Patent citations are references considered important by the patent examiner in determining the novelty and inventive step of a patent application. References made to previous patents are called backward citations, and references received from future patents are called forward citations. Cumulative and knowledge transfer indicators use backward citations, and impact-type indicators are based on forward citations (OECD, 2009). The history of citations in patents dates back to 1947 when examiners in United States Patent and Trademarks Office (USPTO) began citing references considered during the examination process (List, 2010). Researchers have a long history of using citation based indicators. This has helped identify many pitfalls in the use of citation indicators that are addressed in the data analysis pertaining to the thesis. Some of the key issues related to citation analysis are presented below.
Country and industry variations
Citation practices differ significantly across patent offices and across technology fields.
For example, US patents contain three times more references as EPO patents (Michel and Bettels, 2001). This is because European guidelines require examiners to list only the most important references. As a result, the jurisdiction for analysis has been restricted to EPO for most of the studies related to this dissertation.
Right truncation
Forward citation analysis suffer from the phenomenon of ‘right truncation’. More recent patents cannot be analyzed as only citations received so far are known (Hall, et al., 2001).
A common practice adopted by researchers to counter this issue is to count citations received only in the first 5 years of patent filing.
Multiple citations to family members
Patent documents citing each other may not have a one-to-one relationship to inventions.
A given invention can be covered by a number of documents issued by national, regional or supranational offices. Sometimes, different family members that cover the same invention are cited by different examiners. To account for this underestimation, forward citations are counted at the family level (count of families citing families). (OECD, 2009)
Home and language bias
Examiners may succumb to home bias as they are more likely to cite documents from the same office or documents that maybe easy to find / understand (Alcacer and Gittelman, 2006).
3.2.1 Forward and backward citations
Forward citations are citations received by a patent from subsequent patents. More valuable or impactful patents are usually associated with higher number of forward citations. Forward citations have been found to be the most stable indicator of patent value in the literature. Studies have shown that forward citations are strongly correlated to economic value of patents (Trajtenberg, 1990) as well as the social value of inventions (Scherer et al., 2000). Renewal of patents is correlated to high count of forward citations (Putnam, 1996). Lanjouw and Schankerman (1997) showed that patents which receive more citations than average are more likely to be opposed or litigated.
Backward citations, or references added to a patent during patent examination, can potentially provide insights into the exploration process of new technologies (Jaffe, Trajtenberg and Fogarty, 2000; Lukach and Plasmans, 2002). The idea behind utilizing backward citations as a value determinant is in the assumption that combination and knowledge transfer from other technological domains would lead to more valuable patents (Nemet and Johnson, 2012). Backward citations have been a subject of debate among patent value researchers. The results of using backward citations as a value determinant have provided ambiguous results (Hall et al., 2001; Zeebroeck et al., 2011b).
References added to a patent application need not be only to other patent documents. Non-patent literature (NPL) refers to scientific articles in journal and conferences, along with other sources like books, internet links, abstract services, disclosure bulletins, and the like. A large number of NPL citations has been associated with a greater linkage to fundamental scientific research (Narin and Olivastro, 2006; Callaert et al., 2006).
Sapsalis, et al. (2006) have shown that patents citing NPL that are authored by the inventors themselves, are of much higher value. More broadly, they have argued for the importance of distinguishing the sources of citations for patent analyses.
3.2.2 Citation categories
The EPO examination guidelines7 require all cited documents to be identified by a certain letter or a combination of letters. Citation categories are a useful way of assigning relevance to citations. Table 2 below lists some of the most common categories used. X, Y and A are by far the most used categories. Documents assigned X indicate a clear similarity between the claimed invention and the contents of the cited document. X citations thus present a genuine obstacle to the grant of a patent. Y refers to documents
7 EPO search guidelines: http://www.epo.org/law-practice/legal-texts/html/guidelines/e/b_x_9_2.htm
that can be combined with other documents by a skilled person to develop a concept similar to the claimed invention. Documents ascribed A refer to citations that define the state of the art, but are not in conflict with the claimed invention. In simpler terms, an X reference indicates overlapping claims with a previous document; a Y reference means a combination of multiple Y documents can anticipate the claims of the patent being examined; and an A reference mostly defines the state of the art. Previous literature (Czarnitzki, et al., 2011; Harhoff and Wagner, 2009) has combined X and Y references and called them blocking citations. ‘A’ references are usually added by examiners if they are considered useful for the applicant when modifying claims.
Table 2 Most common citation categories. Source: OECD (2009)
3.2.3 Citation origin
Citations can be added to a patent document during different stages of the examination process. Most citations are added during the search phase. Some applicants may request a supplementary search report that may generate more references. Where necessary, these search reports are supplemented by the examiner and/or by third party observations during pendency. Even post-grant opposition and appeal proceedings might uncover and add prior art that may have been overlooked. For PCT applications entering Europe, citations should also be analyzed from the International Search Report (ISR) that is generated by the patent office where the application originated.
Studies related to this dissertation do exploratory empirical work related to citation origin and their influence on patent grant and opposition. Previous work related to citation origin
Citation Category Description
X Particularly relevant documents when taken alone
Y Particularly relevant if combined with another document of the same category A Documents defining the general state of the art
O Documents referring to non-written disclosure
P Documents published between the date of filing and the priority date
T Documents cited for a better understanding of the invention, published after the filing date
E Potentially conflicting patent documents, published on or after the filing date D Document provided by the applicant
L Document cited for other reasons
has been done for US patents. Alcacer, et al., (2009) and Cotropia, et al., (2013) have noted differences in the relevance of citations originating from applicants and examiners.