Science-Policy-Interface

Abstract

The Science-Policy-Interface (SPI) construct captures the mediated interaction between science and politics as represented in news coverage. SPIs are conceptualized as interactions between the scientific and political system that occur across varying degrees of institutionalization and publicness, and on micro (individual actors), meso (organizational), or macro (systemic) levels. The construct integrates two dimensions that prior research has largely treated separately: scientific policy advice and science policy. Scientific policy advice covers instances in which scientific actors provide knowledge as a resource for political decision-making — either in an epistocratic mode (deriving direct political imperatives from scientific findings) or as an honest broker (presenting policy options while acknowledging uncertainty and the limits of scientific authority). Science policy refers to instances in which political actors employ political resources — opportunity structures, financial incentives, or governance mechanisms — to steer, fund, or regulate the scientific system. Beyond the internal structure of the SPI, the construct also captures the journalistic role in the interaction: media coverage may function as an interaction mediator (reporting on a pre-existing interaction) or as an interaction initiator (bringing scientific and political actors into relation through the coverage itself, without a prior real-world interaction). Field of Application/Theoretical Foundation The SPI construct is rooted in the sociology of science and science communication research. Theoretically, the work draws on: Systems theory (Luhmann, 2018; Schimank, 2012): science and politics as functionally differentiated systems with structural coupling and mutual dependencies Science-policy interface literature: normative case-study tradition (Cash et al., 2003; van den Hove, 2007; Watson, 2005) complemented by the analytic typologies of Pielke (2007) on scientists' communication roles (honest broker, epistocrat), and Pregernig (2014) on epistocratic vs. decisionist perspectives Mediatization theory (Esser & Strömbäck, 2014; Blöbaum et al., 2013): both science and politics are mediatized, meaning they adapt their logics and behavior to media demands Science journalism research: distinctions between a narrow understanding of science journalism (reporting new findings) and a broad understanding (contextualizing socio-political events with scientific expertise; Summ & Volpers, 2016) Politicization of science research (Druckman, 2022; Schmid-Petri et al., 2022; Brück et al., 2025): the present construct goes beyond politicization as a specific process and treats the science–policy relationship as a multidimensional, bidirectional phenomenon The construct applies to content analyses of political and science journalism, science communication research, and media framing studies at the intersection of political communication and science communication. Combination with other Methods The study combines the manual content analysis with: Structural Topic Modeling (STM) for the inductive identification of topic areas covered in SPI news coverage (k = 11; implemented in R with the stm package); topics are assigned to coded articles as an additional variable and serve as predictors in subsequent regression analyses Logistic regression (glm function (stats package) in R) with dummy-coded cluster membership as the dependent variable, to examine predictors of SPI coverage patterns (H1–H7) Cluster analysis (Ward method, Euclidean distance, k = 4) for the exploratory identification of coverage patterns across a theoretically motivated subset of variables Example Studies While previous content analyses have examined either science journalism broadly (e.g., Summ & Volpers, 2016; Leidecker-Sandmann & Lehmkuhl, 2022) or the politicization of science (Schmid-Petri et al., 2022) in media coverage specifically (Brück et al., 2024; Schmidt, 2023), no study has yet operationalized the full bidirectional interaction between science and politics as represented in news coverage. The present operationalization of the SPI builds on these approaches — in particular on Brück et al. (2024), whose multidimensional measure captures overlapping dimensions such as dependencies between science and politics (political and scientific resources) and politicized actor structures — while extending them in several respects. First, the present instrument integrates both directions of the science-policy relationship: not only the provision of scientific knowledge as a resource for political decision-making (scientific policy advice), but also the deployment of political resources to govern and fund science (science policy). Second, it adds constructs specific to the structural properties of the SPI, including the degree of institutionalization and publicness of the interaction. Third, it