Abstract:I study the measurement of scientists' influence using bibliographic data. The main result is an axiomatic characterization of the family of citation-counting indices, a broad class of influence measures which includes the renowned h-index. The result highlights several limitations of these indices: they are not suitable to compare scientists across different fields, and they cannot account for indirect influence. I explore how these limitations can be overcome by using richer bibliographic information.
Digital Libraries,Social and Information Networks,Theoretical Economics,Physics and Society
What problem does this paper attempt to address?
The problem that this paper attempts to solve is about how to use bibliographic data to evaluate the influence of scientists, especially the limitations faced when measuring the influence of scientists through citation count indices (such as the h - index). Specifically, the paper focuses on the following points:
1. **Limitations of interdisciplinary comparison**: Existing citation count indices (such as the h - index) are not suitable for comparison between scientists in different disciplinary fields. This is because there are significant differences in citation habits and scales in different disciplinary fields, making direct comparison difficult.
2. **Ignoring of indirect influence**: Existing citation count indices cannot effectively capture the indirect influence of scientists. Indirect influence refers to the situation where a paper is cited by other papers, and these cited papers are further cited by other papers. This cumulative influence is very important in scientific research, but existing indices often overlook this point.
3. **Limitations of the information base**: Existing citation count indices are usually based only on the list of citation times of the author's papers, without making full use of more abundant bibliographic information. This leads to incomplete and inaccurate measurement of the influence of scientists.
To overcome these limitations, the paper proposes a new method to improve the method of measuring the influence of scientists by using more abundant bibliographic information. Specifically, through axiomatic analysis, the paper reveals the assumptions behind the existing citation count indices and explores how to overcome these limitations by introducing new indicators.
### Formulas and concepts
- **Euclidean Index**:
\[
f_a(d)=\left[\sum_{p\in P_a}(C_p)^2\right]^{1/2}
\]
where \(f_a(d)\) represents the influence of author \(a\) in database \(d\), \(P_a\) is the set of papers of author \(a\), and \(C_p\) is the citation count of paper \(p\).
- **Intellectual Debt**:
\[
f_a(d)=\sum_{p\in P_a}\sum_{b\in A}\frac{1}{P_b}\sum_{q\in P_b}\frac{1}{R_q}n(p, q)
\]
This index takes into account the contribution of each author to the entire scientific community and satisfies interdisciplinary comparability.
- **Comprehensive Measure**:
\[
f_a(d)=\sum_{p\in P_a}\sum_{q\in P}\sum_{r\in P}n(p, q)n(q, r)
\]
This index takes into account the influence of indirect citations, but does not satisfy interdisciplinary comparability.
### Main conclusions
- **Axiomatic analysis**: Through the axiomatic method, the paper reveals the basic assumptions of citation count indices, including separability, reference independence, splitting, citation anonymity, and author anonymity.
- **Interdisciplinary comparability**: The paper proposes a weaker concept of interdisciplinary comparability - field comparability, and proves that this property is incompatible with author anonymity. This means that existing citation count indices cannot meet the requirements of interdisciplinary comparability.
- **Improvement measures**: The paper proposes a new method. By introducing the Intellectual Debt index to consider indirect influence and satisfy interdisciplinary comparability.
In summary, this paper aims to reveal the limitations of existing citation count indices through theoretical analysis and the axiomatic method, and propose improvement measures to evaluate the influence of scientists more comprehensively and accurately.