Exploiting Literature-Based Discovery to Study Efects of Bullying

Abstract

Bullying represents an aggressive behavioral problem. It can create severe mental illness in children, adolescents, and adults. Researchers are interested in studying bullying to understand its long-term negative consequences and design better-targeted intervention programs. Broadly researchers in psychology and education have conducted studies on bullying without using advanced analytical features ofered by the computer science domain. However, researchers could benefit from power of advanced analytics for an in-depth understanding of bullying. In this paper, we present graphbased advanced analytics by employing a literature-based discovery (LBD) approach to study the direct and indirect consequences of bullying from published literature. In addition, we describe a graph-based study on the evolution of research on bullying consequences in the last two decades. Our graph-based study reveals a summary view of associations of bullying with other diseases and disorders from published literature. Moreover, we employ literature reference information to quantify the strength of the graph-derived relationships. Furthermore, we propose an interactive Graphical User Interface (GUI) design to study bullying. We demonstrate that LBD is a valuable approach to perform advanced analytics on bullying that helps us to understand the severity of its consequences.

CCS CONCEPTS

• Software and its engineering → Software creation and management. bullying, literature-based discovery, graph analytics IUI Workshops’19, March 20, 2019, Los Angeles, USA Copyright © 2019 for the individual papers by the papers’ authors. Copying permitted for private and academic purposes. This volume is published and copyrighted by its editors.

1 INTRODUCTION

Bullying is an abuse delivered in the form of threatening, demeaning, and belittling a person by a physically stronger individual or one with a higher social status than the victim [ 18 ]. Children, adolescents, and even adults are victims of bullying [ 21 ]. However, young people are at the highest risk [ 1 ]. Bullying can create serious mental distress [ 10 ]. Moreover, combined with additional risk factors, being involved in bullying, increases the possibility of engaging in suicide-related behaviors [ 10 ]. Furthermore, not only the victims but the atmosphere is afected by bullying. Research studies show that individuals who experienced bullying behavior problems at school age have a tendency to participate in criminal behavior as adults, including racial attacks, damage to public property, sexual attacks and even killing [ 8 ].

Bullying is an increasingly interesting topic for researchers, because of growing evidence demonstrating long-term negative outcomes for both bullies and victims. The researchers aim to identify the causes of bullying in order to find efective intervention programs targeting the risk and protective factors related to the involvement in bullying. However, it is dificult to study bullying behavior because it is frequently covered up or fabricated (eheck [ 22 ] for detail) [ 36 ]. Historically, experts in psychology and education have conducted studies on bullying but other fields such as computer science, big data, and virtual communities can provide a significant contribution by ofering new perspectives and innovative approaches to studying bullying and developing eficient bullying interventions [ 15 ].

Despite the fact that the bully-victim problem had existed previously, the first systematic studies were only conducted in the early 1970s [ 26, 34 ]. In the early stage, studies were performed in Scandinavia, US, Japan, Australia, and UK in the, 1970s, 1980s and 1990s [ 27 ]. As stated in US national data, approximately 15% of youngsters experienced cyberbullying and 20% were bullied at school in 2013. Moreover, 8% of students reported being bullied daily [ 15 ]. The rate of bullying among children is fairly consistent; nevertheless, researchers have found bullying among adults as well [ 29 ]. This phenomenon is particularly prominent in the work environment. According to the “Workplace Bullying Survey 2014,” 72% of Americans reported being aware of bullying in the workplace, 21% witnessed it, and 27% sufered abusive conduct at work [ 37 ].

In this paper, we present a summarized view of bullying’s direct and indirect consequences. The literature-based discovery (LBD) method is employed to identify bullying’s short-term and long-term impacts from the published literature. Moreover, we present a systematic discussion on how research on bullying consequences evolves over time. We utilize Semantic MEDLINE, a web application that manages the outcomes of PubMed searches by visualizing SemRep predications collected from MEDLINE citations [ 24 ]. We demonstrate that LBD is a useful approach for domain researchers, to discover logical relationships between bullying and its harmful impacts.

