Some comments on the question whether co-occurrence data should be normalized
In a recent article in JASIST, L. Leydesdorff and L. Vaughan (2006) asserted that raw cocitation data should be analyzed directly, without first applying a normalization such as the Pearson correlation. In this communication, it is argued that there is nothing wrong with the widely adopted practice of normalizing cocitation data. One of the arguments put forward by Leydesdorff and Vaughan turns out to depend crucially on incorrect multidimensional scaling maps that are due to an error in the PROXSCAL program in SPSS.
|Persistent URL||dx.doi.org/10.1002/asi.20647, hdl.handle.net/1765/19262|
Waltman, L., & van Eck, N.J.P.. (2007). Some comments on the question whether co-occurrence data should be normalized. Journal of the American Society for Information Science, 58(11), 1701–1703. doi:10.1002/asi.20647