Theory-Driven Statistics for the Digital Humanities: Presenting Pitfalls and a Practical Guide by the Example of the Reformation
Bibliographic record
Abstract
The Digital Humanities face the problem of multiple hypothesis testing: Evermore hypotheses are tested until a desired pattern has been found. This practice is prone to mistaking random patterns for real ones. Instead, we should reduce the number of hypothesis tests to only test meaningful ones. We address this problem by using theory to generate hypotheses for statistical models. We illustrate our approach with the example of the European Reformation, where we test a theory on the role of opinion leaders for the adoption of Protestantism with a logistic regression model. Given our specific setting, including choice of data and operationalisation of variables, we do not find enough evidence to claim that opinion leaders contributed via personal visits and letters to the adoption of Protestantism. To falsify or to support a theory, it has to be tested in different settings. Our presented approach helps the Digital Humanities bridge the gap between the qualitative and quantitative camp, advance understanding of structures resulting from human activity, and increase scientific credibility.
Fetched live from OpenAlex and de-inverted. Abstracts are not stored in this database: the inverted indexes are 8.6 GB of the frame’s 9.3 GB of text, and the host has 13 GB free.
How this classification was reachedexpand
Full frame machine prediction
Teacher imitationNot calibrated prevalence, not ground truth. Human validation pending. The Gemma side is a direct model label for every work in the frame, read from the title-only record. The Codex side is a classifier learned from the 10,348 direct Codex labels and calibrated to design-weighted sample rates; fields without enough sample support carry no Codex call. Candidate is the union of the two sides; consensus is their intersection. These outputs are machine_predicted_unvalidated and are not human labels.
Distilled classifier scores by category (both heads)
| Category | Codex | Gemma |
|---|---|---|
| Metaresearch | 0.100 | 0.336 |
| Meta-epidemiology (narrow) | 0.002 | 0.002 |
| Meta-epidemiology (broad) | 0.003 | 0.002 |
| Bibliometrics | 0.008 | 0.008 |
| Science and technology studies | 0.003 | 0.020 |
| Scholarly communication | 0.012 | 0.016 |
| Open science | 0.006 | 0.009 |
| Research integrity | 0.009 | 0.028 |
| Insufficient payload (model declined to judge) | 0.010 | 0.008 |
Machine scores (provisional)
The two teacher heads of the student model, read on this work. A score orders the frame for review; it never asserts a category, and the validation status ships verbatim with every row.
Baseline scores from an immature model (maturity gate not passed, 7 training rounds). Scores rank; they never assert a category.
score_only:v0-immature-baseline · verbatim from the scoring run: score_only means the number may rank works, and no category label ships from itClassification
machine, unvalidatedMachine predicted; a candidate call from one source (direct Gemma or distilled Codex), not a consensus.
How this classification was reached, model by model and score by score, is at the end of the page under "How this classification was reached".