We review previous work on LBD in Section 2. In Section 3, we present our research methodology. In Section 4, we provide a detail discussion of our research results. In Section 5, we discuss propose user interface design and Section 6 concludes the paper. 2

BACKGROUND

Literature-based discovery (LBD) uses published literature to identify new interesting relationships between existing knowledge regarding a given phenomenon [ 33 ]. One of the early work on LBD was conducted by Don R. Swanson. The author employed LBD technique to study fish oil [ 33 ]. Swanson apply LBD to find the connection between Migraine and Magnesium as well [ 12 ]. Hristovski, Friedman, Rindflesch, and Peterlin proposed to use semantic predication to increase the strength of the LBD [ 17 ]. Miller et al. employed LBD to do hypothesis testing. In their study, authors found that testosterone and sleep have a connection via cortisol [ 24 ]. Gordon and Lindsay utilized LBD to find the link between Raynaud’s disease and dietary fish oil [ 13 ]. 3

METHODOLOGY

There are two main components of our research. First, we study the direct and indirect efects of bullying. Second, we examine the evolution of the patterns depicted in bullying consequences in published literature. We employ semantic MEDLINE tool to conduct our research. We present description of the semantic MEDLINE and our research components below. Throughout the paper, Unified Medical Language System (UMLS) concept names are italicized.

Semantic MEDLINE

Semantic MEDLINE is a web-based application that provides biomedical document retrieval, summarization, and visualization services [ 30 ]. This application uses MEDLINE citations generated from PubMed searches [ 19, 30 ]. Semantic MEDLINE employs a natural language processing tool called SemRep to find semantic relationships (or predications) from the titles and abstracts of PubMed citations [ 38 ]. SemRep examines each sentence in PubMed titles and abstracts to extract meaningful relationships [ 30 ]. For example, SemRep obtains the relation psychological distress COEXISTS_WITH bullying from the following citation “The Impact of School Social Support and Bullying Victimization on Psychological Distress among California Adolescents (PMID:27708555). However, all the SemRep relations are not meaningful. For example, Semantic MEDLINE shows that citation [ 14 ] creates bullying and vitelliform dystrophy relationship. However, there exists no such relationship in citation [ 14 ]. Thus, it is necessary for us to read the citations carefully to find out the soundness and strength of the relationships.

SemRep relationships are presented in a triple form as it contains three parts: subject, predicate, and object. In the above example, the subject is psychological distress, the predicate is COEXISTS_WITH, and the object is bullying. Subjects and objects represent concepts taken from the UMLS’s Metathesaurus, and predicates are collected from the Semantic Network (SN) [ 24 ]. Semantic MEDLINE provides visualization of the relationships as a graph (or network) by considering the subject and object as nodes, and the predicate as an edge. The graph edge provides citation information and the sentence from which the relationship is extracted. We present an example visualization in Fig. 1 where psychological distress and bullying maintain a COEXISTS_WITH relationship (left side of the figure). The citation sentence from which the relationship is extracted is shown in red on the right side of the figure. Semantic MEDLINE provides numerous node types and relations (edges) based on a citation search. We refer readers to the SemMed documentation for additional discussion on this topic [ 4 ].

Find Efects of Bullying

We utilize semantic MEDLINE to discover direct and indirect efects of bullying. We use the free UMLS license to access the semantic MEDLINE.

To find the direct efects of bullying, we query semantic MEDLINE with the keyword “bullying” and pursue the following steps:

Step 2: We summarize the citation by using Summary Type: Treatment and Disease and UMLS concept that occur in the relations the maximum number of times. We also check the “More Relations” checkbox to utilize all the relations. Summarization provides us relations in a subject, predicate, and object format. It also presents the citation sentence (from where the relation is extracted) and PMID.

Step 3: We visualize the summarized results in a graph where nodes (UMLS concepts) are the subjects and objects, while edges are predicates (from the Semantic Network). We use the COEXISTS_WITH relation in our study.

Step 4: We obtain the corresponding citations for each relation from the graph edge. Finally, we read the citation to identify meaningful relations.

To find indirect efects of bullying, we first search semantic MEDLINE with all the UMLS concepts that co-occur with “bullying” (example: Fig. 2-Level 2) and follow the four steps mentioned earlier. Next, with the results obtained in the first phase, we search semantic MEDLINE again (see example: Fig. 2-Level 3) and repeat the above four steps.

