Background

Communication of science with decision makers serves as a prerequisite for use of science¹ in policy (Akerlof, 2018), but which forms of communication and what types of use? Scholars point to few recommendations for the communication of science to decision makers that are universally generalisable across a wide diversity of circumstances, or even to certain types of contexts (Contandriopoulos et al, 2010; National Research Council, 2012). A 2017 National Academies report on the science of science communication highlighted this deficit within the research literature on communicating with policymakers (National Academies of Sciences, Engineering, and Medicine, 2017). This study brings together concepts from science communication (Fischhoff, 2013), knowledge to action (Contandriopoulos et al, 2010; Lemos et al, 2012), and knowledge use in policy (Weiss, 1979; Whiteman, 1995) to test a new four-dimensional model that builds on these multidisciplinary literatures and explores how barriers to communication and knowledge transfer relate to different types of science use in the US Congress. By evaluating these barriers across two different types of science use and two groups of congressional advisers of differing professional and educational backgrounds, we seek to assess whether our model’s hypotheses hold true.

Ironically, the lack of research on communicating science with Congress comes during a period of time in which the institution itself has become more focused on communication and constituent services to compete for majority status with each electoral cycle, often to the detriment of legislation (Crosson et al, 2021). Since 1980, when the Republicans gained a majority in the Senate following longstanding Democratic control, the two parties have increased their communication capacity in their fight for electoral gains, while lacking incentives to cooperate in policy processes that would hand the majority legislative victories (Lee, 2016). We hypothesise that along with concomitant trends—increased party ideological consistency and organisational strength within Congress (Theriault, 2008)—the hyper-partisan context increases the frequency with which the authority of science is called upon to support established positions, as opposed to formulating new policies, and that this changes the nature of communication of science with decision makers.

Legislative staff and fellows as principal agents in information transfer

In seeking to improve Congress’s capacity for accessing and using scientific and technological (S&T) information for policy, policymakers and scholars alike have called for assessing what sources the institution currently depends on to meet its needs (Fretwell et al, 2019). Congress is predominantly an institution of generalists, not specialists. Elected representatives rely on staff as their principal agents to act on their behalf, staying abreast of new information and vetting it for use by the office within portfolios of specific issues, such as health, environment and natural resources, energy, education, and agriculture (Petty et al, 2018). Staff also lead the development and writing of new legislation (Montgomery and Nyhan, 2017), often in collaboration with external interest groups or other offices. The 2017 and 2019 Congressional Capacity Surveys (Furnas and LaPira, 2019) describe the typical staffer as someone 26–30 years old who has a total tenure of 3.1 years on the Hill. For about half of staff members, tenure also includes time in internships. The range of topics covered by a staff member in a personal office can be large, especially in the House, in which offices have fewer staff. These demographic and institutional factors limit knowledge specialisation and congressional capacity (LaPira et al, 2020).

The relative importance of personal office staff in information gathering and evaluation within Congress has grown as congressional committees, designed to serve as the primary sites for these processes (Smith and Deering, 2014), have become less powerful (Kornberg, 2023). While the number of congressional staff overall has declined, the number of committee staff members relative to those in personal offices has also waned, especially in the House of Representatives (Reynolds, 2022). By 2015, for every one committee staff member in the House and Senate there were approximately five in personal offices (Reynolds, 2020), a ratio that was closer to one to three or four from 1979 to 1994. Further, committee activities—including hearings which play an especially important role in congressional information flow and deliberation (Ban et al, 2023)—have become less frequent (Reynolds, 2022). When hearings are held, they increasingly are positionally framed, contributing to partisanship, and focus on policy problems and government implementation rather than discourses on solutions (Lewallen et al, 2016).

Given this context, the direct placement of science and engineering fellows predominantly within congressional personal offices arguably serves as one of the most effective ways for Congress to increase levels of S&T knowledge within the institution (Golden, 1988; Fainberg, 1994; Stine, 1995), thereby reducing barriers to access and use. But very little research has been conducted on these types of programmes (Alberts et al, 2018). The Congressional Science & Engineering Fellowship, managed by the American Association for the Advancement of Science (AAAS), and the John A. Knauss Marine Policy Fellowship, run by the National Oceanic and Atmospheric Administration (NOAA) National Sea Grant College Program, are among the most prominent. During each of the fellowship years ending from 2015 to 2019, together the two programmes sent approximately 46 master’s and doctoral level scientists and engineers to serve in Congress. The fellows join either the personal offices of members of Congress or committees and augment existing staff, arguably increasing institutional capacity to access and use S&T information.

