shift by distinguishing quantification as a productive and consciously limited reduction from datafication as an epistemic practice in which data begin to substitute for reality itself. From an epistemological and ontological perspective, the paper demonstrates how statistical routines, institutional incentive structures, and technical standardization contribute to a situation in which methodological choices acquire unreflected ontological consequences. Drawing on discussions of the replication crisis, the dominance of statistical significance, the turn toward effect sizes, and examples from medical research, it is argued that many current problems in quantitative research are not the result of individual misuse or incompetence, but rather systemic effects of an underlying ontological narrowing. The central contribution of this paper lies in reframing the limits of statistical methods not primarily as methodological shortcomings, but as expressions of inherently bounded epistemic claims. Effect sizes, meta-analyses, and data-driven predictive models improve statistical precision, yet they do not bridge the ontological gap between measurability and meaning, aggregation and individual reality, or prediction and understanding. The paper concludes by arguing for an explicit limitation of epistemic ambitions and outlines a conception of science in which quantification is recognized as an indispensable – but ontologically non-privileged – instrument of knowledge production.
Brandt et al. (2026) studied this question.