Exploring the Evolution of Research on Bullying

We aim to understand the evolution of bullying research over time. We consider COEXISTS_WITH relationship. We observe from semantic MEDLINE that first UMLS concept COEXISTS_WITH bullying in the year 1995. Next, from 1995, we expand our search date by five-year periods leading to bullying COEXISTS_WITH graph for the years 1995, 2000, 2005, 2010, 2015, and 2017 (Semantic MEDLINE provides data up to the year of 2017). We generate six graphs with numerous relations. To quantify the importance of a relationship, we search Google Scholar [ 2 ] with relationship’s citation title(s) to collect “Cited By” (another citation refers to a relationship citation) information.

Note: “Cited By” information is one of importance measurement metric. However, it may not be appropriate for some cases. 4

DISCUSSION

Efects of Bullying

In the following section, we present our discussion on direct and indirect efects of bullying based on literature-based discovery. We query semantic MEDLINE with keyword “bullying” and use the date range from 01/01/1900 to 06/30/2017, which provides the recent citations on bullying. We obtain many relationships from semantic MEDLINE. Furthermore, we discuss few relations as examples.

Direct Efect . We observe that bullying concept appears directly with the following concepts: 1) suicide, 2) alcohol and other drugs (AOD) use, 3) metabolic syndrome, 4) psychological distress, 5) inflammation , 6) family conflict , and 7) victimizing others. The relationship between bullying and suicide comes from the citation [ 32 ] where authors study “why bullying leads to suicidal thoughts and behaviors.” The study performed on 340 depressed adolescents and their ages were 13-19 years. In this study, it was observed that controlling for internalizing symptoms, age, overt and reputational bullying maintains an association with periodic past month suicide attempts. Moreover, the authors noted that for male victims, victimization leads to risky behavior engagement before suicide attempts whereas for female victims, bullying was directly related to suicide attempts [ 32 ]. Bullying and AOD use relation comes from citation [ 16 ]. In their study [ 16 ], the authors found that in spite of the severity of internalizing problems, substance use maintains independent connections with both negative peer norms and bullying. The citation [ 23 ] provides bullying’s co-occurrence with metabolic syndrome and inflammation . The citation [ 23 ] describes that childhood bullying has no relationship with metabolic syndrome and inflammation . Hence, it means that not all the relationships present in semantic MEDLINE are positively correlated. The relationship between bullying is directly connected to psychological distress, and bullied young people are susceptible to experience serious psychological distress (SPD) twice more than non-bullied young people [ 39 ]. Bullying is also connected to family conflict via psychological distress [ 35 ]. Many citations (e.g., [ 28 ], [ 9 ]) in semantic MEDLINE suggest that bullying concept maintains ISA (sameAS) relationship with victimizing others. Hence, in our study, we consider them synonymous.

Indirect Efect . In the earlier section (Direct Efect), we mention that bullying has a relation with psychological distress [ 39 ]. We search semantic MEDLINE with psychological distress and ifnd that it has a relationship with at risk for suicide, which is shown in citation [ 6 ]. In citation [ 6 ], the presented results showed that Suicidal behavior is positively associated to psychological distress. Next, we search semantic MEDLINE with at risk for suicide. We find that at risk for suicide maintains relationship with diabetes mellitus suggested by citation [ 11 ]. In the citation [ 11 ], the authors observe that diabetes mellitus is strongly connected with an increase in suicidal behaviors and ideation.

While bullying has relation with psychological distress [ 39 ], psychological distress has a relationship with epilepsy [ 20 ]. In the citation [ 20 ], the authors discussed that psychological distress has significant impact on epilepsy. We search semantic MEDLINE with epilepsy and find that cognitive deficit has a relationship with epilepsy [ 31 ].

We discussed the connection between the bullying and suicide before. Now, we search semantic MEDLINE with concept suicide and find that suicide has relationship with eating disorders [ 7 ]. In the citation [ 7 ], via self-image-based prediction, the authors showed that suicide has association with eating disorders. Subsequently, we search semantic MEDLINE with eating disorders and find that it has a relationship with obesity [ 25 ]. In the citation [ 25 ], the authors demonstrate a connection between eating disorders and obesity by distinguishing “Fear of Obesity” and “Self-Esteem Based on Shape and Weight.” However, their study has some limitations.

