Scientific reputation plays a central role in shaping careers, funding decisions, and public trust in research. Today, reputation is largely inferred from bibliographic indicators such as publication counts, journal impact factors, and citation-based metrics like the h-index. While these measures are easy to compute and widely used, they capture only a narrow slice of scientific contribution. As a growing body of meta-scientific literature has shown, overreliance on such metrics creates misaligned incentives that can undermine the quality, transparency, and reliability of science itself.
In our article “A Reputation System for Scientific Contributions Based on a Token Economy”, we propose a different way of thinking about scientific reputation. Instead of treating reputation as a by-product of publications, we conceptualize it as a deliberately designed incentive system—implemented through a decentralized, blockchain-based token economy—that rewards a broad range of valuable scientific activities while discouraging fraud and low-quality output.
We argue that scientific reputation should be reframed as a multi-dimensional, explicitly governed system rather than an emergent artifact of citation patterns. Further, reputation should reflect all meaningful scientific contributions, including:
- Peer reviews
- Replication studies
- Negative results
- Data sets and software
- Preregistration of studies
- Community service and validation work
The Reputation Token Economy: How It Works
At the heart of the proposal is a two-tier token model:
- Spendable tokens, held by approved issuers such as journals or conferences
- Awarded tokens, held by researchers and representing their accumulated reputation
Spendable tokens can be transferred exactly once. When an issuer awards tokens to a contributor—for example, after accepting a paper or validating a dataset—the tokens become awarded tokens and are permanently bound to the recipient’s identity. Awarded tokens cannot be transferred, sold, or exchanged.
The Role of the Consortium
The total supply of spendable tokens is controlled by an independent consortium, envisioned as a network of academic institutions or trusted organizations. Issuers must apply annually to the consortium, specifying how they plan to validate contributions and distribute tokens. The consortium allocates a limited number of spendable tokens to each issuer, enforcing scarcity and preventing reputation inflation.
Introducing “Skin in the Game”: Collateralized Reputation
A particularly novel feature of the system is the introduction of reputation collateral. Researchers may optionally stake a portion of their existing reputation when submitting a contribution. This collateral is at risk: if the contribution is rejected, the staked reputation can be burned; if accepted, the researcher may receive a higher reward.
Formally, issuers distribute their available spendable tokens (\(N\)) across the set of accepted contributions (\(C\)) using the reward function:
Here, \(\beta\) controls how much reputation is distributed equally versus weighted by collateral, and \(\varepsilon\) ensures numerical stability. This hybrid approach avoids disadvantaging early-career researchers while still rewarding confidence and quality.
Issuers themselves are incentivized to use this mechanism responsibly. The number of spendable tokens an issuer receives in the next cycle depends partly on how much collateral was burned due to rejected submissions:
where \(D_{y-1}\) denotes the set of rejected contributions in the previous year. This creates a feedback loop that discourages superficial acceptance and penalizes unfair rejection.
Example
Consider a hypothetical scenario in which a scientific conference (A) and a scholarly journal (B) seek to participate in the proposed reputation token system in order to reward contributors for validated scientific work. To be admitted into the system, both organizations must submit an application detailing their funding model, peer-review procedures, and mechanisms for documenting and auditing their editorial decisions. Upon approval by the consortium—composed of representatives from the scientific community and relevant academic stakeholders—they receive an initial allocation of spendable reputation tokens, which they may distribute in accordance with their approved allocation scheme.
The quantity of tokens allocated to each organization is a function of its prior transaction history within the system, as discussed below. In the absence of such a history, newly admitted participants receive a fixed baseline allocation determined by the consortium (100 tokens in this illustrative example).
Conference A accepts Alice and Bob’s work and rewards them with their allowed tokens which are distributed according to the above formula. Since neither Alice nor Bob put a collateral on their submission, the tokens are evenly distributed among them.
Now they all want to submit to journal B. Since they have some reputation they decide to back their contributions with it to maximize rewards.
Bob and Carlos are accepted. Alice isn’t, so she loses her collateral. The rewarded tokens are split weighted by the respective collateral amount between Bob and Carlos as follows.
After spending the tokens both conference A and Journal B have to renew their application at the consortium. Now upon approval their token budget is partly determined by the amount of collaterals they destroyed in the previous iteration (see second term). Since the journal rejected some contributions backed by collaterals they receive a larger token allowance this time.
Let’s say now conference A accepts a large fraction of submissions because they are run by a paper mill organization. Their token allowance can not grow significantly with respect to their honest competitors.
If on the other hand they randomly reject works to boost their score in the coming iteration, contributors will complain, stop submitting their work in the future and lose trust. The conference might even be turned down by the consortium based on such allegations.
After a certain number of iterations the spendable token allowance should model the value of a venue to the community and the token balance of researchers their reputation. A similar model can also be derived for peer reviewer reputation.
Conclusion
By treating reputation as a designed, auditable, and multi-dimensional asset—rather than a proxy derived from citations—it opens new possibilities for rewarding integrity, rigor, and contribution diversity in science. In the paper, we offer a concrete and technically grounded blueprint for reimagining how trust and value are generated in the scientific enterprise.
Find the paper here:
https://doi.org/10.1007/978-3-031-72437-4_3
2024, Christof Bless, Alexander Denzler, Oliver Karras, Sören Auer.
A Reputation System for Scientific Contributions Based on a Token Economy.
In: Antonacopoulos, A., et al. Linking Theory and Practice of Digital Libraries. TPDL 2024. Lecture Notes in Computer Science, vol 15177. Springer, Cham.
And the PDF here:
-> Download PDF
