Browse Subject Areas

Click through the PLOS taxonomy to find articles in your field.

For more information about PLOS Subject Areas, click here.

  • Loading metrics

Productivity trends and collaboration patterns: A diachronic study in the eating disorders field

  • Juan-Carlos Valderrama-Zurián,

    Roles Conceptualization, Data curation, Formal analysis, Investigation, Project administration, Writing – original draft, Writing – review & editing

    Affiliation Universidad Católica de Valencia "San Vicente Mártir", Instituto de Documentación y Tecnologías de Información, Valencia, Spain

  • Remedios Aguilar-Moya,

    Roles Conceptualization, Data curation, Formal analysis, Investigation, Writing – original draft, Writing – review & editing

    Affiliation Universidad Católica de Valencia "San Vicente Mártir", Departamento de Ciencias de la Educación, Godella, Spain

  • Antonio Cepeda-Benito,

    Roles Conceptualization, Data curation, Formal analysis, Investigation, Supervision, Writing – original draft, Writing – review & editing

    Affiliation The University of Vermont, Department of Psychological Science, Burlington, Vermont, United States of America

  • David Melero-Fuentes ,

    Roles Conceptualization, Data curation, Formal analysis, Investigation, Project administration, Writing – original draft, Writing – review & editing

    Affiliation Universidad Católica de Valencia "San Vicente Mártir", Instituto de Documentación y Tecnologías de Información, Valencia, Spain


  • María-Ángeles Navarro-Moreno,

    Roles Conceptualization, Data curation, Formal analysis, Investigation, Writing – original draft, Writing – review & editing

    Affiliation Universidad Católica de Valencia "San Vicente Mártir", Departamento de Ciencias Experimentales y Matemáticas, Valencia, Spain

  • Asunción Gandía-Balaguer,

    Roles Conceptualization, Data curation, Formal analysis, Investigation, Writing – original draft, Writing – review & editing

    Affiliation Universidad Católica de Valencia "San Vicente Mártir", Departamento Ciencias Aplicadas a la Actividad Física y Gestión, Valencia, Spain

  • Rafael Aleixandre-Benavent

    Roles Conceptualization, Data curation, Formal analysis, Funding acquisition, Investigation, Writing – original draft, Writing – review & editing

    Affiliation INGENIO (CSIC-Universidad Politécnica de Valencia) & UISYS (CSIC-Universidad de Valencia), Valencia, Spain

Productivity trends and collaboration patterns: A diachronic study in the eating disorders field

  • Juan-Carlos Valderrama-Zurián, 
  • Remedios Aguilar-Moya, 
  • Antonio Cepeda-Benito, 
  • David Melero-Fuentes, 
  • María-Ángeles Navarro-Moreno, 
  • Asunción Gandía-Balaguer, 
  • Rafael Aleixandre-Benavent



The present study seeks to extend previous bibliometric studies on eating disorders (EDs) by including a time-dependent analysis of the growth and evolution of multi-author collaborations and their correlation with ED publication trends from 1980 to 2014 (35 years).


Using standardized practices, we searched Web of Science (WoS) Core Collection (WoSCC) (indexes: Science Citation Index-Expanded [SCIE], & Social Science Citation Index [SSCI]) and Scopus (areas: Health Sciences, Life Sciences, & Social Sciences and Humanities) to identify a large sample of articles related to EDs. We then submitted our sample of articles to bibliometric and graph theory analyses to identify co-authorship and social network patterns.


We present a large number of detailed findings, including a clear pattern of scientific growth measured as number of publications per five-year period or quinquennium (Q), a tremendous increase in the number of authors attracted by the ED subject, and a very high and steady growth in collaborative work.


We inferred that the noted publication growth was likely driven by the noted increase in the number of new authors per Q. Social network analyses suggested that collaborations within ED follow patters of interaction that are similar to well established and recognized disciplines, as indicated by the presence of a “giant cluster”, high cluster density, and the replication of the “small world” phenomenon—the principle that we are all linked by short chains of acquaintances.


Eating disorders (EDs) are mental disorders with severe consequences [13]. The better known and possibly the most widely studied of the EDs are anorexia and bulimia nervosa, both of which were included in the DSM-III in 1980 [4,5]. Binge eating disorder (BED) was first covered in the DSM-IV in 1994 under the larger rubric of eating disorder not otherwise specified (EDNOS) [67]. BED has since become the most prevalent of the EDs [8] and is now considered a singular ED in the DSM-5 [1]. Other less common eating and feeding related EDs include avoidant/restrictive food intake disorder, pica, rumination disorder, and unspecified feeding or eating disorder [1].

