Reproducibility, transparency, and open science: a cultural evolution is needed to move from ideals we celebrate to implementation

The paradox: reproducibility, transparency, and open science are widely endorsed across the research community, but poorly implemented

Reproducibility, transparency, and open science are widely endorsed across the research community as essential foundations of trustworthy science [1, 2]. Those factors facilitate scrutiny, error detection, and cumulative knowledge. Yet implementation remains inconsistent despite broad agreement on their importance. Researchers, journals, funders, and institutions frequently express support for these principles, but everyday research practices often fall short. This creates a paradox wherein these values are widely endorsed but still not routinely implemented. The challenge is therefore no longer convincing researchers that reproducibility, transparency, and open science are desirable, but understanding why implementation continues to lag behind endorsement.

The reproducibility crisis is well documented

Evidence demonstrates that reproducibility remains a substantial challenge across scientific disciplines. Replication projects frequently report lower-than-expected replication success, smaller effect sizes than originally published, incomplete reporting, and limited access to underlying data and analytical code. Importantly, despite extensive discussion of the problem, relatively little empirical evidence exists regarding interventions that successfully improve reproducibility [3].

Failure to reproduce findings does not necessarily imply that the original results were false. However, incomplete reporting, inaccessible data, and unavailable analytical code make independent verification challenging, making it difficult to distinguish genuine scientific uncertainty from avoidable methodological shortcomings [3].

Transparency: a critical distinction between formal and meaningful

Transparency is often assessed through visible indicators such as data availability statements (DAS) or preregistration records [4, 5]. However, these markers have little value if the underlying materials remain inaccessible or unusable. This distinction between formal and meaningful transparency is important. Formal transparency refers to visible indicators such as a DAS or preregistration record. Meaningful transparency goes further by ensuring that shared materials are sufficiently documented and made findable, accessible, interoperable, and reusable (FAIR).

Formal compliance may satisfy journal requirements, but without genuine accessibility it contributes little to reproducibility. For example, even journals requiring DAS have reported that most authors ultimately fail to provide usable datasets when requested [4, 5].

Open science promises accessibility, collaboration, and inclusivity

Open science aims to make research outputs more accessible, reusable and collaborative. Considerable progress has been made through open-access publishing, preregistration platforms and reproducibility initiatives. Nevertheless, implementation remains uneven. Barriers include limited infrastructure, insufficient training, technical challenges, and incentive systems that reward publication quantity and novelty more than openness and replication. Consequently, researchers may struggle to implement open science in practice [6].

These barriers are not merely technical. Researchers may lack access to appropriate infrastructure, secure repositories, computational support, or training needed to share data appropriately. Cultural, geographical and socioeconomic barriers are equally important. Concerns about being “scooped”, exposing errors, or investing substantial effort in activities that receive little academic recognition discourage many researchers from adopting open science practices even when they support them in principle.

While reproducibility, transparency, and open science are distinct concepts, they converge on the same goal of making scientific research more trustworthy, verifiable, and useful. Achieving that goal requires moving all three from aspiration to routine practice.

Awareness is not enough: a meaningful reform is needed

Awareness alone is insufficient. Surveys consistently show that researchers value reproducibility and transparency, yet these principles frequently remain secondary to competing career incentives. Lasting improvement therefore requires structural rather than educational reform. Journals should enforce reporting standards and transparency requirements, funders should require open practices where appropriate, and institutions should reward rigorous and transparent research alongside novelty [7, 8]. Educational initiatives remain valuable, but they are unlikely to produce sustained behavioural change unless accompanied by changes in research assessment and reward systems. The problem lies in a persistent misalignment between what the scientific community values in principle and what it rewards in practice.

Cultural evolution is needed: to change the behavior, we must reshape the landscape

The paradox outlined at the outset - widespread endorsement of reproducibility, transparency, and open science alongside limited implementation - is not a reflection of indifference, but of a suboptimal environment. This can be viewed through the lens of cultural evolution - researchers adapt rationally to the fitness landscape in which they work [9]. In evolutionary biology, a fitness landscape is an environment in which some behaviours are rewarded while others carry greater costs [9]. The current scientific landscape rewards publication speed, productivity and novelty more strongly than transparency, replication or careful documentation.

In any complex system, individuals do not just act on their ideals; they adapt to survive and thrive within their specific fitness landscape. If we can come up with creative solutions that would foster reproducibility, transparency, and open science, this could be a textbook example of cultural evolution. To change the behavior, we must reshape the landscape. Researchers generally respond rationally to the systems that evaluate and reward their work. Consequently, changing research behaviour requires changing those systems. Namely, institutions must offer tangible incentives, while journals and funders should move past checkbox mandates. A DAS is not enough; journals and funders should require submission of data and metadata, supported by random audits to ensure the “peak” of success is only reachable through transparent practice.

The research agenda also needs to follow. When we describe the problem, we should move to testing the solutions. We need an experimental meta-research agenda that treats policy as a variable to be tested. This means funding interventional trials that prove which specific requirements - like mandatory preregistration or open peer review - improve research. We also need incentive experiments to see how different grant evaluation criteria change the way scientists design their studies, alongside the development of standardized metrics that can distinguish formal from meaningful transparency. The goal is to make transparency and rigor a standard and rewarded part of the job, rather than an additional burden undertaken by a minority. Without such changes, reproducibility, transparency, and openness in research will remain aspirational rather than operational [10]. Science will only truly open when the systems that evaluate our work finally value the integrity of the process as much as the novelty of the result.

Authors’ contributions

Manuscript design: LP, VR, JŠ, MKS, IB. Manuscript writing: LP, VR, JŠ, MKS, IB. Revising manuscript: LP, VR, JŠ, MKS, IB. Approving final version of the manuscript: LP, VR, JŠ, MKS, IB.

Funding

The authors did not receive any funding for the writing of this editorial.

Data availability

No datasets were generated or analysed during the current study.

Declarations

Ethics approval and consent to participate

Not applicable.

Consent for publication

Not applicable.

Competing interests

LP, MKS, and IB are guest editors of the Collection of articles in BMC Medical Research Methodology, in which this editorial is published. VR and JŠ declare no conflict of interest.