Using Distance Measures to Operationalize Patent Originality

A distance-score methodology (the V-score) for constructing firms' backward-dispersion patent-citation scores improves upon the much-used counting methodology of NBER studies because (a) it incorporates the technology-class codes of prior-art citations that are unlike the technology-class codes of a focal patent's grant in a different way, and (b) it adjusts for dynamics in technological convergence that have occurred over time. When comparing results using the V-score with scores obtained for the same sample by using the counting methodology, we found that the two methodologies produced differently-signed coefficients with differing levels of significance. Backward-dispersion patent-citation scores that were calculated using the counting methodology were more likely to be negatively correlated with firms' performance measures; co-efficients for the distance-score measures were positively-correlated.