The influence of self-citation corrections on Egghe's g index

The g index was introduced by Leo Egghe as an improvement of Hirsch's index h for measuring the overall citation record of a set of articles. It better takes into account the highly skewed frequency distribution of citations than the h index. I propose to sharpen this g index by excluding the self-citations. I have worked out nine practical cases in physics and compare the h and g values with and without self-citations. As expected, the g index characterizes the data set better than the h index. The influence of the self-citations appears to be more significant for the g index than for the h index.