Learning When to Quit: An Empirical Model of Experimentation
Bibliographic record
Abstract
Research productivity depends on the ability to discern whether an idea is promising, and a willingness to abandon the ones that are not. Economists know little about this process, however, because empirical studies of innovation typically begin with a sample of issued patents or published papers that were already selected from a pool of promising ideas. This paper unpacks the idea selection process using a unique dataset from the Internet Engineering Task Force (IETF), a voluntary organization that develops protocols for managing Internet infrastructure. For a large sample of IETF proposals, we observe a sequence of decisions to either revise, publish, or abandon the underlying idea, along with changes to the proposal and the demographics of the author team. Using these data, we provide a descriptive analysis of how R&D is conducted within the IETF, and estimate a dynamic discrete choice model whose key parameters measure the speed at which author teams learn whether they have a good (i.e., publishable) idea. The estimates imply that sixty percent of IETF proposals are publishable, but only one-third of the good ideas survive the review process. Author experience and increased attention from the IETF community are associated with faster learning. Finally, we simulate two counterfactual innovation policies: an R&D subsidy and a publication-prize. Subsidies have a larger impact on research output, though prizes perform better when accounting for researchers' opportunity costs.
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 | 0.022 | 0.103 |
| Meta-epidemiology (narrow) | 0.001 | 0.001 |
| Meta-epidemiology (broad) | 0.003 | 0.002 |
| Bibliometrics | 0.002 | 0.002 |
| Science and technology studies | 0.002 | 0.005 |
| Scholarly communication | 0.005 | 0.009 |
| Open science | 0.004 | 0.002 |
| Research integrity | 0.007 | 0.005 |
| Insufficient payload (model declined to judge) | 0.026 | 0.003 |
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".