Bibliometric studies can provide a window into the interest and research activity associated with specific scientific questions and medical phenomena. For example, Theander [9] used Medline and PsycINFO to study publication patterns related to EDs from 1960 to 1999 and found an almost 27-fold increase in the number of ED references over the 40-year period (which substantively outpaced the average 4-fold growth observed over the same period of time across all medical fields). Clinton [10] searched the database PsycINFO and replicated Theander’s [9] findings, but over a much larger period of time: 1900 to 2009. Clinton [10] also reported that, relative to other psychiatric disorders, the acceleration of research on EDs peaked in the 1980s and stagnated thereafter. Nonetheless, a recent bibliometric study on EDs, Soh and Walter [11] found substantive growth trends in Medline and PsycINFO from 1970 to 2011 (although these authors focused just on publications related to the cross-cultural aspects of EDs).

Given that research teams now dominate the impact of scientific work across various disciplines [12], we believe it would be important to examine the possible role or influence of scientific collaborations in the observed rapid growth of ED research activity. Given that connections between a set of coauthors can be represented as a social network, social network analysis (SNA) can be applied to study how scientific collaborations originate and evolve in parallel with research publications for one or several academic communities [13]. Therefore, the present study seeks to extend previous bibliometric studies on EDs by including a time-dependent analysis of the growth and evolution of multi-author collaborations and their correlation with ED publication trends. To this end, we propose to examine traditional bibliometric indices along with the use of graph theory, which allows for the examination of co-authorship patterns and social network analyses [1418]. We will focus in the 35-year period starting in 1980, which corresponds to the noted relative peak of research activity acceleration on EDs [10]. In addition, we casted a wider search net than previous ED bibliometric studies by using a larger and more comprehensive set of citation and reference databases.


Search strategy and procedures

We selected Web of Science (WoS) Core Collection (WoSCC) (indexes: Science Citation Index-Expanded [SCIE], & Social Science Citation Index [SSCI]) and Scopus (areas: Health Sciences, Life Sciences, & Social Sciences and Humanities) to conduct our search. Our strategy aimed to retrieve the complete collection of articles related to EDs published between 1980–2014 in the health and social sciences by searching journals included in any of the categories or areas listed in Table 1 (WoS Category) or Table 2 (Scopus Subject Areas). To this purpose, we entered a broad collection of ED-related keywords in the fields “Title” or “Topic” (for WoSCC), as well as in the fields “Title” or “Authkey” or “Indexterms” (for Scopus). The keywords used were: “eating disorder*”, or “pica”, or "rumination disorder*", or "avoidant food intake disorder*", or "restrictive food intake disorder*", or "anorexia nervosa*", or "bulimia nervosa*", or "binge eating disorder*". Each of the data based utilized allow for narrowing the search to specific document types. In both cases we limited our search to “Articles”. The search also excluded any articles that contained any of the 223 animal and insect terms included in the Agricultural Thesaurus and Glossary of the National Agricultural Library of the United States Department of Agriculture ( The exclusionary terms were entered in the Topic field of the WoSCC and in the Title-Abs-Key field of Scopus. This search strategy yielded 18,996 bibliographic records from WoSCC and 22,516 from Scopus.

Table 1. Health Sciences and Social Sciences from WoS Categories (SCIE & SSCI).

The search yielded a total of 41,512 bibliographic records that were included in a relational database to purge redundancies and false positives. Following the pre-stablished procedures described elsewhere (see [19]), a total of 16,789 records were removed due to duplications and overlapping (n = 9,343) or false positives (n = 7,446). The signature variations from single authors of the final sample of 24,723 records were then standardized through a manual review of all signatures, checking institutional affiliations in case of doubt. Our sample is the broadest, most comprehensive, and inclusive to date for the period covered. That is, our approach retrieved almost twice as many references as the second largest bibliometric analysis on ED to date (see Table 3).

Bibliometric and social network analysis

To facilitate the time-dependent evolution of the volume of research activity and other bibliometric characteristics, we collected data on 30 different bibliometric and network measures for each of the seven, 5-year periods, or Quinquennials (Q) of the 1980–2014, time span (see Table 4).