Communicative barriers to the usability of science

Usability is defined as ‘a function of both how science is produced (the push side) and how it is needed (the pull side) in different decision contexts’ (Dilling and Lemos, 2011: 681). The study of usability arises from the observation that the conventional ‘loading dock approach’ of publishing research and expecting it to be ‘picked up’ and used does not take into account potential users’ needs, thereby failing to produce usable knowledge (Cash et al, 2003). Lemos and colleagues (2012) specifically distinguish between science that is useful (what scientists believe is needed) and usable (what users know they can use).

The use of science for policy—sometimes also termed ‘evidence-based’ or ‘evidence-informed’ policy (Cairney, 2016)—has been studied predominantly within governance or decision-making networks that operate close to the resources that they manage, and may even exist for that sole purpose, such as crop advisors and stakeholders within NOAA Regional Integrated Sciences and Assessments (RISA) programmes (Lemos et al, 2014a; 2014b). One of the most well-known reviews of case studies in this area explored the mobilisation of science and technology for sustainable development, finding that information salience, credibility, and legitimacy served as key factors for use (Cash et al, 2003).

A focal area of the supply and demand literature—and the related emergent field of evidence-based policy—is identifying barriers that impede the use of research evidence (Cairney, 2016). A number of studies have synthesised this literature (Innvær et al, 2002; Mitton et al, 2007; Lemos et al, 2012; Oliver et al, 2014; Cairney, 2016). In Oliver and colleagues’ systematic review (2014) of the most frequently cited barriers to the use of evidence, availability and access to information were the most frequently cited (63 studies out of 145 total), followed by clarity/relevance/reliability of findings (54), timing/opportunity (42), policymaker research skills (26), and costs (25). These findings largely align with other studies.

Yet few studies of science use in legislatures exist (Wilsdon and Doubleday, 2013; Kenny et al, 2017b; Geddes et al, 2018; Guillot-Wright and Oliver, 2022). A recent systematic review of barriers to the use of research evidence in legislative contexts located only 21 (Ouimet et al, 2023). In the UK Parliament, Kenny et al (2017b; Rose et al, 2020) conducted one of the most extensive studies of factors related to research use. They found that the credibility of the information source was a particularly important factor, as were the availability and accessibility of the information, and the policymaker’s attitudes and time limitations.

Table 1 presents the barriers found by Kenny et al (2017b; Rose et al, 2020) along with those identified in a review conducted by Lemos et al (2012). In this study, we adopt a framework from the Lemos et al (2012) review that details three communication factors that can serve as impediments to usability: fit, or how users perceive knowledge would meet their needs; interplay, or how new knowledge intersects with existing decision routines (negatively and positively); and interaction, or how the collaboration between scientists and decision makers in producing knowledge increases its use. Lemos and colleagues found that these factors do not act independently, but ‘shape each other to increase or constrain usability’ (Lemos et al (2012): 791). Notably, distinctly political factors, such as would reflect contexts in which members are unlikely to stray from established party policy positions, are absent from the list.

Each of the three factors—fit, interplay, and interaction—describe aspects of communication that are hypothesised to impede, or facilitate, the transfer of scientific knowledge to the policymaker. Further, they function as potential pathways for the hypothesised effects of augmenting congressional staff with science and engineering fellows on congressional science use. By communication, we mean one-way or two-way exchanges of information and knowledge that constitute one facet of the myriad forms of engagement in which audiences become invested in issues, characterised by their involvement along any number of cognitive or behavioural dimensions (Moser and Pike, 2015). Information content, the relationship between communicators, and decision-making context appear in communication theory as key to message receptivity (McCombs and Becker, 1979; Windahl et al, 1992). Indeed, as noted by Fischhoff, the first step in crafting science communication is to understand the context, or the nature of the decision at hand (Fischhoff, 2013). From Lemos and colleagues’ framework of science usability (Lemos et al, 2012), we hypothesise that the communicative impediments staff face both in cases of strategic and substantive use will fit into the typology of fit, interplay, and interaction.

