Benford, Zipf and the blogosphere

Abstract

The blogosphere is a relatively recent development on the Internet and, for this reason, has received limited empirical investigation. This short note investigates whether the popularity of the 500 most popular blogs during August and September of 2005, as measured by in-coming links, followed the first-digit distribution attributed to Benford (1938) and the rank-size distribution attributed to Zipf (1949). The evidence suggests that the blogs investigated were not characterized by either empirical regularity, consistent with blog popularity being caused by network externalities.

Notes

¹It is not clear if this increase is a function of Feedster's technology for counting in-coming links or if this reflects an actual increase in the number of in-coming links to the top five hundred blogs.

²The appropriate critical value at the 1% (5%) level, derived by Nishiyama and Osada (2004), is −35.20 (−26.37), rather than the traditional 1.96 (2.58).