The H-Index in Australian Astronomy
Total Page:16
File Type:pdf, Size:1020Kb
Addenda: The h-index in Australian Astronomy Kevin A. PimbbletA,B A School of Physics, Monash University, Clayton, VIC 3800, Australia B Email: [email protected] Abstract: Pimbblet (2011) published an evaluation of the Hirsch h-index in the context of the Aus- tralian astronomical community. This addenda adds treatment of changes of surname to the compu- tation of the h-index and presents derivative data on the m-index. Keywords: Errata, Addenda 1 Change of Surname Table 1: Top 10 h-index for Australian Astron- Pimbblet (2011) published a variety of analyses on the omy, excluding overseas professionals with correc- h-index (e.g. Hirsch 2005) in the context of Australian tion applied for surname changes. Astronomy. Alongside that, a discussion of a number Rank Name h of caveats was also made. One further caveat merits =1 KenFREEMAN 77 explicit attention in the interpretation of the h-index: =1 JeremyMOULD 77 changes of surname. In the Pimbblet (2011) analysis, 3 Karl GLAZEBROOK 71 no attempt was made to track surname changes. This 4 Dick MANCHESTER 68 has the effect of depressing the h-index of individuals 5 MichaelDOPITA 64 who have changed their names over the years. Its effect can readily be seen in Table 1 where we re-compute the 6 Warrick COUCH 61 top ten h-index for Australian Astronomy and account 7 Joss BLAND-HAWTHORN 57 for changing surnames: Bland-Hawthorn’s previous h- 8 MatthewCOLLESS 56 index was suppressed by 4 points by not taking this in 9 Brian SCHMIDT 54 to account. We emphasize that the same dataset was 10 MikeBESSELL 53 used for this re-computation as the original paper. It is impractical to attempt to account for this for the entire membership of the Astronomical Society of Australia, hence we refrain from publishing percentiles of the h- index that account for this correction. of the Astronomical Society of Australia (see Pimb- This correction could potentially affect female sci- blet 2011) are presented in Table 2. It is interesting to entists more than male scientists. We therefore re-issue note how this distribution appears to be shifted toward the caution and caveat emptor that care must be used higher values than generally noted by Hirsch (2005). in the interpretation of any individual’s h-index and A careful examination of the individuals at the upper that the tables published by Pimbblet (2011) have de- end of the distribution reveals the likely cause of this: ficiencies: both known and unknown. many are only one or two years outside of their Ph.D. Finally we note that upon request, ADS is able to award date. Therefore, we recompute the m-index dis- create a synonym for surname changes. Once articles tribution for individuals who have had a minimum of with both surnames exist within ADS, a search upon 5 years of active research time in the last row of Ta- either name will return all results (Grant, priv. comm.). ble 2. This provides a much more realistic treatment of the high end of the m-index distribution. Moreover, it is also broadly consistent with the views expressed 2 m-index in Hirsch (2005) that m ∼ 1 denotes a successful scien- tist, m ∼ 2 are outstanding, and m ∼ 3 or higher are The data gathered by Pimbblet (2011) can also yield truly unique individuals. Again though, we remind the a derived quantity: the m-index. The m-index is for- reader that this analysis is at best an approximation, mally the rate of growth of the h-index, defined as the particularly given the issues noted above. h-index divided by the number of active research years (Hirsch 2005). Thus a scientist with an h-index of 20 and 20 years of active research will have an m-index Acknowledgments of 1. By approximating the number of active research KAP thanks Joss Bland-Hawthorn, Sarah Brough, Mar- years as the time elapsed since the Ph.D. award date, tin Asplund, and Andrew Hopkins for discussion, and we are able to obtain an approximation for the m-index many other correspondants for encouragement. in Australian Astronomy. The percentiles for the m- This research has made use of NASA’s Astrophysics index distribution for various membership categories Data System Bibliographic Services. 