Returns of research funding are maximised in media visibility for excellent institutes

The study examines how institutional research excellence relates to public communication of science and whether competitive research funding brings added value to communication activities. Motivated by funding regimes emphasizing societal impact and public engagement requirements, the authors ask: (RQ1) Does public communication differ between institutes with differing levels of research excellence? (RQ2) How does research funding affect public communication in excellent versus less excellent institutions after accounting for organizational and contextual factors? The context includes evidence that prestigious institutions attract more funding and that inequalities in funding are substantial, potentially leading to expectations of greater communication activity in research-intensive institutes. However, institutional priorities, evaluation systems, and barriers such as lack of time and support may constrain outreach, particularly in highly research-intensive environments. The study uses excellence rankings as a proxy for research intensity, recognizing debates about such metrics, to analyze differences in communication activity and the role of funding across institute types.

Prior research indicates growing institutional engagement in public communication and resource allocation to such activities. Funding bodies increasingly require societal impact and public engagement plans in research proposals. Prestige and funding concentration (e.g., Matthew Effect) can advantage already well-funded institutions, raising questions about how excellence and funding shape communication outputs. At the individual level, more academically productive scientists tend to engage more in outreach. Institutional barriers include undervaluing communication relative to research and incentive systems focused on publication metrics. The literature suggests outreach can compete with research time, especially in institutions striving for top rankings. These strands inform hypotheses that excellent institutes may communicate more due to higher funding but may also face constraints; the study tests these dynamics at the institute level.

Design: Cross-national survey of research institutes examining public communication intensity and its drivers. Sample and data collection: N=1550 institutes from six countries—Germany (355), United Kingdom (172), Italy (351), Portugal (208), Brazil (149), and Japan (315). One questionnaire per institute, completed by unit directors or public communication staff in 2018 (Portuguese data from 2015 as part of a prior study). Sampling employed full populations in smaller countries and stratified probability samples in larger ones, balanced across OECD fields of science. Response rate was 25%. The dataset represents the research institute system in each country. Dependent variables: Three indices of intensity—public events (nine types), traditional media (thirteen channels including press interactions, newsletters, policy briefs, etc.), and new media (six channels: Facebook, Twitter, website, blogs, YouTube, Podcasts). Frequency responses were recoded to estimated annual counts: never=0; annually=1; quarterly=4; monthly=12; weekly=48; social media used 0,4,12,48, and 40 for daily. Indices sum activities per institute, producing estimated annual participation counts. Reliability: Cronbach’s alpha=0.76 (events), 0.86 (traditional media), 0.78 (new media). For driver analyses, continuous factor scores from confirmatory factor analysis (CFA) were used, showing good fit (χ2=627.54, df=142, p<0.001; CFI=0.96; TLI=0.95; RMSEA=0.04; BIC=56,474.49) with higher scores indicating more activity. Independent variables: Excellence in research—binary (1=excellent, 0=less-than-excellent) based on national evaluations: MIUR/ANVUR (Italy), REF 2015 (UK), FCT 2014 (Portugal), RUF 2016 (Brazil), DFG Excellence Initiative 2017 (Germany), NISTEP 2016 (Japan). Context (C) variables—country dummies (reference: Japan), field of science dummies per OECD schema (reference: Humanities), and size (ordinal counts of researchers: <20, 20–40, 41–60, >60). Research funding—ordinal: (1)<€250k, (2)€250–500k, (3)€500k–€1M, (4)>€1M; used as a proxy for competitive funding (average over prior 3 years). Disposition (D) variables—communication staff (count), communication funding (% of annual budget: none, <1%, 1–5%, 5–10%, >10%), communication policy (1=yes, 0=no), and percentage of active researchers engaging in outreach (ordinal recoded to midpoints: none, <10%, 10–20%, 20–40%, 40–60%, 60–100%). Analytic strategy: For RQ1, one-way ANOVAs compared excellent vs less excellent institutes on each intensity index (events, traditional media, new media). For RQ2, hierarchical linear regressions (IBM SPSS 26) were run separately for excellent and less excellent institutes for each dependent variable: Step 1 entered all C and D controls; Step 2 added research funding. Z-tests (Stats Tools Package macro) assessed cross-group differences in coefficients. Partial plots visualized funding effects controlling for covariates.

