https://orcid.org/0000-0002-7037-779X
?Mathematical formulae have been encoded as MathML and are displayed in this HTML version using MathJax in order to improve their display. Uncheck the box to turn MathJax off. This feature requires Javascript. Click on a formula to zoom.ABSTRACT
Clinical relevance
Publications in refereed scientific journals provide a rigorous research base that underpins clinical optometric practice. Leading optometrists who generate this literature can be identified and ranked using standardised citation indicators.
Background
This work seeks to identify and rank all optometrists included in a Science-Wide author database of standardised citation indicators (S-W) and to compare this ranking with the Global Optometrist Top 200 Research Ranking (T200).
Methods
A search was conducted for the names of all optometrists in the T200 who were included in the S-W, which is a world-wide listing of the top 2% of scientists in each of 174 subfield disciplines, ranked according to a composite citation indicator (cns) that excludes self-citations and corrects for multiple authorships and author order.
Results
The names of 66 optometrists are found in the S-W. Of these, 58 are designated as working in the primary sub-field ‘Ophthalmology & Optometry’; this listing, in rank-order of cns, is referred to as the ‘S-W-derived Optometrist Research Ranking’ (S-WORR). Australian optometrist Nathan Efron is ranked #1 in the S-WORR. The number (%) of optometrists in the S-WORR from each country is: the United States – 26 (45%), Australia – 12 (21%), the United Kingdom – 11 (19%), Canada – 5 (9%), Spain – 2 (3%), Hong Kong – 1 (2%) and South Africa – 1 (2%). The universities housing the equal highest number of optometrists in the S-WORR (five each) are the University of California, Berkeley, USA; the University of New South Wales, Australia; and Queensland University of Technology, Australia. There is a moderately strong correlation between T200 and S-WORR rankings (ρ = 0.6017, N = 58, p < 0.0001).
Conclusions
The S-WORR represents an elite cohort of optometrists who ought to be celebrated for their outstanding, leading and impactful contributions to optometric research.
Introduction
The Global Optometrist Top 200 Research Ranking (T200),Citation1 constructed by the present authors, has attracted considerable attention in the nine months since being published on-line in January 2021, as evidenced by this paper having been downloaded 3078 times from the website of the publisher. This makes it the most downloaded paper from Clinical and Experimental Optometry in 2021, by a margin of 1016 downloads from the second most downloaded paper. Over the same nine month period, there were 5335 unique views of the T200 companion website (www.optomrankings.com).
There are many reasons for the high interest in the T200: (a) it is a novel and unique resource for optometry; (b) universities, funding bodies, individual researchers in various disciplines, and others, may derive benefit from having access to an objective analysis of the research impact of leading optometrists in this field; (c) research optometrists can benchmark their own bibliometrics against those in the T200; and (d) a wider audience of clinical optometrists, vision scientists, optometry students, and the broader public, are able to discover, from a single resource, optometrists at the forefront of research in the field.
After publishing the T200, a number of colleagues contacted the authors asking how the ranking order would be impacted if standardised for confounding influences in publication and citation behaviour, as discussed extensively in the original T200 paper.Citation1 As it turns out, at the time the T200 was being constructed, Ioannidis et al.Citation2 from Stanford University published a ‘Science-Wide author database of standardised citation indicators’ (S-W). This standardisation approach corrects for several confounding influences inherent in a simple ranking by h-index alone, as is the case in the T200.Citation1 Specifically, the composite citation indicator used by Ioannidis et al.Citation2 accounts for differences in impact between the various disciplines in which authors work, and corrects for self-citations, multiple authorships and author order.
The purpose of this work is to identify and rank all optometrists included in the S-W according to their composite citation indicator, and compare and contrast this ranking with the T200. An answer can therefore be provided to the question raised by colleagues as to how rankings change when confounding influences are accounted for and corrected using standardised citation indicators.
Science-wide author database of standardised citation indicators (S-W)
Ioannidis et al.Citation2 identified the top 2% of scientists in the world, who have published at least five papers, in each of 174 ‘subfields’ (disciplines). Subfields were defined according to a Science-Metrix classification of journals, with multi-disciplinary journals classified using a character-based convolutional deep neural network.Citation2–4 A total database of 159,683 scientists (i.e. the top 2% of all scientists in the world) was generated and each author was determined—according this classification algorithm—to be working in a primary and secondary subfield. Citation metrics were derived from the Scopus database (Elsevier), with a data freeze of 6 May 2020, in order to assess scientists for career-long citation impact up until the end of 2019.