A model of science communication and use in Congress

While ‘achieving impact’ in communication of science with decision makers is typically interpreted as influencing policy decisions (Oliver and Cairney, 2019; Cairney and Oliver, 2020), the use of research can take many forms. Weiss (1979) identified seven meanings of research utilisation, ranging from a linear ‘knowledge-driven’ model in which basic research informs applied research, development, and application, to a ‘research as part of the intellectual enterprise’ model in which science and policy mutually influence each other, and are influenced by societal trends. Whiteman (1985) developed a simpler typology that recognises that the use of information does not just occur before a policy decision, but often after the decision has already been made. Substantive use of information contributes to the development of a policy position in the absence of a strong prior commitment. Elaborative use extends or refines a position that has already been formed. In strategic use, information serves to advocate for, or reconfirm, a position that has already been defined.

This study focuses on the two ends of Whiteman’s spectrum—substantive and strategic use—as elaborative and substantive use occur under similar conditions. In his 1985 analysis of use of information from the Office of Technology Assessment (OTA) in congressional committees, Whiteman found that strategic use of information from the assessments occurred at all stages of the policy process: formulation, information gathering, modification, and deliberation, and was the most common form of use (Whiteman, 1985). Substantive use, alternately, only occurred during the formulation and information-gathering stages. Thirteen cases of substantive use were observed (17%), compared to 35 cases of strategic use and 30 of elaborative use (Whiteman, 1985: 299). Moreover, strategic use was most common in high-conflict issue areas, while substantive and elaborative use occurred predominantly under low-conflict conditions. More use of OTA reports in all forms: strategic, elaborative, and substantive, occurred for issues of higher salience, as judged by committee staff.

These findings align with the work of Contandriopoulos and colleagues (2010). They contended that two of the largest sources of variation in how science is used for policy are the dimensions of political polarisation, and science producer and user willingness to contribute time and resources to knowledge exchange. By overlaying Whiteman’s typology with a similar figure that Contandriopoulos et al based on Weiss’s research utilisation categories, in Figure 1 we illustrate how the science communication factors identified by Lemos et al (2012) relate to the conditions under which substantive and strategic use are most likely to occur in Congress.

View Full Size

Four dimensions capture the dynamics of science communication in Congress: issue politicisation, who bears the costs of information transfer, type of information use, and the characteristics of the communication itself

Citation: Evidence & Policy 20, 3; 10.1332/17442648Y2023D000000013

• Download Figure
• Download figure as PowerPoint slide

Model-derived hypotheses

Three dimensions of our model relate to the decision context: who bears the cost of information transfer, issue polarisation, and type of information use (Figure 1). The fourth dimension represents communication factors that lower barriers to use of the information, and hence increase its usability. In order to assess the model, we evaluate the frequency of information use among congressional staff and science and engineering fellows, and whether there are differences in the science communication needs of these decision makers in contexts of supporting or defending a previously established policy (‘strategic’ use), or substantiating a new policy decision (‘substantive’ use) (Whiteman, 1985) (Figure 1).

We hypothesise that: 1) four decades after Whiteman conducted his foundational research in Congress, strategic use of science will remain more frequent than that of substantive use; 2) impediments in using science for policy will differ based on the type of use; and 3) communication factors can reduce barriers to knowledge transfer (Figure 1). We also hypothesise that fellows, given their scientific fluency and expertise: 1) should experience fewer barriers to science use for policy than other staff; and 2) because their year-long presence arguably builds office resources and capacity for more in-depth issue exploration, they will engage in more substantive use.

When offices are developing positions on policy issues, we anticipate that staff members and fellows will receive little outside assistance in acquiring scientific information, but that for fellows this attention deficit will prove less problematic due to their scientific fluency. Once offices align with a policy direction, such as one promoted by their party, other organisations will help bear the costs of gathering, synthesising, and making this information accessible to promote a specific legislative agenda.