captures the communication modes of scientific actors (epistocrat vs. honest broker; Post & Bienzeisler, 2024) and the communicative acts of both scientific and political actors. Fourth, it explicitly accounts for the journalistic role in the interaction, distinguishing between media coverage as an interaction mediator — reporting on a pre-existing science-policy interaction — and as an interaction initiator, in which the coverage itself constitutes the interaction. Information on Brück et al. (2025) Authors: Janise Brück, Julia Serong, & Lars Guenther Research question: How can the politicization of science in and through media coverage be theoretically conceptualized and empirically operationalized? In a pilot study, the theoretical concept was partially tested by asking: To what extent has science been politicized in German COVID-19 media coverage? Object of analysis: 262 science-related online media articles about the COVID-19 pandemic from three German quality outlets (Frankfurter Allgemeine Zeitung, Süddeutsche Zeitung, Die Zeit), collected as an artificial week across two pandemic waves in 2020 (t1: March 2nd–April 19th; t2: August 17th–October 4th). Variables: The politicization of science in media coverage was measured using three overarching indicators: (a) thematic blurring of politics and science — including the presence of political perspectives, story origin, and dependencies between science and politics (political resources such as commissions and funding; scientific resources such as policy advice); (b) politicized actor structure — mentions and citations of political actors, scientific actors, and scientific administrations; (c) politicized news values — emphasis on different types of science-related uncertainty (deficient, technical, consensus, scientific) with or without political (de)legitimization. Level of analysis: news article. Reliability: Intra-coder reliability was assessed for approximately 15% of the sample (Holsti's CR > .81; Krippendorff's α > .72). Codebook: Brück (2025); https://doi.org/10.17605/OSF.IO/83XKY Information on Zils (2026) Research question: (FF1) Which thematic fields and science areas are covered in relation to SPIs, and how do these differ over time and across outlets? (FF2) Which interaction dimensions (science policy/ scientific policy advice) are represented in coverage, and how do these differ over time and between tabloid and quality press? (FF3) Which structural properties of SPIs — including the trigger of coverage and the journalistic role as mediator or initiator — are represented, and how do these differ over time and across outlets? (FF4) How are scientific and political actors communicatively addressed and represented? (FF5) Which patterns of SPI coverage characterize German news reporting? Object of analysis: Print editions of three German national daily newspapers — Süddeutsche Zeitung (SZ), Frankfurter Allgemeine Zeitung (FAZ), and BILD — covering the period January 1, 2019–December 31, 2023. Articles were identified via a broad keyword search followed by manual relevance screening based on a predefined access criterion (presence of an SPI). A validation study comparing specific and broad search strings confirmed that a broad string with manual post-selection yielded superior recall. Final sample: N = 1,343 relevant articles (SZ: n = 340, FAZ: n = 402, BILD: n = 601) out of approximately 5,400 screened articles. Level of analysis: News article. Table 1: Variable/Construct overview Dimension Variable Coding Level of Analysis Relevance Filter SPI present Dichotomous Article Formal Outlet Nominal Article Date Metric Article Word Count Metric Article Journalistic format (descriptive vs. opinion) Nominal Article Thematic Science field according German Research Foundation (5 dichotomous variables) Dichotomous   Article Topic Determined via structural topic model Article Trigger of coverage (scientific / political / SPI itself / not identifiable)   Dichotomous   Article SPI structure Institutionalization: 0/1; Publicness: 0/1 → 4 interaction dimensions Dichotomous Article Interaction type: science policy/ scientific policy advice Dichotomous Article Communication mode of scientist: epistocrat/ honest broker Dichotomous Article Political resources: opportunity structures, financial incentives, governance Dichotomous Article Actor level science/politics :1 = micro, 2 = meso, 3 = macro Ordinal Article Hierarchy level science/politics: 1 = no quotes, 2 = direct/indirect quotes, 3 = interview, 4 = guest article Ordinal Article Communicative acts science/politics: criticize, appreciate, explain, demand (4 dichotomous variables each) Dichotomous Article Substantiating statements: empirical results cited, method mentioned (2 × 2 dichotomous variables for science and politics separately) Dichotomous Article Addressees of science/politics: population, economy, politics, science (4 × 2 dichotomous variabl