Ioannidis et al.Citation2 derived an equation to calculate the composite indicator for career-long impact, excluding self-citations, cns (where c is the composite indicator and the subscript ‘ns’ refers to ‘no self-citations’). The equation sums the ratio of the natural log of the indicator value + 1, over the natural log of the maximum of those indicator logs (designated with the subscript ‘max’) for six indicators (nc, h, hm, ncs, ncsf and ncsfl) as defined in , and is expressed as follows:
Table 1. Component indicators used in the calculation of the composite indicator for career-long impact excluding self-citations (cns)
Comparisons of citation metrics and consequent rankings are more meaningful when made within the same subfield,Citation1–4 and the S-W facilitates this. The various indicators incorporated into EquationEquation 1(1)
(1) are designed to derive an equitable assessment of impact. For example, with the trend towards mega-authored papers, an author of a given paper might be one of hundreds of authors. That author generally would be considered as having made less of a contribution to the paper than if the same author was the sole author of the paper. Citations to multi-authored papers are therefore ‘discounted’ by using the indicator hm, which is the Schreiber downward-adjusted correction of h.Citation5
Other indicators in EquationEquation 1(1)
(1) accord greater weight for total citation count, as well as citations to papers upon which the author is the sole, first or last (senior) author, because these author positions suggest pivotal contributions to the work.
Global Optometrist Top 200 Research Ranking (T200)
An extensive global search for the names of leading optometric researchers (with a notional h-index ≥ 15), conducted by the present authors between September and December 2020, resulted in the identification of 480 names.Citation1 A custom-designed bibliographic search tool was developed to interrogate the Scopus database (Elsevier) and extract publication metrics using the unique Scopus Author Identifier number for each of these 480 optometrists. On 13 January 2021, the full list was ranked by h-indexCitation6 and then truncated to reveal the top 200 optometrists; this listing constitutes the originally-published T200.Citation1 The output from the custom tool automatically updates the bibliometrics of all 480 optometrists every 24 hours, but only the top 200 are published on-line, at www.optomrankings.com.
Methods
The essence of the methodology, described in detail below, was to (a) capture a current, updated version of the T200; (b) search for all names on this T200 listing that appear in the S-W; (c) construct a rank-order listing, by cns, of the optometrists found on the S-W and (d) compare and contrast the ranking of optometrists on the T200 versus those found on the S-W.
Capturing an updated T200 listing for the present analysis
Since establishing the T200 website in January 2021, the authors have received notification of the names of approximately 50 optometrists, via the reporting form on www.optomrankings.com, who may have been ‘overlooked’ when the listing was first constructed. Following verification, the names of optometrists with h ≥ 15 were added to the master database. Eight of those suggested now appear in the T200 because their h-indices (and other metrics if tie-breaking is required) are higher than that of the optometrist ranked #200.
For the purpose of the present work, the T200 listing was downloaded from www.optomrankings.com on 15 July 2021, and is shown in Table S1 of the Supplementary Material. This table represents an accurate, complete and up-to-date listing of the top 200 research optometrists in the world, ranked according to h-index.
Generating a rank-order list by cns of T200 optometrists derived from the S-W
Various versions of the S-W are available from Ioannidis et al.,Citation2 via a link to the Mendeley website, in the form of searchable Excel spreadsheets. For the purposes of this work, ‘Table S6-career-2019ʹ was downloaded and searched for all 200 names from the current version of the T200 (i.e. the 200 names in Table S1 of the Supplementary Material). The names of 66 optometrists from the T200 (33%) were found in the S-W. It is apparent that a T200 h-index of approximately 35 is required to be included in the S-W, notwithstanding variances arising from shifts in ranking due to metric standardisation, as discussed in detail below.
Results
Of the 66 optometrists who were found in the S-W, 58 were determined by Ioannidis et al.Citation2–4 to be working in the primary subfield ‘Ophthalmology & Optometry’, according to journals in which they primarily publish their work, as described earlier. These optometrists, who are among the top 2% of scientists in the world, are listed in rank-order of cns in , which shall be referred to as the ‘S-W-derived Optometrist Research Ranking’ (S-WORR). The ranking positions (according to h-index) of these optometrists in the T200 captured on 15 July 2021 (taken from Table S1 of the Supplementary Material) are also displayed in the last column of the for ready comparison.
