Were has the cumulative advantage gone? Some observations about the frequency distribution of scientific productivity, of duration of scientific participation, and of speed of publication

Frequency distributions of scientific productivity are usually based on cross section cuts of the investigated population of scientists. Therefore, some of the registered scientists are involved for the whole period of time, but there are many fractional authors, too. If one compares only scientists active in a specialty for the same length of time, the typical bibliometric skewness of the distribution vanishes. But also the duration of participation of a cohort of scientists which began their career in the same year is not distributed in a Lotkean manner. Furthermore, the speed of publication-which might be a better statistical indicator of scientific capacities than publication output as such—has more similarity to a normal distribution than to a lognormal one.