Methods

In order to address the deficit of information on science communication and use in Congress, and test our model, we collected data in three stages. First, in a series of in-depth interviews (n=16, 12/2017–3/2018), we explored how personal office staff members serving members on energy, natural resource, and science committees use science for policy, and the barriers they experience in doing so. Committee membership serves as an indicator of member legislative interests and that the office has personal staff responsible for issues within the committee’s jurisdiction. We created coded categories representing the range of responses from staff and broadened the interviews to a larger number of House and Senate staff members who served in energy, natural resource, and science portfolios during the 115th Congress (n=42, 9/2018–2/2019).

The response rate for the second sample is 9%, which is typical for congressional studies (Hertel-Fernandez et al, 2019). As Long et al note, surveys with this population are particularly difficult to conduct (Long et al, 2021). The same questions were also used in an online survey with Congressional Science & Engineering and Knauss fellows who participated in cohorts that completed their year in Congress between 2015 and 2019 (n=68, 4/2018–1/2019, response rate 29%). Each set of respondents was asked to identify the top five major policy issues on which they worked and whether the office already had established a position on it. They were then asked about their use of science and what barriers they had experienced in doing so. The questionnaires are available through the Open Science Framework (OSF) at https://osf.io/e62s7/.

The study protocol was approved by George Mason University’s Institutional Review Board (1150984-1). Congressional staff were identified through calls to their office and recruited for interviews by email. Fellows were recruited through their scientific societies, AAAS, and NOAA Sea Grant.

Content analysis

Between two to three people coded barriers from the staff interviews using a codebook developed both from previous literature and a close reading of the texts (Supplementary Materials, Table 1, https://osf.io/e62s7/). They obtained a Krippendorff’s α of 0.82 or greater across all variables, measured using a program for intercoder reliability analysis developed by Hayes and Krippendorff (2007). To ensure consistent interpretability across studies, reliabilities of α > 0.80 have been recommended (Krippendorff, 2004). These lists of barriers served as response options within the fellows survey; respondents could also add barriers not on the list. The content analysis of staff in-depth interview data also provided the basis for the categorisation of policy issue areas and the ways that staff and fellows use scientific information in Congress.

Determination of strategic and substantive use

We asked both staff and fellows to identify at least five issues on which they were currently working, had been working until very recently (if they were still in the position), or on which they had worked. Next, they were prompted to categorise the issues as to whether the office policy position was already established and whether scientific information was relevant. Issues for which scientific information was relevant were categorised as ‘use of science’—these were further divided by whether the office had an established position (strategic use), or not (substantive use), an adaptation of the typology developed and operationalised by Whiteman (1985). We then conducted a chi-square test to evaluate whether staff experienced differing frequencies of barriers to use of science between these two categories.

Samples: congressional staff and fellows

Findings

Types of use

We asked respondents to provide additional detail on one of the five policy issues on which they had worked closely, divided approximately in half between strategic and substantive issues. During the in-depth interviews with staff members, we asked how they used scientific information in that case. We derived nine categories for common types of ‘use’ (Figure 2). Notably, ‘use’ in five of the nine categories represents a form of communication: floor speeches, hearings, media releases, internal office discussions, and external conversations with stakeholders. Congressional staff members are, in fact, science communicators.

View Full Size

Conceptual use of science (‘background on the issue’) ranks high in both strategic and substantive use

Citation: Evidence & Policy 20, 3; 10.1332/17442648Y2023D000000013

• Download Figure
• Download figure as PowerPoint slide

In the short-form interviews with staff, we again asked: “What have you done with the information you obtained from all [helpful] sources you named?” Interviewees (n=41) responded to the list of potential forms of use described above. The most common types of science use described by the staff members were providing background on the issue and internal office communication with the member of Congress or staff. Virtually all staff used the information for these purposes (respectively, 92%/96%, strategic use; 100%/94%, substantive use). One of the other ways in which researchers have traditionally classified research use, in addition to the Weiss (1979) and Whiteman (1985) typologies, is as conceptual, instrumental, and symbolic (Pelz, 1978; Albæk, 1995). Forms of conceptual use to inform the staff members, others in the office, and stakeholders, about the science of the policy issue tended to be more common than instrumental forms of use such as developing legislation, media generation, oversight, and committee activities. Both strategic and substantive use occurred across all activity categories with no significant difference between the two.