Table 2. Ranking (by cns) of optometrists included in the S-W primary subfield of ‘Ophthalmology & Optometry’
According to the S-W database subfield classification algorithm used by Ioannidis et al.,Citation2–4 four optometrists were revealed to be working in the primary subfield ‘Experimental Psychology’ and four in ‘Neurology & Neurosurgery’. It is a well-established principle of bibliometrics that metric values, or ranking positions, of authors from different disciplines cannot be directly compared.Citation3,Citation4,Citation7 Thus, it is not possible to directly compare cns values of those in different sub-fields. Accordingly, separate rank-order lists have been constructed for the optometrists determined to be working in the primary subfields of ‘Experimental Psychology’ and ‘Neurology & Neurosurgery’, which are presented in , respectively.
Table 3. Ranking (by cns) of optometrists included in the S-W primary subfield of ‘Experimental Psychology’
Table 4. Ranking (by cns) of optometrists included in the S-W primary subfield of ‘Neurology & Neurosurgery’
It is pertinent to observe that only four of the eight optometrists (50%) who work in subfields other than ‘Ophthalmology & Optometry’ (i.e. all those listed in ) are located in optometry settings; Gerald Westheimer works in a department of molecular and cell biology, Robert Hess and Donald Mitchell work in psychology departments and Tailoi Chan-Ling works in a department of medicine. In contrast, 50 of the 58 optometrists (86%) listed in the S-WORR () are located in optometry settings. These proportions are significantly different (χ2 = 6.20, p = 0.013), suggesting that the above observations in respect of work settings are broadly consistent with the subfield classification paradigm employed by the S-W.
Given that this work seeks to focus on leading optometrists actually working in the primary subfield ‘Ophthalmology & Optometry’, attention is directed to the S-WORR in . Of the 58 optometrists in the S-WORR, 12 (21%) are female.
The top-ranked optometrist is Nathan Efron,Citation8 emeritus professor in the School of Optometry and Vision Sciences at the Queensland University of Technology, Australia. His primary research interest is in the ocular response to contact lens wear. He has also published extensively on ophthalmic markers of diabetic neuropathy.
Interestingly, the three top-ranked optometrists – Nathan Efron (rank #1), Konrad Pesudovs (#2) and David Atchison (#3) – are Australian, and all three graduated in optometry from the University of Melbourne.
The distribution of countries where these 58 optometrists in the S-WORR are currently working, or worked before they retired or died, is as follows: the United States – 26 (45%), Australia – 12 (21%), the United Kingdom – 11 (19%), Canada – 5 (9%), Spain – 2 (3%), Hong Kong – 1 (2%) and South Africa – 1 (2%).
The 58 optometrists listed in the S-WORR conduct their research in 30 different work locations (mainly universities, but also independent laboratories/practices, or companies). Three of these work locations (all universities) house the equal highest number of optometrists in the S-WORR (five each); these are the University of California, Berkeley, USA; the University of New South Wales, Australia; and Queensland University of Technology, Australia.
Discussion
The simple reason why 134 optometrists in the T200 were not found in the S-W is that their cns scores were of insufficient magnitude to be ranked among the top 2% of scientists in the S-W primary subfield in which they were found to have worked, according to the algorithm used by Ioannidis et al.Citation2–4 The present authors are confident that the cohort of 66 optometrists found in the S-W represents a complete listing of all optometrists in the S-W, because this cohort is essentially a subset of the most impactful optometrists from the T200. Putting this another way, it is highly unlikely that an optometrist with a h-index that is insufficient for inclusion in the T200 (i.e. h < 21) would appear in the S-W, which has a much higher threshold for inclusion (around h = 35).
As a further check that no optometrists in the S-W have been missed, an additional search was made of the names of senior colleagues and others among those ranked 201–300 in the extended, unpublished section of the T200 master database captured on 15 July 2021 (who have h-indices ranging from 16 to 20). Specifically, a search was made for those who theoretically may have been ranked high enough for inclusion in the S-W according to the cns metric – such as Anthony Cullen, Elwin Marg, Merton Flom, John Lovasik, George Woo, Barry Collin and Glenn Fry – by virtue of publishing a significant number of sole author or small-author-group papers. This search failed to reveal the names of any further optometrists in the S-W.
An advantage of the cns metric is that it is a continuous variable; therefore, tie-breaking is not required for ranking. In contrast, rankings in the T200 are by h-index,Citation6 which are whole integers. Numerous optometrists can therefore have the same h-index; for example, as can be seen in the T200 listing in Table S1 of the Supplementary Material, 18 optometrists are tied with a h-index of 23. A three-level tie-breaking hierarchy, designed to reflect impact, was devised, based on alternative metrics and used to determine the rank order of the otherwise-tied optometrists.Citation1 In the S-WORR shown in , it has been necessary to cite cns values to 5 decimal places to avoid ties.