Bibliometric results

Table 5 lists the diachronic progression of productivity measures per 5-year period, or Q. Overall, productivity, measured as the number of articles per Q increased by a factor of about 6 from Q1 (1980–1984) to Q7 (2010–2014). The final sample of 24,723 articles were authored or co-authored by 48,167 investigators, most of whom authored or co-authored a single manuscript (Transience Index = 73.89%, n = 35,593), with a substantive number of them authoring or co-authoring 2 to 9 articles (23.91%, n = 11,515), and very few authoring or co-authoring 10 or more publications (1.25%, n = 1,059). The substantive increase in the number of publications over the 35-year span does not appear to be due to a change in productivity measured categorically as the percentage of authors/co-authors with 1, 2 to 9, and more than 9 articles. That is, across Q2 to Q7, 76.85% to 78.47% of authors published a single article, 20.12% to 21.79% published between 2–9 articles, and 1.09% to 1.65% published 10 or more articles. Thus, the almost 6-fold growth in the number of publication in eating disorders over the 35-year span appears to be driven, on the surface, by the overall growth in the number of active authors per Q, which experienced an almost 10-fold increase from 2,111 in Q1 to 20,145 in Q7. Said differently, the growth of the number of authors/co-authors outpaced the growth of the number of total publications.

The number of authors per Q is the direct sum of newcomers (authors without a publication in the previous Q) plus continuants (authors with at least a publication in the previous and current Q). The number and percentage of newcomers and continuants are given in Table 5, which also shows the overall net gain from Q to Q. Table 5 shows a steady increase in the overall number of authors per Q of about 30% from Q to Q. This steady increase is the result of newcomers consistently outpacing terminators, and to a lesser extent a small but sustained increase in the percentage of continuants that remains from Q to Q (from 11% at Q2 to 17% at Q5, Q6 and Q7) that counteracts the small but sustained decline in the percentage of newcomers that appear from Q to Q (from a high of 89% at Q2 to a low of 83% at Q5, Q6 and Q7).

Table 6 lists the results per Q for each collaboration indicator. A total of 85.2% or 21,065 articles were co-authored by two or more authors, and only 14.8% or 3,658 articles were single authored. Four authors per article was the average over the 35-year period (M = 4.03, SD = 3.11). The percentage of co-authored articles increased with each new Q, from 61.97% in Q1 to 91.73% in Q7. Likewise, the average size and size range of the collaborating teams, or number of authors per article, also increased with each Q, from a low of M = 2.43 (SD = 1.67; range = 1 to 10) at Q1 to a peak of M = 4.79 (SD = 3.81; range = 1 to 176) in Q7.

Similarly, the relative number of authors who engaged in collaborative efforts also increased from Q to Q, from a minimum 79.20% to a peak of 94,41% of authors publishing exclusively as co-authors in Q1 and Q7, respectively. Consequently, the percentage of authors publishing both single and co-authored articles, or only single authored publications, decreased considerably over time (see Table 6). In fact, only about 4% of authors between 2010 to 2014 could claim they had single authored a publication. Whereas the observed increases in collaboration rates and size of collaborating teams co-occurred with an increase in the total number of published articles noted earlier, the overall number or articles per author (articles:author) steadily declined from Q1 (0.57) to Q7 (0.36).

Social network results

Table 7 shows the diachronic progression of co-authorships through network collaboration measures. The vast majority of authors (n = 46,472, 96.48%) was connected to the network at least within one of the seven 5-year periods. The number of vertices, or count of authors who are connected to at least to one other author in the network, increased from 1,818 in Q1 to 19,766 in Q7, with a mean average increase of 2,991 (SD = 2,045) vertices between consecutive Qs. Likewise, the total number of edges or connections that link any two vertices within the network, increased from 3,280 in Q1 to 93,618 edges in Q7, which corresponds to a mean average increase between consecutive Qs of 15,056 (SD = 15,295).

Interestingly, the number of connections (edges) per author decreased over time, from M = 3.61 (SD = 2.77) in Q1 to M = 1.75 (SD = 3.47) in Q7 (although this average appears to have been stabilized over the last decade; M = 1.69 to 1.75; SD = 2.75 to 3.47). Consequently, network density decreased from 0.2679 at Q1 to 0.0019 at Q7.

The average, shortest distance between the vertices of the various network components initially increased from M = 2.79 to M = 7.78 edges (Q1 to Q4), but then decreased to M = 5.38 edges (Q4 to Q7). That is, it appears that from 1980 to 1999 collaboration networks grew with little interconnectivity among vertices, but from 2000 to 2014 the interconnectivity between co-authors increased.

Regarding vertices, the percentage of vertices per Q that shared a single article in common varied from about 83% to 91% but without following a time-line dependent pattern. Likewise, the proportion of vertices that authored between 2 to 9 articles jumped considerably across Qs without following an ascending or descending time pattern, from a minimum of 9.24% at Q1 and a maximum of 16.58% at Q4. The only time-dependent progression observed for productivity between vertex pairs was for those coauthors sharing 10 or more publications, which ranged between 0.09% and 0.26% during the first four Qs, and between 0.38% to 0.44% during the Qs representing the most recent 15 years.

