Professor Bhaven N. Sampat of the Mailman School of Public Health at Columbia University has recently distributed an article on “Determinants of Patent Quality: An Empirical Analysis.” In the paper, he analyzes a dataset of examiner and applicant prior art references from 502,687 utility patents issued between January 1, 2001 and December 31, 2003 in an effort to answer three questions:
Do examiners face particular difficulties in identifying certain types of prior art?
Do applicants’ incentives to search for prior art vary across industries?; and
Do applicants’ incentives to search for prior art vary across inventions? (p. 8)
His findings confirm the underlying assumptions behind the Community Patent Review project, namely that “patent examiners have a comparative disadvantage in searching for non-patent prior art and foreign patents, suggesting that all else equal [sic], patents are likely to be of lower quality for technological areas for which most prior art is not embodied in U.S. patents.” (p. 3)
Interestingly, while patent examiners account for 41% of the citations to previous U.S. patents, they account for only 10% of references to non-patent prior art. “If an applicant does not search for prior art and thereby does not report a piece of relevant prior art on his/her information disclosure statement, the examiner is less likely to discover it if it is codified in the non-patent literature or a foreign patent than if it is codified in a U.S. patent, since examiner capabilities for searching for U.S. patents exceed their capabilities for searching other sources of prior art” (p. 13).
This empirical study provides useful background and evidence to support the need for open, community-based review to dig up non-patent prior art, in particular.
A copy of the draft article can be downloaded here: http://siepr.stanford.edu/programs/SST_Seminars/patentquality_new.pdf_1.pdf
Comments