In the issues described in-depth by fellows in the survey, science was similarly used most frequently as background information (88%, strategic use; 84%, substantive use), in internal office communication (81%; 81%), and for developing or amending legislation (73%; 84%).

Barriers to use of science by context

We identified 13 barriers in our initial staff interviews that largely fall across the three communicative criteria of information fit, interplay with the user’s decision context, and interaction or communication (Figure 3). Lack of trustworthiness and perceived bias can be an attribute of either the information or the communicator, and hence can fall either in fit or interaction. But the difficulties that staff members say they experience in communicating science (58%, strategic use; 59%, substantive use) fall outside of the Lemos et al (2012) typology, pointing to the needs of the decision maker to convey science to their own external audiences. The finding above that much of the use of science in the policy space is communication gives further credence to scientific fluency as a new facilitator/barrier, and aligns with the increasing emphasis within Congress on communication over legislation.

View Full Size

Reported barriers in strategic and substantive use fall predominantly under fit, interplay, and interaction. Lack of trustworthiness and perceived bias can describe either information fit or interactions with an informant. There are significant differences between the barriers to substantive and strategic use for staff: ^a X2(1, n=41)=4.377, p=0.04; ^b X2(1, n=41)= 7.712, p= 0.01; ^c X2(1, n=41)=4.361, p= 0.04; ^d X2(1, n=41)= 14.473, p<0.01; but no significant differences for fellows.

Citation: Evidence & Policy 20, 3; 10.1332/17442648Y2023D000000013

• Download Figure
• Download figure as PowerPoint slide

The most frequent barriers to use of science by congressional staff were not having enough time (strategic, 67%; substantive, 94%), information source bias (71%; 76%), and the complexity of science (67%; 76%) (Figure 3). Lack of time was the most frequently reported problem. Staff members said that the lack of time is caused by both overwork: “It is not a secret that congressional staff are underpaid and overworked”, and the pace of congressional hearing and voting schedules, in which there is little lead time: “We may not get a notice of a hearing until a couple days in advance. Then we are scrambling”.

Source bias was cited by staff members as a constant factor in their work that came with their jobs on Capitol Hill. As one said: “They will present us with what helps their argument the most. We try to take it for what is”. Indeed, it wasn’t even necessarily a bad thing, said some: “I’m not necessarily using it in the pejorative”.

The inherent complexity of science, in some fields more than others, was the third most frequently cited barrier. It interacted with other barriers, such as time and expertise. More complex science required more time from staff members: “If science is tough, it takes more time to understand it, and that makes complex science difficult to work with”. And the less topical expertise that staff have, the more complex the science becomes: “Without a background in that, the science of it is complex”.

Staff experienced more barriers in substantive use of science, in which their offices did not have policy positions, than with strategic use, in which they did (Figure 3). Of 13 barriers to substantive use, only one was mentioned by less than half of respondents: transparency and legitimacy of the scientific process (29%). Staff were more likely to say they encountered problems with lack of time (94%), presentation of the information (88%), difficulties finding or accessing the information (71%), and knowing whom to contact for information (71%).

In contrast, fellows said they experienced far fewer barriers in using science for policy than staff, and the types of barriers for strategic and substantive use were similar. While science communication was a problem that staffers said they faced, the same was not true for fellows (0%, strategic use; 3% substantive use). The only barrier that more than half said was relevant to the issue they explored in depth in the survey was that other decision factors were more important than the science (58%, Figure 3). Inconclusive evidence (30%) and biased information sources (28%) were cited by more than a quarter of the fellows as impediments. While bias is among the top three barriers for both staff and fellows, it is much less frequently cited by the latter.

Discussion

The portrait of congressional use of science by staffers and fellows that this study paints is nuanced. Science use is typical for energy, environment, and science issues, at least as background information and in office discussions. But the ways in which science is used, often in support of policy positions that have already been made (staff, 73% of issues; fellows, 68%), suggests that the authority of science is being brought to bear most frequently as a rhetorical tool, not unlike Whiteman’s findings decades ago. The ratio of substantive use of science (staff, 19%; fellows, 18%) is similar to that found by Whiteman (17%) among use in committees from 1979 to 1980 (Whiteman 1985), and aligns with the findings from Ouimet and colleagues’ systematic review (Ouimet et al, 2023).

