The Royal Statistical Society (RSS) Discussion Meetings Committee (DMC), along with the RSS Medical Section, invites submissions of discussion papers on ‘Advances in evidence synthesis: integrating granular and aggregate data’.
Papers selected for publication will form the basis of a programme of RSS events on evidence synthesis and data integration. The programme will begin with a Discussion Meeting at the RSS International Conference in September 2028 and continue with dedicated Discussion Meetings and related activities through August 2029. Accepted papers will be published in the most appropriate series of the Journal of the Royal Statistical Society (JRSS), together with the associated discussion contributions and authors’ responses.
Proposal Abstract
Evidence synthesis combines information from multiple data sources to estimate robust effects and is a cornerstone of reliable statistical modelling. Models like model-based meta-analysis and multilevel network meta-regression are well established in health and epidemiology, but the challenge of combining individual-level data with aggregate, population-level statistics arises in many disciplines. Related problems are addressed through approaches such as spatial downscaling in geography and ecological inference in political science.
Across these fields, researchers face complex theoretical and computational challenges. These include modelling complex hierarchical and dependent data structures, adjusting for differences between populations, and transporting estimates to new target populations. The challenges become more difficult when populations or underlying relationships change over time, data are missing through complex mechanisms, or models must extrapolate beyond well-represented regions of the sample space.
To encourage exchange between disciplines that have often developed methods separately, the RSS invites proposals for Discussion Papers on the methodology, theory and application of evidence synthesis and data integration. Submissions should address questions of broad statistical interest and have clear potential to stimulate discussion across disciplines and application areas.
The aim is to advance the statistical foundations, theory and applications in evidence synthesis involving individual-level and aggregate data, while encouraging exchange between disciplines in which related approaches have often developed separately.
Topics of interest include (but are not limited to):
-
Compatibility, identifiability and conflicting evidence: Establishing the conditions under which data from different sources can identify the target quantity of interest; assessing whether populations, variables, measurements and assumptions are sufficiently compatible; and developing methods to detect, explain and accommodate disagreement between sources.
-
Cross-disciplinary methodological translation: Bridging synthesis techniques used in health and epidemiology, such as model-based meta-analysis and multilevel network meta-regression, with related data-integration approaches developed in geography, political science, economics, and other fields, including spatial downscaling, ecological inference, multilevel regression and poststratification, and g-computation.
-
Population rebalancing, transportability and covariate adjustment: Developing statistical methods for aligning disparate cohorts and transporting or generalising estimates to new target populations, including settings in which covariate information or overlap between populations is limited.
-
Temporal dynamics and distributional drift: Developing methods to synthesise evidence across data sources when populations, measurement processes, baseline risks or underlying relationships evolve over time.
-
Missing data and extrapolation: Addressing complex missing-data mechanisms when combining heterogeneous datasets and quantifying uncertainty when extrapolating into sparsely represented regions of the sample space. This may include sensitivity analysis for selection bias, measurement error, unmeasured confounding and model misspecification.
-
Simulation frameworks and synthetic data: Developing realistic data-generating mechanisms to benchmark synthesis methods, create synthetic populations, evaluate bias and uncertainty under complex correlation structures, and identify the settings in which methods perform well or fail.
Proposals will be assessed for their relevance to the call, scientific quality, breadth of appeal and potential to stimulate discussion across statistical disciplines and application areas. All full-paper submissions will undergo the standard editorial assessment and peer-review process for RSS Discussion Papers, with any papers submitted to JRSS Series B being screened and handled by the Research Section. A full paper that is judged suitable for publication but not for inclusion in the Discussion Paper programme may, with the agreement of the authors and the relevant journal editors, be considered as a regular journal submission.
Full papers must not exceed 16 pages, excluding supplementary material, and must follow the standard JRSS formatting guidelines.
Submissions will follow a two-stage process: initial consideration of a short abstract, followed by editorial assessment and peer review of invited full papers.
The submission and review process will be as follows:
1. Abstract submission. Authors are invited to submit a one‐page abstract of no more than 400 words outlining the proposed Discussion Paper to journal@rss.org.uk by 15th November 2026.
2. Invitation to submit a full paper. Authors of selected proposals will be notified by 31st November 2026 and invited to submit a full paper for consideration. An invitation to submit a full paper does not guarantee acceptance for publication.
3. Full-paper submission. Full papers should be no more than 16 pages, excluding supplementary material, and prepared according to the standard JRSS formatting guidelines. Papers should be submitted via Manuscript Central to the most appropriate JRSS series (A, B or C) selecting the “Discussion Paper” option. The deadline for full paper submission is 15th April 2027.
4. Editorial assessment and peer review. All full-paper submissions will undergo peer review according to the Society’s standard criteria for discussion meeting papers, considering both scientific quality and potential for discussion.
5. Pre-print circulation. Final versions of accepted papers are expected to be ready for circulation as preprints in advance of the relevant events in the 2028/29 programme.
6. Participation in the programme. Authors of accepted papers will be expected to present their work at an event within the programme of RSS activities on evidence synthesis and data integration. At least one author should therefore be available to present the paper and participate in the discussion. Further details of the programme, including any activities held as part of the RSS International Conference, will be provided in due course.
Informal enquiries about the call are welcome and may be directed to the Discussion Papers Editor, Ben Swallow, bts3@st-andrews.ac.uk.