Accelerating the American Scientific Enterprise

The most serious mistakes are not being made as a result of wrong answers. The truly dangerous thing is asking the wrong questions. —Peter Drucker, Management: Tasks, Responsibilities, Practices (1974)

OSTP’s RFI on ways to “Accelerate the American Scientific Enterprise” benefits from asking the right questions. Accelerating scientific discovery and technology-based innovation across the United States, however, is not best achieved through policy tweaks to grants (e.g., lotteries), directed handouts, or quantitative “science of science” modeling alone. Success will come from drawing insights from historical and contemporary programs, and from designing policy experiments and interventions that allow agencies to learn, adapt, and scale what works. These should be structured to produce results from which agencies can learn from and then either scale successful programs or shutter ones that are not having desired outcomes. Science and technology (S&T) policy must more systematically learn from failure.

This response to the OSTP RFI outlines policy options to strengthen the U.S. innovation system by supporting translation from discovery to application, scaling regional innovation ecosystems, supporting novel institutional models for research, and ensuring that the benefits of federally funded research reach all Americans. The response draws on more than 25 years of research, policy engagement, and institutional experimentation by the Arizona State University Consortium for Science, Policy & Outcomes (CSPO), including sustained collaboration with federal program managers, academic leaders, philanthropic sponsors, industry, and public stakeholders. CSPO’s work has focused on understanding and improving how the American Scientific Enterprise operates in practice, particularly through the development and testing of policy-relevant frameworks and methods such as Real-Time Technology Assessment (Guston and Sarewitz, 2002), Reconciling Supply of and Demand for Science (Sarewitz and Pielke, 2007), Public Value Mapping (Bozeman & Sarewitz, 2011), Anticipatory Governance (Guston, 2014), and Participatory Technology Assessment (Kaplan et al., 2021). Together, these efforts inform the recommendations offered below for strengthening translational capacity, regional coordination, institutional innovation, and public benefit in federal science and technology policy.

A central takeaway from this body of work is that “science will be made more reliable and more valuable for society today not by being protected from societal influences but instead by being brought, carefully and appropriately, into a direct, open, and intimate relationship with those influences” (Sarewitz, 2016). We find OSTP’s framing of the current challenge closely aligned with our perspectives and approaches.

First, we wholeheartedly agree with and laud the three-pronged approach centered on accelerating science, enabling technological innovation, and ensuring beneficial outcomes for all Americans. These aims align with well-established understandings of how innovation systems evolve, particularly with respect to the pace, direction (including market uptake), and character or public value of technological change.

Second, we agree that, not only in the last century but since the birth of the Republic, American exceptionalism has advanced not by remaining institutionally static in the face of changing demand and circumstances, but by introducing incremental, radical, and sometimes revolutionary changes that are “fit for purpose.” These would include giving birth to citizen science where volunteers worked alongside public officials as early as the Louis and Clark expeditions of the early 19th century, to cross-sectoral and inter-disciplinary co-production in the middle and later 1800s through the Morrill and Hatch Acts through the American land-grant university systems, and the public funding of basic research as foundational to advancing the “Endless Frontier” through the National Science Foundation in the mid-20th century.

Third, we agree that the better of these course corrections has come not by retreating from the previously set mission, but by making strategic, forward-facing programmatic additions and institutional innovations—not the “either” but the “and” approach. Shifting from a “science for science” base through the National Foundation Act (1950) to a “science for markets” course adjustment through the Bayh-Dole Act (1980) was as good for science as it was for markets. We believe the third aim of ensuring benefit to all Americans requires an approach that integrates the “science for science” component with the “science for market” component through a “science for society” component. Therefore, although we focus our answers on questions ii (2), iii (3), vii (7), and xii (12), we believe they will also help address the challenges articulated in questions vi (6) and ix (9).

ASU-CSPO responses to specific questions in the RFI

(ii) Supporting translation from discovery to application

Federal policy too often treats ‘translation’ as a downstream, linear process in which discoveries move from laboratories to markets primarily through patent licensing or startup formation. Government programs seek to translate knowledge into services or technologies directly (e.g., via NSF I-Corps or agency-backed startup programs) or indirectly through substantial tax advantages for venture capital. Evidence from decades of science and innovation policy research suggests, however, that translational failures are more frequently institutional and social than technical or even financial. Translation accelerates when research agendas, user contexts (and direct testing with user communities), regulatory environments, and societal values are engaged earlier and more deliberately across the research lifecycle. Federal policy should therefore expand translational support beyond technology transfer offices to include governance mechanisms that enable sustained, iterative exchange among researchers, public agencies, industry, and knowledge and technology users, particularly for services and products with significant societal implications.