Comparing the rank-order of optometrists in the T200 versus S-WORR
The h-indices for the 58 optometrists listed in the S-WORR () are lower than their corresponding h-indices listed in the T200 (Table S1 of the Supplementary Material) for two reasons. First, self-citations are excluded from the S-W h-index calculations, necessarily resulting in lower h-indices compared with h-indices unadjusted for self-citations listed in the T200 (unless there were no self-citations, in which case the S-W and T200 h-indices would be identical). Accordingly, cns values are always lower than c values unadjusted for self-citations.
Second, the bibliometrics used to construct the T200 and S-WORR were generated about 14 months apart; specifically, the data freeze for the T200 for this analysis was 15 July 2021, whereas that for the S-W was 6 May 2020. Because authors continue to publish papers, and citations continue to accrue over time, the magnitude of h-indices in the T200 generally will be higher than the corresponding h-indices in the S-WORR. However, the 14-month time difference is unlikely to have had a significant impact on the relative ranking of optometrists between the two databases, which is the primary focus of this work.
The ranking positions of the 58 optometrists in the S-WORR were compared with their corresponding ranking positions in the T200 (the first and last columns, respectively, in ) using Spearman’s rank order correlation coefficient. The result of this test demonstrates a moderately strong correlation (ρ = 0.6017, N = 58, p < 0.0001).
The differences in ranking positions of the 58 optometrists listed in the S-WORR compared with their ranking positions in the T200 are illustrated in the form of a slope plot in . As can be seen from this plot, a large number of optometrists have similar rankings; for example, 20 of the 58 optometrists have a difference of ≤ 4 ranks between the S-WORR and T200.
Figure 1. Slope plot comparing the rankings of 58 optometrists in the T200 versus S-WORR. Line colouration facilitates differentiation but is otherwise arbitrary.
However, a number of optometrists have a greater ranking differential. Forensic examination of the values of the component indicators used to calculate cns for a given optometrist, as presented in , can reveal the reason for the relative ranking positions of that individual in the T200 versus S-WORR. In general, rankings will be higher in the S-WORR for individuals who have a higher total number of citations and a higher h-index; less frequently cite their own work; publish more often as a single author and/or as one of a small author group; and publish more papers as first or last (senior) author.
Two examples are offered to demonstrate the reason for ranking differentials between the T200 versus S-WORR. The optometrist with the greatest improvement is Charles McMonnies, whose ranking jumped from #181 in the T200 to #24 in the S-WORR. The primary reason for this improved ranking in the S-WORR is that McMonnies has published a large number of single-author papers (100) that have been extensively cited (934 times). As well, his S-WORR h-index of 19 drops by only a small margin, to 17, when Schreiber-adjusted for multiple co-authorships, indicating that when not publishing alone, McMonnies generally writes papers with small groups of co-authors, more often as first or last (senior) author.
The optometrist displaying the greatest decrease is Debra Schaumberg, whose ranking of #6 in the T200 drops to #46 in the S-WORR. The primary reason for this decreased ranking in the S-WORR is that Schaumberg has not published any single-author papers and thus has zero single-author citations. Also, her S-WORR h-index of 50 drops substantially to 16 when Schreiber-adjusted for multiple co-authorships, indicating that she generally publishes with large groups of co-authors, and less often as first or last (senior) author.
The five optometrists who have had the greatest improvement in ranking position in the S-WORR compared with the T200 – Charles McMonnies (rank difference 151, discussed above), William Harris (116), Ian Bailey (109), Gordon Ruskell (105) and Theodore Grosvenor (76) – are more senior colleagues who commenced publishing in, or prior to, 1970 and have thus emerged from an era during which single or small group authorship was the approach du jour. Over the past three decades, science has generally evolved towards large-scale, multinational, multi-disciplinary collaborations, typically reported by large teams of authors. The more traditional approach of single or small group authorship has seemingly benefited these optometrists on the S-WORR.
Caveats to this work
Some important caveats to this work need to be considered. The analysis reported here was conducted in respect of ranking lists and databases that were ‘frozen in time’. There is good reason for this: as time goes on, authors continue to accrue citations and publish papers, which essentially renders these rankings to be dynamic. While the h-index can only increase, the cns metric can increase or decrease. Indeed, this dynamic nature of citation metrics is the very reason why a constantly-updating version of the T200 (www.optomrankings.com) was established by the present authors. Re-analysis of the T200 and S-WORR would likely produce different rankings if repeated with updated metrics in a few months time. Nevertheless, the ongoing dynamic nature of citation metrics does not invalidate the present analysis, which demonstrates how the use of standardised citation indicators can impact, and arguably correct, rankings based on unadjusted h-indices.
