Inventor Triad Dynamics and Invention Quality Dataset
Variables for regression of the study on the relation between inventor triad dynamics and invention quality
Steps to reproduce
Our primary data source is the National Bureau of Economic Research (NBER) patent dataset (Hall, Jaffe, and Trajtenberg 2001) complemented by two other data sources: Harvard Patent Network Dataverse and WRDS Compustat database. NBER patent data includes all U.S. utility patents granted during 1963-2006. Harvard Patent Data offers the cleaned and disambiguated inventor names, and WRDS Compustat database includes firm attributes. Our sample consists of U.S. biotechnology utility patents that were granted between 1987 and 2003. We aim to create a sample of open triad relations that subsequently become closed triad relations (treatment group) as well as a comparable sample of open triad relations that did not become closed triads (control group). This design identifies the treatment effect of the treatment. In our study, it allows us to test whether inventors in the open triad experience the predicted change in their invention quality when they are involved in a triadic closure process. By merging the treated and control groups, we obtained a dataset of 26,486 cases.