The evolution of semiconductor manufacturing in the United States illustrates this point. Early federal investments did not simply fund basic research; they supported sustained coordination among universities, national laboratories, defense procurement agencies, and industry consortia, enabling rapid iteration between discovery, design, and production. In the 1960s and 1970s, firms such as Fairchild and Intel benefited directly from Cold War imperatives as semiconductor chips became central to weapons systems and national security infrastructure. Cost-plus contracts and mission-driven procurement allowed the industry to scale rapidly, driving down costs and diffusing computing capacity across the economy. Success came not from siloed research and offshore manufacturing, but from tight integration of R&D and production. Intel in its early years had no separate research labs, just as General Electric in the early twentieth century physically connected its research facilities to manufacturing plants.

Beginning in the 2000s, a combination of federal policy choices and economic rationales associated with globalization encouraged institutional and geographic separation of research, design, and manufacturing. Greater emphasis on frontier basic research, combined with globalized manufacturing and production strategies, weakened linkages between discovery and implementation. Translational bottlenecks followed, especially in areas such as health technologies, transportation systems, and other domains that may not appear “leading-edge” to academic researchers but shape Americans’ daily lives. For all the success of Bayh-Dole in incentivizing patenting by recipients of federal funds, the singular emphasis on patents as the primary technology transfer mechanism has shifted attention away from manufacturing capability and services implementation. Earlier technology transfer scholarship remains instructive, particularly in distinguishing between vertical transfer (moving from discovery to downstream application) and horizontal transfer (adapting existing technologies across contexts) (Brooks, 1968). Historically, the United States led in collaboration between manufacturing firms—often small and medium-sized—and nearby university laboratories to solve production problems, develop new equipment, and refine methods. In services, including for software pioneers, transfer similarly relied on exchange of know-how, universities acting as experimental partners and collaborators with the private sector, rather than orienting all engagement through the lens of patent licensing.

To accelerate translation, federal programs need to look beyond technical readiness levels (TRLs) in the relevant engineering field and in the technology itself to the institutional capacity of national labs, universities, and community-based centers to integrate social context, regulatory alignment, manufacturing or service delivery constraints, and user engagement. Programs that embed translational intermediaries rather than relying solely on intellectual property transfer will result in greater durability, broader uptake, and stronger alignment with public needs.

Our recommendation is that federal agencies should pilot and scale new translational programs that explicitly support institutional integration rather than only IP transfer or licensing. These could include competitively funded Translational Partnership Platforms that require sustained collaboration among research institutions, manufacturers or service providers, public agencies, and end users (represented via both geographically based and online community organizations), with funding structured to support iterative design, demonstration, and deployment rather than one-time commercialization milestones. Agencies should also expand support for manufacturing- and service-embedded research—including co-located facilities, shared staff appointments, and applied research consortia—particularly in regions with existing federal research assets. Finally, federal policy should rebalance translational incentives by complementing patent-centric metrics with support for horizontal transfer, workforce-embedded learning, and problem-driven collaboration, recognizing that many socially valuable innovations depend less on new patents than on institutional capacity to adapt, produce, and implement technologies at scale.

(iii) Scaling regional innovation ecosystems

Invention and innovation clusters have a long history in the United States, from nineteenth-century precision manufacturing in the Hartford region, to early twentieth-century automobile innovation in Detroit, to the mid-to-late-twentieth-century semiconductor cluster in Silicon Valley (the model subsequently studied, fetishized, and followed to varying degrees of success by policymakers worldwide), and today’s biotechnology concentration in Greater Boston. These clusters were never exclusive—multiple regions at any given time combined firms of different sizes, educational institutions, sources of risk capital, and skilled workers—but in each era, one or two clusters became dominant. Importantly, the historical record also shows that many government-led efforts to intentionally create new clusters have failed. New federal programs aimed at regional innovation should therefore proceed with humility and plan for a high rate of failure to achieve lofty goals. Regional innovation ecosystems succeed not because they replicate a formula, but because they evolve through place-based, relational processes that build upon local histories, institutional capacities, and social networks. Regions requiring a constant influx of government funding for technology development are not examples of successful ecosystem design and implementation.

