Research Focus
The h index and career assessment by numbers

https://doi.org/10.1016/j.tree.2006.01.005Get rights and content

Growing demand to quantify the research output from public funding has tempted funding agencies, promotion committees and employers to treat numerical indices of research output more seriously. So many assessment exercises are now conducted worldwide that traditional peer assessment is threatened. Here, we describe a new citation-based index (Hirsh's h index) and examine several factors that might influence it for ecologists and evolutionary biologists, such as gender, country of residence, subdiscipline and total publication output. We suggest that h is not obviously superior to other indices that rely on citations and publication counts to assess research performance.

Introduction

Several different citation-based indices are used to measure research performance (e.g. the number of highly cited papers published, the mean number of citations per paper and the total number of citations). There are valid reservations about using these indices to measure performance because papers are cited for reasons that are unrelated to the quality or utility of a study. For example, researchers cite papers more often by potential reviewers or editors [1] or by colleagues from the same country [2] (reviewed in [1]). Comparing researchers using these indices is problematic because citation patterns vary among scientific disciplines [3]. For biologists, there might even be a taxon effect. For example, herpetologists write conceptually broader introductions than do ornithologists, who focus more narrowly on birds [4]. This should generate an asymmetry in taxon-based citations, resulting in higher citation rates for biologists working on ‘popular’ organisms. It is therefore important to interpret these indices cautiously and to recognize their limitations.

Section snippets

The latest citation-based index

Recently, Nature [5] and Science [6] promoted a new measure of research performance developed by Jorge Hirsch called the ‘h index’ [7], defined as the maximum number of papers h by a scientist where each paper has received h or more citations. It can be calculated using a database such as Thomson Scientific's Web of Science© (http://isiknowledge.com) and sorting publications using the ‘times cited’ option: scroll down the output until the rank of the paper (in terms of citations) is greater

Measuring h for ecologists and evolutionary biologists

The rate at which papers accumulate citations varies across disciplines: cell biology publications accumulate citations more rapidly than do ecological publications [3]. The average number of papers per researcher and references per paper also vary across disciplines. How does this effect h, which depends on the number of papers published and the citations they have at the time of assessment? Hirsch noted that highly cited biological and biomedical scientists, as ranked by Thomson Scientific,

What factors influence h?

Although there was a linear relationship between h and scientific age, we used a log-log plot because it provided a better fit than the untransformed values (Box 1) (Figure 1a). Importantly, the fit between the residuals from this regression (=age-corrected h) and m was high (R2=80–94%) for all journals, except J. Vert. Paleontol. (R2=32%). The assumption of linearity and using m to control for scientific age are therefore appropriate. In support of the argument that some fields have lower m

Important facts about the h index

  • The h index is closely correlated with total publication output; thus, it will generally result in the same assessment as one based on counting publications. Therefore, if the link between publication rate and h is causal, we are unlikely to see researchers shift to producing fewer papers if h is widely deployed.

  • Female scientists produce fewer papers [9], which affects their h index. Assessors should be aware that the h index also shows a gender effect.

  • Comparison of highly cited scientists

References (10)

There are more references available in the full text version of this article.

Cited by (0)

View full text