HN archives are stored in BigQuery, so it would be possible to do an analysis that was more rigorous than a human reading the thousands of front pages to see a pattern of difference.