Effective regional ecosystems involve sustained coordination among universities, workforce systems, local governments, small and medium-sized firms, and communities—particularly in regions with existing federal research assets such as national laboratories or federally funded research centers. Historical cases such as Research Triangle Park and Silicon Valley demonstrate that success emerged over decades through aligned investments in education, research, workforce development, and public-sector coordination, rather than through isolated grants or short-term initiatives. Contemporary experience reinforces this lesson. For example, Baltimore’s recent progress in strengthening its technology and entrepreneurship ecosystem has depended less on the scale of individual federal grants than on bottom-up coordination, dense social ties, and sustained convening among local actors (UpSurge Baltimore, 2025). By contrast, many recent “hub” initiatives struggle when they emphasize branding, competition, or startup counts over governance capacity and durable coordination.

These lessons suggest that federal policy should prioritize long-term regional capacity-building over winner-take-all models. One promising approach is to support new institutional roles, what we term innovation extension agents, analogous to the cooperative agricultural extension system that successfully diffused knowledge and technology across diverse local contexts for more than a century. Federal agencies could pilot and scale programs that fund locally embedded professionals and organizations responsible for ecosystem coordination: making connections among firms, universities, workforce providers, and communities; adapting national innovation priorities to local comparative advantage; and maintaining continuity across political and funding cycles. Rather than asking each region to reinvent this function or relying on short-term consultants, federal programs should codify best practices, support training and professionalization, and provide sustained funding for these connective roles. Such an approach would increase the likelihood that place-based innovation investments translate into durable regional capacity, rather than isolated successes or symbolic hubs.

(vii) Supporting novel institutional models for research

The complexity, scale, and societal implications of contemporary scientific challenges exceed the capacity and strategic focus on national labs and intramural research and technology development programs at classic federal agencies. They point to the need for research institutions that go beyond traditional ‘breakthrough’ or ‘leading-edge only’ models. Problems such as climate adaptation, energy transitions, public health improvement, and regional economic revitalization require sustained coordination across disciplines, institutions, and communities, as well as iterative engagement with users and affected publics. While national labs and universities remain indispensable sites of discovery and training, many of the institutional assumptions embedded in federal research funding, especially short grant cycles, discipline-based review, and limited support for boundary-spanning roles, constrain the ability of the research system to address use-inspired and societally embedded challenges. Novel institutional forms, including boundary organizations, mission-oriented consortia, and participatory research intermediaries, can enable long-term coordination, integrate diverse forms of expertise, and sustain engagement beyond individual grant funding cycles.

Historically, the federal government has repeatedly created new institutional models when existing arrangements proved inadequate. National laboratories, agricultural experiment stations, cooperative extension, and large-scale mission agencies emerged to address challenges that could not be met by individual investigators or industrial research alone in the first half of the 20^th century. More recently, agencies have experimented with community-based and participatory research initiatives that treat communities not only as sites of data collection or consultation, but as co-producers of knowledge. Programs such as NOAA’s long-standing regional climate adaptation partnerships and the Department of Energy’s emerging consent-based nuclear siting initiatives illustrate both the promise and the fragility of these models. Where federal support has been sustained over time, communities and researchers have developed durable relationships, shared methods, and decision-relevant knowledge. Where initiatives remain project-based or episodic, as in some newer programs, capacity building and institutional learning are harder to sustain, even when individual projects succeed.

These experiences suggest that federal policy should move beyond treating novel research institutions as temporary experiments and instead recognize them as essential components of the national research infrastructure. Agencies could support this shift by creating funding mechanisms that explicitly reward boundary-spanning functions, provide stable support for community-engaged research intermediaries, and enable integration across programs and agencies. This includes supporting training and career pathways for professionals who connect scientific expertise with community priorities; designing evaluation criteria that value legitimacy, trust-building, and learning alongside technical outputs; and coordinating investments so that community-based research capacity persists beyond individual funding cycles. By institutionalizing these models rather than repeatedly reinventing them, federal policy can strengthen the scientific enterprise’s ability to address complex societal challenges while rebuilding public legitimacy and accountability in publicly funded research.

