Requirements for a cocitation similarity measure, with special reference to Pearson's correlation coefficient

@article{citeulike:881958, abstract = {Author cocitation analysis (ACA), a special type of cocitation analysis, was introduced by White and Griffith in 1981. This technique is used to analyze the intellectual structure of a given scientific field. In 1990, McCain published a technical overview that has been largely adopted as a standard. Here, McCain notes that Pearson's correlation coefficient (Pearson's r) is often used as a similarity measure in ACA and presents some advantages of its use. The present article criticizes the use of Pearson's r in ACA and sets forth two natural requirements that a similarity measure applied in ACA should satisfy. It is shown that Pearson's r does not satisfy these requirements. Real and hypothetical data are used in order to obtain counterexamples to both requirements. It is concluded that Pearson's r is probably not an optimal choice of a similarity measure in ACA. Still, further empirical research is needed to show if, and in that case to what extent, the use of similarity measures in ACA that fulfill these requirements would lead to objectively better results in full-scale studies. Further, problems related to incomplete cocitation matrices are discussed.}, address = {Swedish School of Library and Information Science, 501 90 Bor\r{a}s, Sweden; KHBO, Department of Industrial Sciences and Technology, B-8400 Oostende, Belgium}, author = {Ahlgren, Per and Jarneving, Bo and Rousseau, Ronald }, citeulike-article-id = {881958}, comment = {Conclusion We have shown that if the main objective of a cocitation study is to find clusters or maps of authors with a similar cocitation pattern, and to explain changes of such maps over time, the Pearson correlation coefficient is probably not an optimal choice. ---=note-separator=--- Leydesdorff and Zaal (1988): These scientists performed a co-word analysis of a group of biochemistry articles. They used Pearson’s correlation coefficient, Jaccard’s measure, Euclidean distances, and Salton’s cosine measure as similarity measures, did not make much difference ---=note-separator=--- vector similarity measure (association measure) == pearson correlation == Chi-Squared Distance == Salton's Cosine Measure ---=note-separator=--- two “orthogonal methods for similarity measure One method only uses the A-B cocitation frequency; the other one uses all cocitation frequencies (involving A and B), except the A-B cocitation frequency (pearson's r). }, doi = {http://dx.doi.org/10.1002/asi.10242}, journal = {Journal of the American Society for Information Science and Technology}, keywords = {co-citation, similarity}, number = {6}, pages = {550--560}, posted-at = {2006-12-06 10:50:39}, priority = {0}, title = {Requirements for a cocitation similarity measure, with special reference to Pearson's correlation coefficient}, url = {http://dx.doi.org/10.1002/asi.10242}, volume = {54}, year = {2003} }

TY - JOUR ID - citeulike:881958 L3 - citeulike-article-id:881958 TI - Requirements for a cocitation similarity measure, with special reference to Pearson's correlation coefficient JF - Journal of the American Society for Information Science and Technology VL - 54 IS - 6 SP - 550 EP - 560 AD - Swedish School of Library and Information Science, 501 90 Borås, Sweden; KHBO, Department of Industrial Sciences and Technology, B-8400 Oostende, Belgium N2 - Author cocitation analysis (ACA), a special type of cocitation analysis, was introduced by White and Griffith in 1981. This technique is used to analyze the intellectual structure of a given scientific field. In 1990, McCain published a technical overview that has been largely adopted as a standard. Here, McCain notes that Pearson's correlation coefficient (Pearson's r) is often used as a similarity measure in ACA and presents some advantages of its use. The present article criticizes the use of Pearson's r in ACA and sets forth two natural requirements that a similarity measure applied in ACA should satisfy. It is shown that Pearson's r does not satisfy these requirements. Real and hypothetical data are used in order to obtain counterexamples to both requirements. It is concluded that Pearson's r is probably not an optimal choice of a similarity measure in ACA. Still, further empirical research is needed to show if, and in that case to what extent, the use of similarity measures in ACA that fulfill these requirements would lead to objectively better results in full-scale studies. Further, problems related to incomplete cocitation matrices are discussed. KW - co-citation KW - similarity N1 - Conclusion We have shown that if the main objective of a cocitation study is to find clusters or maps of authors with a similar cocitation pattern, and to explain changes of such maps over time, the Pearson correlation coefficient is probably not an optimal choice. N1 - Leydesdorff and Zaal (1988): These scientists performed a co-word analysis of a group of biochemistry articles. They used Pearson’s correlation coefficient, Jaccard’s measure, Euclidean distances, and Salton’s cosine measure as similarity measures, did not make much difference N1 - vector similarity measure (association measure) == pearson correlation == Chi-Squared Distance == Salton's Cosine Measure N1 - two “orthogonal methods for similarity measure One method only uses the A-B cocitation frequency; the other one uses all cocitation frequencies (involving A and B), except the A-B cocitation frequency (pearson's r). AU - Ahlgren, Per AU - Jarneving, Bo AU - Rousseau, Ronald PY - 2003/// UR - http://dx.doi.org/10.1002/asi.10242 ER -