The number of components and the number of vertices within the most populated component increased with each Q, from n = 414 components and n = 110 vertices (Q1) to n = 1,901 components and n = 9,992 vertices (Q7). Likewise, the average number of vertices per component increased from M = 4.39 (SD = 6.6) to M = 10.4 (SD = 10.4) from Q1 to Q7, respectively. In contrast, the other two measures of central tendency, the median and the mode, remained stable across Qs (Median ranged from 3 to 4; Mode ranged from 2 to 3). That is, whereas the mean average and the range of the number of vertices per component has increased considerably over time, the most common component or group size of connected authors has remained relatively small at about 3 vertices. The low number of vertices per component has resulted in high interconnectivity within the components, which have averaged clustering coefficients between 0.94 (Q1) and 0.91 (Q4).

Fig 1 depicts the components’ evolution of the social network Q by Q. Each vertex is represented by a gray dot and each pair of vertices are connected by a gray line. As the number of vertices in a component increases, and as the members within a component become more interconnected (denser) the dots become darker and larger in diameter. When vertices from different components connect with each other, they merge and become very large, and more central to the larger social network. The density (darkness) of the merged components increases as the density or interconnectivity between the vertices of the merged component increases. The figure clearly depicts the progressive increase in vertex density of the central components, as well as a relative increase in the density of components with few vertices per component in the periphery. In addition, Fig 1 shows that, starting with Q4, the central components began to connect with (absorb) the peripheral components.

The distribution of components per vertex per Q revealed that with each successive Q, components become larger and larger components become more frequent. With time, the largest component eventually takes a much larger share of the universe of vertices within the network. For example, the largest component of the network takes up about 6% of all the vertices in Q1 and Q2; however, this proportion consistently increases Q to Q from 28% to 51% from Q3 to Q7 (see Fig 2). Conversely, the combined sum of all the vertices made up by all components with five or fewer co-authors takes up more than 50% of all the vertices at Q1, but this combined sum only contributes 21% of the vertices by Q7 (see Fig 2). A similar phenomenon can be seen for the share of publications attributable to the largest component. In Q1, the combined sum of all components with five or fewer co-authors publishes more than 50% of all articles; whereas the largest component only publishes 8.4% of the articles (see Fig 3). However, by Q3 the largest component takes a larger share of the publications than all the components with five or fewer co-authors combined; and starting with Q5 (2000–2004), the single largest component publishes more than 50% of all the articles (Fig 3).

Fig 2. Within each Quinquennial, share of the total number of vertices taken by the largest component (dashed line) and the sum of components with 5 or fewer co-authors.

Fig 3. Within each Quinquennial, share of the total number of publications by the largest component (dashed line) and all components with 5 or fewer coauthors (solid line).


Using traditional bibliometric indexes of productivity such as number of peer reviewed publications and number of published authors, our results showed a steady net increase in the productivity (number of publications) of scientific research from the first Q (1980–1984) to the most recent Q (2010–2014). Given that the period of time covered in our study overlaps but also extends three years beyond the most recent bibliometric study on EDs (i.e., [11]), our findings contradict Clinton [10] in that we can affirm that the increased interest in the scientific study of EDs that started in the 1960s remains strong beyond the first decade of the 21st Century (see also [12, 21, 22, 28]).

The progressive increase in the number the scientific articles published during the overall 35-year period studied is likely the combined effect of a number of factors: (a) a possible increase in the global prevalence of EDs [29, 30], (b) increase in the awareness or recognition of eating disorders as a grave public health problem [31], and (c) other factors that may influence the augmentation of scientific activity regardless of scientific discipline [32]. Whereas our analyses cannot parse out all the specific factors that contributed to the noted increase in scientific productivity in ED, we can provide important observations about the extent to which individual and collaborative efforts may have accounted for the noted diachronic increase in the overall scientific activity in ED [12].

We can confidently rule out changes in individual author productivity (articles per author) as responsible for the noted augmentation in scientific activity. First, the Transience Index (percentage of authors with a single publication), as well as the proportions of authors with 2 to 9 and more than 10 publications did not substantively vary from Q2 to Q7 (see Table 5). In fact, the noted increase in overall scientific productivity occurred in lieu of a gradual decline in the ratio of articles to authors.

Over the 35-year period examined, there was a 10-fold increase in the number of published authors and 6-fold increase in the number of publications per Q. Thus, we can assert with confidence that the number of published authors correlated positively with the number of publications. This increase in the number of authors was the result of author incorporation rates (newcomers) that outpaced author attrition rates (terminators), and a small but sustained increase in the longevity of the existing authors (continuants).