Staff report both more barriers in substantive use than strategic use, and different ones, aligning with our hypothesised model. We anticipated that users, i.e. staff, would have to pay the ‘costs’ for acquiring new scientific information on these issues which are less likely to be politically polarised and for which fewer issue advocacy organisations facilitate information access. Indeed, lack of time, contacts, access, and difficult presentation are more challenging with substantive use for staffers. These barriers are quite similar to those identified in the UK Parliament (Kenny et al, 2017a)—credibility of the information source, availability and accessibility of the information, and time limitations—but they don’t align with those identified by Cash et al (2003), particularly salience and legitimacy. This suggests that the specific context relating to the institution and decision, as described in the model we present, may be important in understanding the dynamics of science communication in Congress.

Further, as hypothesised in the model, fellows said they faced fewer barriers in using science for policy than staff members, illustrating their fluency in scientific communication and success in transferring contacts and access within the institution. But fellows showed great similarity with staff in how they used science, reporting using science more often in support of policies that have already been decided by the office (strategic use) than in areas where a decision has not already been made (substantive use), in almost the exact same ratios as staff. This speaks to the limits imposed by the structure and culture of the institution itself, demonstrating both the strengths of expertise and its limitations to set norms and decision-making routines.

A new type of barrier not previously identified in other studies, including the recent review by Ouimet and colleagues (Ouimet et al, 2023), is that of fluency: the difficulties staff members face in communicating science. While training for scientists in science communication is increasingly de rigueur, similar programmes are rare for staff in Congress. The focus of Congress on communication over legislation further emphasises the importance of staff as science communicators.

Limitations

The scope of the study posed some acknowledged limitations. The rationale for focusing on staff members who work on science, energy, and natural resource portfolios was to ensure that science was relevant to the issues before them. However, the sample selection may also have influenced the relative rates of substantive to strategic use that we found, due to the partisan nature of the most common issue areas on which staff were working (energy and climate change). Further, because members participate in more than one committee, staff may do so as well. As a result, this study does not try to assess potential effects on staff behaviour from the many House and Senate committees, and their comparative composition, culture, and partisanship. In the future, researchers may wish to evaluate how these differences may influence personal staff science information needs and use.

Conclusion

The finding that strategic use of science predominates among both personal office staff and congressional science and engineering fellows aligns with other research demonstrating that members have become more interested in staking out partisan positions, and less interested in using information for problem solving (Lewallen et al, 2016), regardless of who advises them. Indeed, political polarisation is often faulted for the lack of evidence-based policymaking (Bogenschneider and Corbett, 2010). Differences in barriers to the use of science perceived by staff and fellows also point to structural factors within the institution that contribute to capacity, such as policies regarding personnel. Many have identified the need to increase the pay and retention of staff as a key component of modernising Congress to bring it into the information age (Graves and Schuman, 2020; Select Committee on Modernization of Congress, 2022). Remembering that the structure of the institution throws a long shadow may help scientists, and science organisations, to focus on science communication not just as a problem of building interpersonal connections (Crowley et al, 2021), but reshaping the broader policymaking environment.

Disclaimer

This paper is independent research that was supported by a subcontract with Harmonic International under Contract No. GS00F0003S, Order No. WC-133R-14-BA-0026/C-0008 to the National Oceanic and Atmospheric Administration (NOAA). The views expressed in this publication are those of the authors and not necessarily those of NOAA.

Funding

This research was supported by a subcontract with Harmonic International under Contract No. GS00F0003S, Order No. WC-133R-14-BA-0026/C-0008 to the NOAA Climate Program Office in the Science Related Communication Services for the Climate Program Office (CPO).

Research ethics statement

The study protocol was approved by George Mason University’s Institutional Review Board (1150984-1).

Contributor statement

KLA wrote the first and subsequent drafts of the manuscript, with comments from MCL, ETC, EH, SN, JH, and KMFT. KLA, MCL, ETC and EH conceptualised the study. KLA designed the study. KLA conducted data analysis and interpretation, with contributions from SN, JH, and KMFT.

Conflict of interest

The authors declare that there is no conflict of interest.