RQ1 (group differences): Excellent institutes reported significantly higher communication activity than less excellent across all domains: public events F(1,1548)=4.893, p<0.05; traditional media F(1,1548)=6.288, p<0.05; new media F(1,1548)=18.642, p<0.001. Estimated annual averages: public events—excellent 34 (median 21) vs less excellent 30 (median 19); traditional media—excellent 50 (median 29) vs 42 (median 21); new media—excellent 165 (median 60) vs 119 (median 32). RQ2 (drivers and funding effects):

• Context (C) factors: Country and size were robust predictors across models; field effects were mostly nonsignificant except Engineering and Technology, which showed negative associations for new media in both institute types and for traditional media among excellent institutes. Size positively predicted events and traditional media; effects on new media were not significant.
• Disposition (D) factors: Communication policy, active researchers, communication staff, and communication funding all positively predicted activity across domains and institute types, with policy and active researchers among the strongest contributors. Z-tests generally showed no significant differences in D-factor effects between excellence levels.
• Research funding effects: Adding funding (Step 2) improved model fit for all outcomes, especially among excellent institutes. Funding was positively associated with all three outcomes, with stronger effects on media channels than on public events. Unstandardized coefficients (Table 1, Model 2): • Public events—less excellent b=0.081 (SE=0.025), p<0.001; excellent b=0.104 (0.027), p<0.001. • Traditional media—less excellent b=0.089 (0.026), p<0.001; excellent b=0.163 (0.027), p<0.001. • New media—less excellent b=0.074 (0.024), p<0.01; excellent b=0.176 (0.028), p<0.001. Funding explained 5.7% of variance in traditional media among excellent institutes vs 1.8% among less excellent; for new media 6.4% vs 1.4%, respectively. Overall, Models 2 explained roughly 30–40% of variance across outcomes.
• Active researchers’ effect strengthened when funding entered the models, particularly for new media (becoming significant in Model 2), suggesting a link between funding and researcher participation.
• Funding availability: Excellent institutes had significantly higher funding levels (t(901.017)=-5.762, p<0.001; mean funding scale: excellent M=3.20, SD=1.623 vs less excellent M=2.61, SD=1.508), with differences most marked at extreme funding levels.
• Additional notable coefficients: Engineering & Technology showed negative coefficients for traditional and new media in excellent institutes (e.g., traditional media b=-0.326, p<0.01; new media b=-0.386, p<0.01). Several countries (e.g., Germany, Italy, UK, Brazil) had positive associations relative to Japan across domains.

The findings address RQ1 by showing that excellent institutes engage more in public communication across events, traditional media, and new media, with the strongest differences in media channels. Addressing RQ2, research funding contributes significantly to communication intensity beyond organizational and contextual factors, with effects markedly stronger in excellent institutes and concentrated in traditional and new media rather than public events. This pattern implies that funding yields higher media visibility returns for already research-intensive institutes, potentially due to their reputational status attracting media attention and encouraging investment in communication for profiling and visibility. The increased impact of active researchers when funding is included suggests that well-funded researchers may engage more publicly, aligning with prior evidence that more productive scientists do more outreach. These dynamics point to an intensified medialisation of science within excellent institutes and raise concerns about cumulative advantage (Matthew effect) in public visibility, whereby funding amplifies the prominence of already visible institutions. At the same time, the modest funding effects on public events indicate limited translation into dialogic public engagement, suggesting that communication may be instrumentally oriented toward visibility rather than participatory engagement. Institutions should reflect on communication goals and outcomes to ensure alignment with public engagement objectives rather than solely reputation-building.

This study demonstrates that research excellence is associated with higher public communication activity and that competitive research funding disproportionately enhances media-related communication in excellent institutes. Returns on funding are maximized in traditional and new media visibility rather than in public events. Institutional commitment—policies, researcher participation, staff, and dedicated communication funding—remains the strongest driver across institute types. The results imply that current funding and evaluation systems may reinforce visibility advantages for already excellent institutes. Future research should investigate the goals and content of institutional communications in funded contexts, the values and incentives shaping communication strategies, and the extent to which activities foster genuine public engagement versus instrumental visibility. Expanding models to include attitudes toward publics, engagement rationales, and qualitative assessments of communication practices could enhance explanatory power and inform policy.

• Explanatory scope: While models accounted for approximately 30–40% of variance, other unmeasured factors (e.g., views on publics, engagement goals, institutional values) likely influence communication and should be incorporated in future work.
• Response rate: The 25% response rate is typical for organizational online surveys but may introduce nonresponse bias.
• Measurement constraints: Excellence is proxied by national evaluations, and research funding is measured ordinally as institute-reported averages over three years, which may limit precision.
• Temporal heterogeneity: Portuguese data were collected in 2015 (earlier than others in 2018), which could introduce temporal inconsistencies.