Among the many possible factors that may have enabled or contributed to the noted remarkable growth in the net number of publications and authors per Q, we can point to the transformation or development of collaborations or author networks. To our knowledge, our study is the first to document within ED the dominant tendency in science toward multi-authored publications in detriment of sole authorships [33]. This can be observed in the increase in the percentage of authors who never published single authored articles, which grew from 79% to almost 96%, as well as in the decline of both authors with only single authored publications (14% to 2%) and authors with both single and co-authored publications (7% to 2%) (see Table 6). Thus, it is not surprising that the overall increase in the number of scientific publications observed occurred in parallel with the growth of works produced in collaboration. This noted tendency toward collective authorships, or increase in the average number of signatures per article, replicates results found in other disciplines including drug addiction research (e.g., [34]). For example, transience index values higher than 70% have been observed in Biomedicine and the Social Sciences [3538].

The observed time-dependent growth in the number of signatures per article might be a response to the increasing need or demand to investigate phenomena from an interdisciplinary perspective to ensure a more complete and comprehensive advancement of scientific knowledge [3941]. A high degree of collaboration within a discipline could be an indicator of scientific strength and vibrancy, as collaborations may improve both the quantity [35] and quality [42] of the research findings, elements that are all indispensable for the development of science and the advancement of knowledge [43]. Unfortunately, our database does not allow us to investigate with precision the extent to which the collaborations noted in ED are interdisciplinary, a question that deserves exploration in future research.

The network analyses combined with the collaboration measures provide a number of interesting observations. Most notably, the results show that the detected interest in collaborative work leads to the proliferation of research components with a high degree of clustering (i.e., high interconnectivity within individual components), in part because most article collaborations involve a reduced number of coauthors within relatively small components. In addition, our analyses suggest the presence of a mature, co-author network of ED that is similar other well developed and stablished networks across various disciplines.

Although the number of edges grew proportionally at a higher rate than the number of authors, the density of the components decreased over time. Nonetheless, the average distance among reachable pairs in the last 5 Qs is similar to those reported by Bordons and collaborators [44] in a study that examined social network activity between 2006–2008 within various disciplines. This means that the average distance, or minimum number of pathways need to connect two unlinked authors is relatively short within the ED network (see also [45]).

According to Fatt and collaborators [46] the component formed by the largest number of authors (or giant component) might be considered the network nucleus of any given of research area. The relative size of the largest component matters, as the size is an indicator of the existence of a “core” field in the research community, and an indicator that knowledge flows a relatively fast pace (see [13]). In general, there appears to be strong consensus that well- formed giant formed giant components include more than 40 to 90% of the vertices [13]. Within the present study we have noted and illustrated that not only the nucleus grew considerably from 6% (1980–1984) to over 50% (2010–2014), but that the producivity attributed to the nucleus surpassed the 50% mark by the fifth Q (1995–1999).

The high clustering coefficients (0.91 and 0.94) within each Q suggest that throughout the entire 35-year period there is high interconnectivity between the vertices, which is made possible by a high percentage of small clusters and the existence of structures that recreate the small-world phenomenon (The small-world phenomenon refers to the often found fact that any two individuals within any given social network are most likely connected by a short sequence of intermediate acquaintances) within the giant and larger components [26]. As reported by others [4749], authors tend to create collaboration groups that connect fully all of its members. It has been argued that high clustering within the Health and Social Sciences is a necessary condition for the advancement of knowledge [50]. Our findings are similar to those reported by Barabásis and collaborators [15] with regard to the field of Neuroscience, or by Bordons and collaborators [44], for various disciplines.


Although our search approach yielded the largest sample or ED-related articles to date, our sample likely excluded a not insignificant number on non-English journals not covered by the WoS Core Collection database. This limitation may have been lessened by our decision to also mine Scopus, which includes a wide coverage of journals in languages other than English [51]. It is also possible but unlikely that we may have failed to include journals that might be covered by other databases such as PsycINFO. Nonetheless, WoS Core Collection and Scopus are the databases used most often in bibliometric studies because they include scientific literature of greater international impact [52] and other national databases play a more peripheral role in communicating the scientific literature [53].

Our search strategy may have been narrower than we portray it because we limited our search keywords to feeding and eating disorders terms, but excluded other eating disorder-related terms such as “restrictive eating” or “purging behaviors” or “body image disturbance”. We therefore checked the impact of leaving out the terms noted above and found we potentially missed adding up to 154 articles to our final sample (or a 0.35% net increase, assuming none of the 154 articles were redundant with any of the articles already in our sample). Our search strategy may have also been wider than it needed to be. That is, some researchers may find that our inclusion of ‘pica’ and ‘rumination’ disorders, which accounted for about 3% of all the articles in our search, may not be highly relevant to the ED field. However, these disorders are listed in the DSM-5 and previous DSM versions alongside anorexia and bulimia under the “feeding and eating disorder” categorization. Thus, while acknowledging our search could have been broader or more comprehensive in some respects, and narrower in others, we believe that the sum of potentially missed articles and perhaps extraneous articles is too small to have made much of a difference in the overall pattern of findings.

