Systematic review maps how researchers build scenarios of possible futures

A systematic review of academic literature published between 2014 and 2023 has mapped, in unusual detail, the methods researchers actually use when they construct futures scenarios, the planning tools that explore alternative futures rather than predicting a single outcome. The review, published in the European Journal of Futures Research, analyzed 115 articles indexed in the Scopus and Web of Science databases and found a field dominated by qualitative, participatory approaches, but one that is steadily absorbing quantitative techniques, computational models, and early applications of artificial intelligence.

The study was conducted by Pedro Leon Cruz Aguilar, Javier Enrique Medina Vasquez, and Olga Patricia Marin Arroyave, who applied the PRISMA protocol for systematic reviews. Their search string combined the keywords foresight, future, and scenarios across the two databases, restricted to articles and conference papers published from 2014 through 2023. The initial retrieval produced 1,320 records, 610 from Scopus and 710 from Web of Science. After removing 409 duplicates with a reference manager and screening out 796 papers on the basis of titles and abstracts, 115 articles remained for full systematization in a structured extraction matrix.

The headline finding concerns which methods dominate. Environmental scanning in its PESTEL variant, participatory workshops, the Delphi method, structural analysis using the MICMAC software, and Causal Layered Analysis emerged as the most frequently used techniques. Four dominant methods appeared in 49 percent of the analyzed exercises, a pattern the authors read as evidence of the enduring strength of qualitative and participatory methodologies. These approaches favor narrative construction of alternative futures over formal modeling, treating scenarios as coherent stories about how the future might unfold rather than as forecasts.

Yet the review also documents a clear movement toward combination. The authors identified seven recurring patterns of methodological pairing, ranked by how often the constituent methods co-occur in the same article. The most common is the intuitive logics pattern, which chains literature review, environmental scanning, identification of drivers of change, the classic two-by-two matrix, and scenario storylines. A second pattern follows the French School, combining morphological analysis, actor-strategy analysis, and the Regnier Abacus, typically supported by software such as MICMAC, MACTOR, and Morphol to reduce subjectivity in strategic decisions.

The remaining five patterns extend the repertoire considerably. Expert consultation with quantitative triangulation pairs Delphi panels and focus groups with statistical summaries such as medians, quartiles, and percentages, alongside documentary and bibliometric checks. Multi-level critical foresight integrates Causal Layered Analysis, the Futures Triangle, and the six pillars approach to deconstruct the metaphors and worldviews that quietly steer decision-making. An implementation-oriented pattern brings together backcasting, technology roadmapping, flex points, and operational scenarios to connect anticipation with concrete action. A computational pattern applies bibliometrics, text mining, system dynamics, Monte Carlo simulation, and spatial modeling to large textual, numerical, or georeferenced datasets, while a participatory, experiential pattern uses serious games, LEGO Serious Play, card games, and landscape simulations to build shared visions through play and collective meaning-making.

To describe this combinatorial behavior, the authors propose a new concept: methodological aggregation capacity, defined as the ability of a scenario exercise to incorporate techniques of different epistemic natures in a coherent, functionally integrated way while preserving internal consistency and plausible results. The authors position aggregation as an intermediate level of coupling, sitting between the mere coexistence of independent techniques and full integration into a single formal framework. Unlike triangulation, which seeks validation through convergence, aggregation distributes complementary functions among methods organized around an anchor technique. In their reading, this structuring is what allows the field to absorb new tools without losing coherence or strategic purpose.

One striking negative result: not a single article in the corpus relied exclusively on quantitative methods. Of the 115 exercises, 48, or 42 percent, used only qualitative methods; 40, or 35 percent, used quantitative and semiquantitative methods; and 27, or 23 percent, combined qualitative and quantitative work. The authors interpret this absence as an implicit consensus among practitioners, namely that quantitative data gains strategic value only when embedded in a narrative framework that makes it socially meaningful and actionable. Computational approaches remained marginal overall, accounting for 7.8 percent of the corpus, and the review found no sustained upward trajectory in their use during the period studied.

The review also charts where scenario exercises are being applied. Roughly 70 percent of the analyzed cases concentrate in one cross-cutting process and six sectors, while the remaining 30 percent spread across additional domains. The authors report that scenarios have migrated beyond large corporations and the economic sphere into public health, higher education, governance, environmental sustainability, agriculture, and territorial development. This sectoral expansion, they argue, is turning scenario methods into a public methodological tool available to communities and organizations of all kinds, though it also raises a capability challenge: not every adopting sector has technical teams able to guarantee the coherence, consistency, and plausibility of the resulting scenario statements.

Artificial intelligence marks the temporal boundary of the corpus. Of the 115 articles, 88, or 76.5 percent, were published between 2014 and 2021, and 97, or 84.3 percent, appeared before 2023, indicating that AI-supported techniques such as data mining, text mining, and computational modeling still had only an incipient presence in the scenario literature. The authors argue that AI integration will let analysts process volumes of textual and numerical data that would overwhelm human experts, surfacing weak signals that cognitive biases and procedural limits might otherwise hide. Crucially, the reviewed exercises treat such tools as complementary to human judgment rather than as substitutes, improving the quality of inputs without displacing collective deliberation or narrative construction.

The authors acknowledge limitations that shape how the findings should be read. The search equation prioritized English-language outputs containing the term foresight, which may have underrepresented Latin American and Francophone contributions. Method names were not standardized during data extraction, so each of the 115 articles registered a unique methodological composition, and the number of distinct methods kept rising as the corpus grew, meaning the review reached saturation across sectors but not across methods. The authors conclude that the maturity of the field rests on the systemic convergence of data-driven rigor and narrative depth, and they frame the emerging skill in foresight as a shift from manual technique application toward the design and critical curation of AI-assisted analyses, with expert judgment retained as the certifying authority for plausibility.

Subject of Research: Which foresight methods researchers use to formulate futures scenarios, based on a systematic review of 115 articles from 2014 to 2023

Article Title: A systematic review of foresight methods used in the formulation of futures scenarios: evidence from articles indexed in scopus and web of science (2014–2023)

Article References: Aguilar, P. L. C., Vasquez, J. E. M., & Arroyave, O. P. M. (2026). A systematic review of foresight methods used in the formulation of futures scenarios: evidence from articles indexed in scopus and web of science (2014–2023). European Journal of Futures Research, 14(1), Article 35. https://doi.org/10.1186/s40309-026-00298-4

Image Credits: AI Generated

DOI: 10.1186/s40309-026-00298-4

Keywords: foresight, futures scenarios, systematic review, PRISMA, Delphi method, PESTEL, Causal Layered Analysis, MICMAC, backcasting, artificial intelligence, methodological triangulation, strategic planning