Replication Data for \"Ladder or trap_how incremental innovation spurs breakthroughs in AI technology\"

This dataset contains all materials and data supporting the findings of the associated manuscript. It includes: (1) The full sample data (“[Data]LCPtoAI_arcsinh.dta”) is used to generate Tables 3–6. (2) The scripts in this dataset do not include the categorization method. However, using the full sample data and applying Equation (11) from the manuscript, one can easily compute the relative ratio of LCPs to HCPs (LCP ratio) for categorization. (3) The categorized data with tags (“_HR”, “_HR no CHN”, “_LR”, “_LR add CHN”) are used to generate Tables 8–11. (4) For simplicity, categorized data for other nations/regions are not included in this dataset. Repetitive analyses can be performed using the full sample data and applying the same grouping method as for the CHN sample, as shown in the manuscript.