Another limitation is that we could not address or describe the likely interdisciplinary makeup of the ED co-author network, and thus cannot speak to the extent to which attempts to advance knowledge in ED contributed to the observed increase in collaborative work over time.

Future research

As possible future lines of research, we believe that it would be important to conduct more in-depth analyses of the evolution of specific areas within ED contrasted against each other, as well as comparing the relative penetration of ED research across related sciences such as Health Sciences, Education Sciences or Social Sciences, or across associated disciplines and bodies of knowledge, such as epidemiology, clinical, pathology and prevention. It might also be important to examine in greater depth the factors or motivations that drive researchers to collaborate, as for example the hypothesized desire and need to understand phenomena fully by incorporating interdisciplinary perspectives.

We hope our investigation inspires other lines of inquire, such a more direct examination of how collaborative research across investigators and research centers might be associated with external funding success, or with the ability to publish in high impact journals. Similarly, of great interest to the profession, at a time when interdisciplinary collaborations have become a buzz word, it would important to study how social networks contribute or become barriers to promotion or tenure within universities and research institutions.

Finally, EDs, once conceptualized as “culture-bound” syndromes rooted in Western culture [54,55], are or have become more prevalent among non-Western cultural groups and nations that previously thought [56]. Although not within the scope of the current study, it would be interesting and insightful to investigate the extent to which the realization that EDs are not phenomena circumscribed to Anglo/Western cultures is the result of increased collaborations across investigators across borders, or vice versa [30].


