The Hirsch-index and related impact measures

Abstract

The Hirsch-index (or h-index), introduced in 2005 by Jorge Hirsch, is one of the most popular indicators in information science and in informetrics. Hundreds of articles have been written on the h-index or related h-type indices. To quote Ball (2007): ”… the h-index does seem to be able to identify good scientists, and is becoming widely used informally, for example to rank applicants for research posts”. A review on this important topic is hence not superfluous. In the first section we give the definition of the h-index and present advantages and disadvantages of this index. To avoid some of the disadvantages of the h-index, several other, “h-type” indices (also called impact measures) have been introduced. These are discussed in the second section. Applications of these indices are given in the third section. These applications deal with the definition and use of h-type indices to cases different from authors (e.g. journals or topics). Also an overview of case studies are given. In the fourth section we study impact measures from the point of view of their intrinsic properties they have or should have. We remark that it is not possible to exactly define what a good impact measure is. Axiomatic characterizations of some impact measures are given. We also look at the influence of production on some impact measures. In the fifth section we discuss some informetric models for these h-type measures (e.g. their dependence on the total number of articles and on the total number of citations). Also dynamical aspects of these indices are described: the influence of transformations on h-type indices. Also distributions of h-type indices are studied and it is noted that these are of a different nature than the distributions of the impact factor. In the last section (before the conclusions section) we study h-type indices in function of time (e.g. the career of an author) and examine the possible shapes of these functions (several h-index sequences are studied). Case studies are presented. We close with some conclusions and open problems. This review does not intend to describe many case studies (although an attempt is made to present a complete reference list up to the time of the manuscript’s final submission) but focusses on the “philosophy” behind h-type indices as impact or performance measures and their potentialities as new informetric indicators. As always, this review contains some personal opinions.