CoChat

The DOI matches its Crossref record; 18 of the 18 references it cites were found in a registry or an index; none of the 70 sources its claims cite is checked for support yet.

Data storytelling is not storytelling with data: A framework for storytelling in science communication and data journalism

Open access
Matei, Sorin Adam, Hunter, Lucas · 2021 · The Information Society

Abstract

Storytelling teaches. All good storytelling is good teaching. Storytelling is a form of teaching and learning because it asks the readers or listeners to replace well-established explanations that are considered facts with new, unexpected ones. We always learn new things when listening to good stories. Good stories do so by violating expectations and surprising the listener or the reader. Surprise breeds suspense, which generates engagement, which is a catalyst for learning – even truer for scientific and data-driven stories. In this paper we show how understanding the nature of good stories by focusing on the novelty they introduce and assumption they violate helps us to do effective scientific work and tell excellent data-based stories. We start by providing definitions of “story,” “storytelling,” and “good stories.” We then outline a methodology for building stories and provide an illustrative example of an effective data-based story from the history of medicine.

Overview

CoChat Summary

TL;DR

This conceptual essay argues that “good” stories—including scientific and data stories—violate expectations by replacing a known causal explanation with an unexpected one, and that data storytelling is not merely adding data to a narrative. It proposes a framework based on expectation violation, Davis’s taxonomy of interesting research, and causal-inference rules, illustrated through the history of cholesterol/heart-disease research and astronomy.

Key findings

  • A good story connects an unexpected cause to a known effect, or a known cause to an unexpected effect; surprise drives curiosity, engagement, and learning.
  • The paper adapts Murray Davis’s strategies for generating interesting research, including violating assumptions about organization, composition, abstraction, generalization, stabilization, function, evaluation, correlation, co-existence, opposition, and causal direction.
  • Data storytelling is defined as narrative that upends public assumptions about a causal relationship using quantitative causal-inference rules: cause must precede effect, third factors must be controlled, and effects should be large and significant.
  • In the cholesterol example, the paper reports that Finnish participants died of heart attacks at a rate 13 times higher than Japanese participants; it also uses Castelli’s Framingham data to show that total cholesterol can mislead—e.g., corner B has total cholesterol 185 with under 5% risk, while corner D has total cholesterol 125 with over 15% risk. Increasing HDL from low to high decreases risk from 20 to under 5, while lowering LDL from high to low decreases risk from about 15 to 10.

Methods

  • Design: conceptual/theoretical essay, not an empirical study.
  • Data or sample: no original dataset; the argument draws on historical and illustrative examples, especially the history of heart-disease/cholesterol research (Keys, Framingham, Castelli) and astronomy/gravity (Copernicus, Galileo, Newton).
  • Main analysis: argumentative synthesis of narrative theory, Murray Davis’s taxonomy of expectation-violating research strategies, and causal-inference criteria for data storytelling.

Limitations

  • The paper states that opposing assumptions and identifying a strong causal mechanism does not guarantee success, and that stories defying theoretical assumptions and common thinking may struggle to enter the zeitgeist.
  • It also states that what works for the public may not work for experts, and that causal relationships are hard to prove because correlation is not causation.
  • Evident from the text: the framework is not empirically tested, the examples are selectively illustrative rather than systematically sampled, and no data collection or formal validation of the proposed storytelling method is reported.

Model-written summary. Check it against the paper before relying on it.

Integrity signals

Registry

Crossref · 10.1080/01972243.2021.1951415

Cites

18 of 18 references verified0 citation errors · 0 not checkable

Statistics

Nothing to recompute.

Verification

Claims the paper makes

0 Valid · 6 Not valid · 44 Not checked · 68 Own results and interpretations

  • Probably not a substantial claim. 112
  • It may state two separate claims as one. 30

Create a free account to see these claims.

Self-check

Percentages, sums and reported differences re-computed from the paper’s own numbers.

Each contradiction between two passages listed here was confirmed by an independent check.

2 of 2

Citations

Cites

Show all 18 references

Research works printed but not linked: 2

  1. Banister F., 2001, School Science Review, 83, 75
  2. Molière, 1998, Tartuffe and The Bourgeois Gentleman

Cited by

See all 61

Metadata

Authors
Matei, S. A., & Hunter, L.
Affiliations
Purdue University West Lafayette (US)
Corresponding author
Matei, Sorin Adam
Venue
The Information Society
Published
2021-08-06
Volume
37
Issue
5
Pages
312-322
ISSN
  • 0197-2243 (ISSN-L)
  • 1087-6537
OpenAlex
W3190635930
Cited by
47
References
17
Citations by year
2026: 11 · 2025: 11 · 2024: 8 · 2023: 5 · 2022: 10 · 2021: 2
FWCI
2.25
Type
journal article
Language
en
Open access
Locations
The Information Society (publishedVersion, cc-by-nc-nd)
Funders
Purdue University Integrated Data Science Initiative
Topics
  • Physical Sciences › Computer Science › Computer Graphics and Computer-Aided Design › Data Visualization and Analytics
  • Social Sciences › Library and Information Sciences › Academic Publishing and Open Access
  • Social Sciences › Communication › Media Influence and Health
Keywords
data storytelling, storytelling, science communication, data journalism, engagement, history of medicine, narrative framework

About these checks

Registry, retraction, Cites and statistics results are computed, not judged, with one exception: a printed reference that no index lists is read once by a model before it counts as a citation error. Claims are extracted by a model and kept only when their quote appears verbatim in the text; each cited source is then checked, with a model’s help, for whether it says what the claim relies on. Those verdicts can be wrong — read the claim’s quote and the cited source before relying on one.

A language model drew the claims from the paper on 2026-11-11, under version 3 of the claim definition (model deepseek-v4.1-flash, prompt extract-v3).

A second model labelled each claim on 2026-11-11 (model jev-1.13-20260917, definition revision gold-guideline-v0.2).

Statistics and arithmetic are recomputed, not judged.

Confidence bands are shown only for fields calibrated on our annotated test set. None is calibrated yet, so no band is shown.

What counts as a claim ›How we measure confidence ›How we check this ›

Are you an author, or do you see a problem with this page? Request a review

How to cite

Cite CoChat for the checks on this page, and the paper for its own findings.

This page
CoChat. (2026). Citation check: Data storytelling is not storytelling with data: A framework for storytelling in science communication and data journalism. https://app.cochat.ai/papers/10.1080/01972243.2021.1951415
The paper
Matei, S. A., & Hunter, L. (2021). Data storytelling is not storytelling with data: A framework for storytelling in science communication and data journalism. The Information Society. https://doi.org/10.1080/01972243.2021.1951415

Data sources

Bibliographic data: Semantic Scholar, OpenAlex, Crossref.

Contains information from the Semantic Scholar Academic Graph Datasets, made available under the ODC Attribution License (ODC-BY 1.0).

Semantic Scholar’s data is described in: Kinney, R., et al. (2023). The Semantic Scholar Open Data Platform. arXiv:2301.10140.