  1. 1. American Psychiatric Association. American Psychiatric Association Diagnostic and Statistical Manual of Mental Disorders, 5th Edition, DSM-5. Washington, DC: American Psychiatric Association; 2013.
  2. 2. McClelland J, Bozhilova N, Campbell I, Schmidt U. A systematic review of the effects of neuromodulation on eating and body weight: evidence from human and animal studies. European Eating Disorders Review. 2013; 21:436–455. pmid:24155246
  3. 3. Long JC, Cunningham FC, Carswell P, Braithwaite J. Patterns of collaboration in complex networks: the example of a translational research network. BMC Health Services Research. 2014; 14. pmid:24885971
  4. 4. American Psychiatric Association. American Psychiatric Association Diagnostic and Statistical Manual of Mental Disorders, 3rd Edition, DSM-III. Washington, DC: American Psychiatric Association; 1980.
  5. 5. Lancelot C, Brooks-Gunn J, Warren MP, Newman DL. Comparison of DSM-III and DSM-III-R bulimia nervosa classifications for psychopathology and other eating behaviors. International Journal of Eating Disorders. 1991; 10:57–66.
  6. 6. Wonderlich SA, Crosby RD, Joiner T, Peterson CB, Bardone-Cone A, Klein M, et al. Personality subtyping and bulimia nervosa: psychopathological and genetic correlates. Psychological Medicine. 2005; 35:649–657. pmid:15918341
  7. 7. Spitzer RL, Devlin MJ, Walsh BT, Hasin D, Wing R, Marcus MD, et al. Binge eating disorder: To be or not to be in DSM-IV. International Journal of Eating Disorders. 1991; 10:627–629.
  8. 8. Wonderlich SA, Gordon KH, Mitchell JE, Crosby RD, Engel SG. The Validity and Clinical Utility of Binge Eating Disorder. FOCUS. 2014;12(4):489–505.
  9. 9. Theander SS. Literature on Eating Disorders during 40 Years: Increasing Number of Papers, Emergence of Bulimia Nervosa. European Eating Disorders Review. 2002; 10:386–398.
  10. 10. Clinton D. Towards an ecology of eating disorders: creating sustainability through the integration of scientific research and clinical practice. European Eating Disorders Review. 2010; 18:1–9. pmid:20054876
  11. 11. Soh NL-W, Walter G. Publications on cross-cultural aspects of eating disorders. Journal of Eating Disorders. 2013; 1:4. pmid:24764527
  12. 12. Wuchty S, Jones BF, Uzzi B. The increasing dominance of teams in production of knowledge. Science. 2007 May 18;316(5827):1036–9. pmid:17431139
  13. 13. Kumar S. Co-authorship networks: a review of the literature. Aslib Journal of Information Management. 2015; 67:55–73.
  14. 14. Van Raan AFJ. Evaluating the Scientific Excellence of Research Programmes: a Pivot of Decision-Making. The IPTS Report. 1999; 40:30–37.
  15. 15. Barabásis AL, Jeong H, Néda Z, Ravasz E, Dchubert A, Vicsek T. Evolution of the Social Network of Scientific Collaborations. Physica A. 2002; 311:590–614.
  16. 16. Newman MEJ. Coauthorship networks and patterns of scientific collaboration. Proceedings of the National Academy of Sciences. 2004; 6:5200–5205.
  17. 17. Aleixandre-Benavent R, Alonso-Arroyo R. Indicadores bibliométricos, patología del aparato respiratorio y reducción del consumo de tabaco. Revista de Patología Respiratoria. 2011; 14:1–3.
  18. 18. Pino-Díaz J, Jiménez-Contreras E, Ruíz-Baños R, Bailón-Moreno R. Evaluación de redes tecnocientífi cas: la red Española sobre Áreas Protegidas, según la Web of Science. Revista Española de Documentación Científica. 2011; 34:301–333.
  19. 19. Valderrama-Zurián JC, Aguilar-Moya R, Melero-Fuentes D, Aleixandre-Benavent R. A systematic analysis of duplicate records in Scopus. Journal of Informetrics. 2015; 9:570–557.
  20. 20. Guardiola-Wanden-Berghe R, Sanz-Valero J, Wanden-Berghe C. Medical subject headings versus American Psychological Association index terms: indexing eating disorders. Scientometrics. 2013; 94:305–311.
  21. 21. Soh N, Walter G, Touyz S, Russell J, Malhi GS, Hunt GE. Food for Thought: Comparison of Citations Received from Articles Appearing in Specialized Eating Disorder Journals versus General Psychiatry Journals. International Journal of Eating Disorders. 2012; 45:990–994. pmid:22729837
  22. 22. Theander SS. Trends in the Literature on Eating Disorders over 36 Years (1965–2000): Terminology, Interpretation and Treatment. European Eating Disorders Review. 2004; 12:4–17.
  23. 23. Kawamura M, Thomas CDL, Tsurumoto A, Sasahara H, Kawaguchi Y. Lotka’s law and productivity index of authors in a scientific journal. Journal of Oral Science. 2000; 42:75–78. pmid:10989588
  24. 24. Price DJS, Gürsey S. Studies in scientometrics. Part I: Transience and continuance in scientific authorship. International Forum on Information and Documentation. 1976; 1:17–24.
  25. 25. Lawani SM. Quality, collaboration and citations in cancer research: A bibliometric study. PhD thesis. Florida State University, Tallahassee. 1980.
  26. 26. Watts DJ, Strogatz SH. «Collective dynamics of 'small-world' networks». Nature. 1998; 393:409–410. pmid:9623993
  27. 27. Jacomy M, Venturini T, Heymann S, Bastian M. ForceAtlas2, a Continuous Graph Layout Algorithm for Handy Network Visualization Designed for the Gephi Software. PLoS ONE. 2014; 9:e98679. pmid:24914678
  28. 28. Aguinaga M, Fernández LJ, Varo JR. Trastornos de la conducta alimentaria. Revisión y actualización. Anales Sistema Sanitario de Navarra. 2000; 23:279–292.