There are many ways in which standardised citation indicators can be derived, and different models would likely produce different rankings. The S-W generated by Ioannidis et al.Citation2 was used in the present analysis because it is apparently the only such database published to date. It serves as a valuable comparator for the T200, which is based on raw h-indices, and helps demonstrate the extent to which research rankings are affected by correcting for potential confounding patterns of publishing.
In the entire primary subfield ‘Ophthalmology & Optometry’, many scientists rank higher than the 58 optometrists listed; these primarily would be ophthalmologists, but also vision scientists.
Although Ioannidis et al.Citation2 cited cns values to nine decimal places in order to unambiguously rank 159,683 scientists, it was only necessary in this work to cite cns values to five decimal places to avoid ties among the 58 optometrists in the S-WORR. Citing cns to the required number of decimal places to break ties is a valid approach, because by definition, higher cns values indicate greater impact. However, the practical significance of differences at the 5th decimal point is open to debate. For example, the most closely ranked optometrists in the S-WORR – Suzanne Fleiszig (rank #39) and Anthony Adams (#40) – have a cns differential of 0.00001. When considering the overall academic work of these two optometrists, and others who are similarly closely ranked, some may view such small numeric differences in cns as immaterial.
Conclusions
Presented here is a rank-order list, derived from the S-W, of the 58 most impactful optometrists in the world, determined to be publishing in the subfield of ‘Ophthalmology & Optometry’. These optometrists are among the top 2% of scientists world-wide. All of those listed in the S-WORR (presented in ) are therefore members of an elite cohort who ought to be celebrated for their outstanding, leading and impactful contributions to optometric research.
The S-WORR accords higher ranks to optometrists who have a higher total citation count and h-index; minimise self-citations; tend to publish more as sole author or in small author groups; and appear more often as first or last (senior) author of published works. Ostensibly, therefore, this listing more accurately reflects the hierarchy of impact and achievement of those listed, compared with the T200, which despite its utility and currency, has some deficiencies, which have been recognised previously and discussed at length.Citation1 The present work highlights the potential impact of those deficiencies, demonstrates how they can be rectified with a more refined metric (cns in this instance), and reinforces the need to exercise caution when interpreting any research ranking list.
While differences do exist between the ranks of individual optometrists in the T200 versus the S-WORR, the overall ranks of the 58 optometrists who appear in both lists are moderately well correlated, essentially providing cross-validation to each ranking algorithm. Given this observation, and the impracticality, for technical reasons, of maintaining a live on-line version of the S-WORR, the authors intend on maintaining, into the future, the constantly-updating T200 website (www.optomrankings.com), as a freely-available and reasonably accurate listing of the most impactful research optometrists in the world.
Supplemental Material Table 1
Download PDF (145.5 KB)Disclosure statement
The authors stand to derive reputational benefit from this work by virtue of being ranked among the most impactful research optometrists world-wide.
Supplementary material
Supplemental data for this article can be accessed here
References
- Efron N, Morgan PB, Jones LW, et al. Global Optometrist Top 200 Research Ranking. Clin Exp Optom 2021; 104: 471–485.
- Ioannidis JPA, Boyack KW, Baas J. Updated science-wide author databases of standardized citation indicators. PLoS Biol 2020; 18: e3000918.
- Ioannidis JP, Klavans R, Boyack KW. Multiple citation indicators and their composite across scientific disciplines. PLoS Biol 2016; 14: e1002501.
- Ioannidis JPA, Baas J, Klavans R, et al. A standardized citation metrics author database annotated for scientific field. PLoS Biol 2019; 17: e3000384.
- Schreiber M. A modification of the h-index: the hm-index accounts for multi-authored manuscripts. J Sci Inform 2008; 2: 211–216.
- Hirsch JE. An index to quantify an individual’s scientific research output. Proc Natl Acad Sci 2005; 102: 16569–16572.
- Patience GS, Patience CA, Blais B, et al. Citation analysis of scientific categories. Heliyon 2017; 3: e10243.
- Cole BL. Nathan Efron. Charismatic academic returns to Australia. Clin Exp Optom 2005; 88: 430-433.