ASU’s institutional structure—characterized by transdisciplinary research institutes, strong community partnerships, and experience coordinating large, multi-stakeholder initiatives—enables rapid experimentation with new organizational forms while maintaining rigorous evaluation and policy feedback. ASU thus offers a role model for new institutional research models even as CSPO could serve as a national resource for documenting best practices, training boundary-spanning professionals, convening agency and community partners, and developing evaluation frameworks that help federal agencies move successful community-based and participatory research initiatives from episodic pilots to durable elements of America’s research infrastructure.

(xii) Ensuring benefits of federally funded research reach all Americans

The American scientific enterprise is a complex system comprising many actors and forces that continually reshape how research is conducted and while demanding “continuous improvement in how the Federal government supports scientific research.” We advocate a new and integrated approach, which directs science and policy toward points of convergence among scientific, market, and public values and outcomes. This approach is necessitated by recent innovation dilemmas that can be classified as complex (e.g., power grid stability and cascading failures, and ecosystem responses to climate intervention technologies), wicked (e.g., urban housing affordability crises and global greenhouse gas reduction), or post-normal (e.g., climate adaptation planning for low-lying coastal cities and pandemic response measures). Responding to these challenges requires transdisciplinary, cross-sectoral, and distributional governance approaches that are anticipatory, adaptive, and reflexive.

Participatory Technology Assessment (pTA) is a platform engagement methodology that was developed, analyzed, operationalized, and implemented by CSPO in collaboration with its academic, informal science education, participatory science, and policy partners over the last fifteen years, encapsulating all four of these agreements. In April 2010, CSPO brought together a group of social science, public engagement, and science policy scholars and practitioners to launch the Expert and Citizen Assessment of Science and Technology (ECAST) network. ECAST combines expert and social assessment of science and technology with community, stakeholder, and public participation to inform policy and decision-making through a reflexive, inclusive, and adaptable pTA engagement methodology.

The pTA approach developed by CSPO and partners is built around three integrated layers of engagements: (1) academic partner-led layer of problem framing engagement with experts, stakeholders and lay publics to co-define the topics, contents and publics for engagement, (2) museum partner-led layer of inclusive, informed and deliberative engagement with target publics to produce useable outcomes for different academic, education and decision-making audiences, and a (3) policy think tank partner-led final layer of results integration engagement with the intended audience. The objective of pTA can vary: from filling democratic gaps and contributing to public dialogue to enriching policy advice and supporting institutional transformations. To date, federal sponsors have supported CSPO-led pTA projects for four specific objectives: (1) informing policy and decision-making (planetary defense by NASA), (2) improving scientific literacy (community climate hazards by NOAA), (3) mapping public values (human gene editing by NIH), and (4) innovating public engagement (nuclear waste siting by DOE). Philanthropy sponsors have also supported pTA research and application in cases involving solar geoengineering research and carbon dioxide removal (Alfred P. Sloan Foundation) and autonomous vehicles (Charles Koch and Kettering Foundations). Additional details can be found at: https://cspo.org/areas-of-focus/pta/ 

Finally, in a “Day One” Policy memo drafted in December 2020 and updated in January 2025, CSPO policy scholars (Weller, Govani, and Farooque, 2025) recommend “a robust, adaptable and scalable participatory assessment capacity to address complex issues at the intersections of science, technology, and society.” Specifically, we recommend establishing a special unit within the Science and Technology Policy Institute (STPI)—an existing federally funded research and development center (FFRDC)—to provide evidence-based, just-in-time, and fit-for-purpose capacity for Participatory Technology Assessment (pTA) to the White House Office of Science and Technology Policy and across executive branch agencies.

Rigorous participatory and multi-stakeholder engagement supports responsible decision making where neither science nor existing policy provides clear guidance. Participatory technology assessment (pTA) is an established, evidence-based process that assesses public values, manages sociotechnical uncertainty, integrates living and lived knowledge, and bridges democratic gaps in contested science and technology issues. By incorporating broader community expertise and experience, pTA surfaces plausible alternatives and solutions that are often overlooked in technocratic policy- and decision-making processes, thereby strengthening legitimacy, transparency, and accountability. When systematically integrated into research and development processes, pTA enables anticipatory governance by assessing socio-technical futures, engaging communities and stakeholders, and guiding decisions, policies, and investments toward desirable outcomes.

A pTA unit within STPI would build and sustain a shared repository of knowledge and best practices across government while providing pTA design, development, implementation, integration, and training services for the executive branch on emerging science and technology issues. By integrating public and expert value assessments, the next administration can ensure that federal science and technology decisions deliver the greatest possible benefit to society.