SciPost Submission Page
Reanalysis of "Competition and Innovation: An Inverted-U Relationship"
by Andrew Gelman
Submission summary
| Authors (as registered SciPost users): | Andrew Gelman |
| Submission information | |
|---|---|
| Preprint Link: | scipost_202510_00058v3 (pdf) |
| Date submitted: | Dec. 16, 2025, 1:48 a.m. |
| Submitted by: | Andrew Gelman |
| Submitted to: | Journal of Robustness Reports |
| Ontological classification | |
|---|---|
| Academic field: | Multidisciplinary |
| Specialties: |
|
| Approach: | Computational |
Abstract
The inverted U held up in reanalysis, but the result is sensitive to the assumption of a common curve for all industries.
Author comments upon resubmission
Thank you for the comments. I revised the paper accordingly.
List of changes
I revised the paper as follows: 1. I explained the way in which the quadratic and nonparametric curves in the original paper do not fit the data. 2. I changed the log(1+y) model to sqrt(y), a standard variance-stabilization transformation for count data. 3. I reduced the discussion of the measurement issues and made it clear that I do not address them in my reanalysis. 4. I have emphasized that the result relies on an assumption that was in the original paper and also in my reanalysis, an assumption of a common curve across industries. 5. I have given more details about the priors in the Bayesian model.
Current status:
Refereeing in preparation