1 2 Publications of the Astronomical Society of Australia Table 2: Percentiles of m-index distribution. The sample labelled ‘5’ are those researchers in M+F-R-O categories who have had five or more years active research as defined by their Ph.D. award date – see text for detail. Sample N Percentile 25 50 75 90 95 97.5 99 M+F+R-O 242 0.64 1.25 1.89 2.43 3.00 4.20 4.87 M+F-R-O 206 0.91 1.47 2.04 2.50 3.18 4.00 4.98 5 183 0.83 1.37 1.93 2.36 2.84 3.35 4.03 References [1] Hirsch, J.E. 2005, Proc. Nat. Acad. Sciences 102, 16569 [2] Pimbblet, K.A. 2011, PASA, 28, 140 CSIRO PUBLISHING Publications of the Astronomical Society of Australia, 2011, 28, 140–143 www.publish.csiro.au/journals/pasa The h-index in Australian Astronomy Kevin A. Pimbblet School of Physics, Monash University, Clayton, VIC 3800, Australia. Email: [email protected] Received 2011 January 29, accepted 2011 March 10 Abstract: The Hirsch h-index is now widely used as a metric to compare individual researchers. To evaluate it in the context of Australian astronomy, the h-index for every member of the Astronomical Society of Australia (ASA) is found using NASA’s Astrophysics Data System Bibliographic Services. Percentiles of the h-index distribution are detailed for a variety of categories of ASA members, including students. This enables a list of the top ten Australian researchers by h-index to be produced. These top researchers have h-index values in the range 53 , h , 77, which is less than that recently reported for the American Astronomical Society membership. We suggest that membership of extremely large consortia such as the Sloan Digital Sky Survey may partially explain the difference. We further suggest that many student ASA members with large h-index values have probably already received their Ph.D. and need to upgrade their ASA membership status. To attempt to specify the h-index distribution relative to opportunity, we also detail the percentiles of its distribution by years since Ph.D. award date. This shows a steady increase in h-index with seniority, as can be expected. Keywords: sociology of astronomy — publications, bibliography — astronomical databases: miscellaneous 1 Introduction et al. (2011) presentation and seeks to determine the The modern research academic is judged as never before: typical range of the h-index in Australian astronomy, a large variety of metrics are now employed to determine which may be of use for future employers and employees the worth and merit of researchers, particularly when it in the community. The format of this work is as follows. In comes to hiring. Anecdotally, one of the chief metrics Section 2, we give an overview of the dataset that we use: used is the Hirsch index (h-index; Hirsch 2005). The the membership of the Astronomical Society of Australia. h-index is formally defined as follows: ‘A scientist In Section 3 we determine percentiles of the h-index has index h if h of his or her Np papers have at least h distribution for a variety of ASA membership categories, citations each and the other (Np 2 h) papers have # h including students. To attempt to normalize relative to citations each’ (Hirsch 2005). Its modest simplicity is opportunity, we re-evaluate the h-index distribution as probably a prime factor in its rapid pick-up by major a function of time elapsed since Ph.D. award date in publishers (Anon. 2005; Ball 2005). Moreover, this index Section 4. Our conclusions are presented in Section 5. is particularly useful as it has superior predictive power (in terms of productivity) for the future of researchers 2 Data compared to the total number of career citations, career To determine the h-indices of Australian astronomers, we publications and mean citations per paper (Hirsch 2007). make use of the Astronomical Society of Australia (ASA) Although other metrics and analyses exist (cf. Pearce membership list. The membership list is a fair represen- 2004; Kurtz et al. 2005; Egghe 2006; Jin 2006; Kosmulski tation of the Australian astronomical community: the 2006; Blustin 2007; Jin et al. 2007; Bornmann, Mutz & majority of professional astronomers are members. Daniel 2008; Wu 2010; Zyczkowski 2010), the h-index Membership of the ASA comes in several different cat- remains as the most prominent of its class in the field. egories, each of which we indicate with a single letter as Recently, Conti et al. (2011) presented work on the detailed in Table 1. The advantage of the ASA member- astronomer’s H-R diagram (number of Google search ship list is that we can distinguish between different results versus citations and h-index) for members of grades of members (i.e. amateurs and professional the American Astronomical Society (AAS). Contained astronomers who actively publish) to better probe the within that presentation are a number of interesting h-index in these sub-categories. concepts: a top-ten list of AAS members by h-index For each ASA member, we then implement a search (spanning the range 94 , h , 118) and the h-indices of in NASA’s Astrophysics Data System (ADS) to return a all AAS members.