Science and Climate Change Discourse on Dutch Twitter/X (2023): A Coded PDF Dataset

This dataset contains a systematically filtered and anonymized collection of Dutch-language Twitter/X posts from 2023 that explicitly reference science, scientific knowledge, or scientific authority in the context of climate change. The dataset is designed to support qualitative, quantitative, and mixed-methods research on climate communication, the sociology of scientific authority, and the use of religious and moral vocabularies in public discourse.

Data collection and corpus construction

The dataset was constructed using a two-step taxonomy-based retrieval and filtering procedure. First, a broad climate-related corpus was compiled by collecting Dutch-language Twitter/X posts containing the keyword klimaat (“climate”) over the calendar year 2023. This initial retrieval produced a maximal climate-relevant pool of approximately 698,000 posts, which serves as the background corpus for subsequent filtering.

Second, a science-focused taxonomy was applied to this climate corpus to extract tweets that explicitly foreground science rather than treating it as an implicit or background assumption. This filtering step yielded a subset of 56,696 tweets, which constitute the present dataset. The taxonomy groups Dutch-language keywords into four functional categories:

1. Scientific discourse and authority (e.g. wetenschap, consensus, peer reviewed);

2. Scientific findings and indicators (e.g. CO2-concentraties, zeespiegelstijging);

3. Scientific institutions and actors (e.g. universiteiten, klimaatwetenschappers, IPCC);

4. Challenges to science and counter-science vocabularies (e.g. pseudowetenschap, klimaatontkenning).

This taxonomy functions as an explicit filtering mechanism, ensuring that all included tweets engage directly with science as an object of evaluation, authority, or contestation.

Data format and structure

The dataset is provided as a PDF file in which tweets are presented in coded and sequential form. Each entry contains the original Dutch-language tweet text alongside an internal reference code used for analytical traceability. The PDF format is intentionally chosen to balance transparency and ethical responsibility, allowing close reading and citation while reducing risks of large-scale redistribution or re-identification.

Anonymization and ethical considerations

All tweets in the dataset have been fully anonymized. Usernames, profile information, direct links to accounts, and other personal identifiers have been removed. Only the textual content of tweets is retained. The dataset includes exclusively publicly available posts and follows established ethical guidelines for social media research. The dataset is intended for scholarly analysis rather than for profiling individual users.

Analytical relevance

The dataset underpins a mixed-methods research design that combines distributional analysis (e.g. keyword frequency, hashtag patterns, sentiment signals) with qualitative thematic and rhetorical analysis. It is particularly suited for examining how science is constructed, sacralized, or delegitimized in public climate discourse. In the associated research, the dataset is used to identify two recurring frames:
(1) science as a substitute for religion, where scientific authority is invoked to authorize moral imperatives and collective obligation; and
(2) science as a false religion, where scientific institutions and consensus are recoded as dogma, belief, or coercive orthodoxy.

Scope and limitations

The dataset reflects discourse on Twitter/X in Dutch during 2023 and does not claim to represent the Dutch population as a whole. Social media debates tend to amplify polarized positions, and the focus on science-related keywords may over-represent moments of controversy and authority conflict. Nevertheless, the dataset provides a robust empirical window into how scientific authority is publicly negotiated under conditions of uncertainty, polarization, and moral urgency.

Reuse and citation

The dataset is released for research and educational purposes under an open license. Users are encouraged to cite the dataset when using it for secondary analysis, replication, or comparative research on climate discourse, science communication, religion-related framing, or digital sociology.