  29. 29. Pike KM, Dunne PE. The rise of eating disorders in Asia: a review. Journal of eating disorders. 2015 Sep 17;3(1):33.
  30. 30. Qian J., Hu Q., Wan Y., Li T., Wu M., Ren Z., & Yu D. (2013). Prevalence of eating disorders in the general population: a systematic review. Shanghai archives of psychiatry, 25(4), 212. pmid:24991159
  31. 31. El Ghoch M, Soave F, Calugi S, Dalle Grave R. Eating Disorders, Physical Fitness and Sport Performance: A Systematic Review. Nutrients. 2013; 5:5140–5160. pmid:24352092
  32. 32. Jones AW. The distribution of forensic journals, reflections on authorship practices, peer review and role of the impact factor. Forensic Science International. 2007; 165:115–128. pmid:16784827
  33. 33. Bordons M, Gómez I. Collaboration networks in science. In: Cronin B, Atkins HB, editors. The web of Knowledge: a festschrift in honor of Eugene Garfield. Medford: Information Today; 2000. pp. 197–213.
  34. 34. Baker TB, Hatsukami DK, Lerman C, O’Malley SS, Shields AE, Fiore MC. Transdisciplinary science applied to the evaluation of treatments for tobacco use. Nicotine and Tobacco Research. 2003; 5:s89–99. pmid:14668089
  35. 35. González-Alcaide G, Melero-Fuentes D, Aleixandre-Benavent R, Valderrama-Zurián JC. Productivity and collaboration in scientific publications on criminology. Journal of Criminal Justice Education. 2013; 24:15–37.
  36. 36. López-Muñoz F, Alamo C, Rubio G, García-García P, Martín-Agueda B, Cuenca E. Bibliometric analysis of biomedical publications on SSRI during 1980–2000. Depression and Anxiety. 2003; 18:95–103. pmid:12964177
  37. 37. González-Alcaide G, Aleixandre-Benavent R, Navarro-Molina C, Valderrama-Zurián JC. Coauthorship networks and institutional collaboration patterns in reproductive biology. Fertility and sterility. 2008; 90:941–956. pmid:18177648
  38. 38. González-Alcaide G, Park J, Huamaní C, Belinchón I, Ramos JM. Evolution of Cooperation Patterns in Psoriasis Research: Co-Authorship Network Analysis of Papers in Medline (1942–2013). PloS one. 2015; 10:e0144837. pmid:26658481
  39. 39. Klein JT. Interdisciplinary needs: the current context. Librarian Trends. 1996; 45:134–154.
  40. 40. Bordons M, Zulueta MA. La interdisciplinariedad en los grupos españoles de investigación en el área cardiovascular. Revista Española de Cardiología. 2002; 55:900–912.
  41. 41. Chan HF, Önder AS, Torgler B. The first cut is the deepest: repeat interactions of coauthorship and academic productivity in Nobel laureate teams. Scientometrics. 2016; 106:509–524.
  42. 42. Rodríguez de Fonseca F. La Investigación en Drogodependencias. Introducción. In: Ballesteros J, Torrens M, Valderrama JC, editors. Manual Introductorio a la Investigación en Drogodependencias. Valencia: Sociedad Española de Toxicomanías; 2006. pp. 9–18.
  43. 43. Agudelo D, Bretón-López J, Buela-Casal G. Análisis bibliométrico de las revistas de Psicología Clínica editadas en castellano. Psicothema. 2003; 154:507–516.
  44. 44. Bordons M, Aparicio J, González-Albo B, Díaz-Faes AA. The relationship between the research performance ofscientists and their position in co-authorship networks in three fields. Journal of Informetrics. 2015; 9:135–144.
  45. 45. Newman MEJ. The structure of scientific collaboration networks. Proceedings of the National Academy of Sciences. 2001; 98:404–409.
  46. 46. Fatt CK, Abu Ujum E, Ratnavelu K. The structure of collaboration in the journal of finance. Scientometrics. 2010; 85:849–860.
  47. 47. Kretschmer H. Author productivity and geodesic distance in bibliographic co-authorship networks, and visibility on the Web. Scientometrics. 2004; 60:409–420.
  48. 48. Yan EJ, Ding Y, Zhu QH. Mapping library and information science in China: a coauthorship network analysis. Scientometrics. 2010; 83:115–131.
  49. 49. Yin LC, Kretschmer H, Hanneman RA, Liu ZY. Connection and stratification in research collaboration: an analysis of the COLLNET network. Information Processing & Management. 2006; 42:1599–1613.
  50. 50. Lambiotte R, Panzarasa P. Communities, knowledge creation, and information diffusion. Journal of Informetrics. 2009; 3:180–190.
  51. 51. Leydesdorff L. World shares of publications of the USA, EU-27, and China compared and predicted using the new Web of Science interface versus Scopus. El profesional de la Información. 2012; 21:43–49.
  52. 52. Bartol T, Budimir G, Dekleva-Smrekar D, Pusnik M, Juznic P. Assessment of research fields in Scopus and Web of Science in the view of national research evaluation in Slovenia. Scientometrics. 2014; 98:1491–1504.
  53. 53. López Illescas C, Moya Anegón F, Moed HF. Coverage and citation impact of oncological journals in the Web of Science and Scopus. Journal of Informetrics. 2008; 2:304–316.
  54. 54. Warren CS, Gleaves DH, Cepeda-Benito A, Fernandez MDC, Rodriguez-Ruiz S. Ethnicity as a protective factor against internalization of a thin ideal and body dissatisfaction. International Journal of Eating Disorders. 2005; 37(3):241–249. pmid:15822090
  55. 55. Prince R, Thebaud EF. Is anorexia nervosa a culture-bound syndrome?. Transcultural Psychiatric Research Review. 1983; 20(4):299–302.
  56. 56. Miller MN, Pumariega AJ. Culture and eating disorders: A historical and cross-cultural review. Psychiatry: Interpersonal and biological processes. 2001; |64(2):93–110.