Evolution of Bullying Research Literature

Our study shows that bullying research is continuously growing as indicated in Fig. 3, which shows relationship of bullying with other concepts that have been established within various time frames. According to Fig. 3(a), aggressive behavior is the first disorder that COEXISTS_WITH with bullying in the year 1995. The 1995 graph contains only two nodes and one edge. After five years, in the year 2000, in addition to aggressive behavior, bullying is co-occurring with psychiatric problem, victimizing other and social fear (Fig. 3(b)). The 2000 citation graph contains five nodes and three edges. That is, the research work in the time-frame 1995-2000 has established these additional relationships. (b) Bullying research graph from 01/01/1900 to 06/30/2000. (d) Bullying research graph from 01/01/1900 to 06/30/2010. (e) Bullying research graph from 01/01/1900 to 06/30/2015. (f) Bullying research graph from 01/01/1900 to 06/30/2017.

Considering bullying and victimizing other synonymous, as shown in Fig. 3(c), in the year 2005, the number of edges (relationships) doubled from 2000. Moreover, Fig. 3(c) illustrates some of the serious consequences of bullying like violence, learning disability, and depressive disorder. The research efort on bullying increases significantly in the year 2010 as the corresponding graph contains 15 nodes and 17 edges (Fig. 3(d)). Dangerous relations like completed suicide, at risk for suicide, and retinoblastoma are occurred in this graph. Also, Fig 3(d) presents some interesting secondary relationship of bullying. For example, bullying occurs with aggressive behavior and aggressive behavior occurs with obesity and alcohol and other drugs (AOD) use. In the graph for 2005, depressive disorder maintains an indirect relationship with bullying (depressive disorder–victimizing others–violence–bullying). However, in the 2010 bullying graph depressive disorder maintains a direct relationship with bullying. In the year 2015, the graph expands by a large margin, which contains 21 nodes and 24 edges. This graph shows a direct relation between bullying and suicide. We consider it an important relationship because numerous research studies are performed on this relationship. A number of interesting concepts appear in the graph (Fig. 3(e)); in the following we provide some examples: autism spectrum disorder, paranoia, drug abuse, narcissism, etc. Besides, like earlier, in the 2010 bullying graph psychological distress maintains an indirect relationship with bullying (psychological distress–victimizing other–bullying). However, psychological distress maintains a direct relationship with bullying in 2015 graph. In 2017 graph (Fig. 3(f)), we have 21 nodes and 21 edges. In this case, the growth is smaller as the gap between the year 2015 and 2017 is only two years. However, in these two recent years, interesting concepts such as family conflict , self-injurious behavior, chronic disease, paranoid ideation, and inflammation appear in the graph.

From Semantic MEDLINE, we collect citation titles of the relations available (shown in Fig. 3). We search Google Scholar with these titles to find their “cited-by" count. The cited by information is collected on April 3, 2018. The citation with title “Bullying: its roots in pre-school aggression.” represents the single edge in the graph shown in Fig. 3(a). We did not find this citation in Google Scholar. Fig. 3(b)’s top citation title is “Bullying behaviour and psychosocial health among school students in New South Wales, Australia: cross sectional survey”, that established the relation between bullying and psychiatric problem, has been cited 581 times. For 2005 graph (Fig. 3(c)), we identify relationship between victimizing others and depressive disorder very important because the corresponding citation (title: Does bullying cause emotional problems? A prospective study of young teenagers) is cited 1,057 times. We consider the connection between bullying and harassment are important for (Fig. 3(d)) because corresponding citation “Does online harassment constitute bullying? An exploration of online harassment by known peers and online-only contacts” is cited 533 times. Similarly, for Fig. 3(e) the connection between victimizing others and psychiatric problem is important because corresponding citation “Cross-national time trends in bullying victimization in 33 countries among children aged 11, 13 and 15 from 2002 to 2010” is cited 73 times. Finally, for the year 2017 (Fig. 3(f)), we acknowledge Victimizing others and Psychosocial problem is a vital relationship (citation title: “Consequences of bullying victimization in childhood and adolescence: A systematic review and meta-analysis”, cited by 20 times). We observe that many mental diseases and disorders are co-occurring with bullying over the time leading to strong ties between bullying and mental disorders. 5

USER INTERFACE

A Graphical User interface (GUI) can be developed on top of semantic MEDLINE and Google Scholar citation data for interactive data analysis. Semantic MEDLINE data is available in [ 5 ]. “Scholarly” software [ 3 ] can be employed to get Google Scholar data. Next, it is possible to create a Resource Description Framework (RDF) database based on semantic MEDLINE and Google Scholar datasets and a GUI on top of the RDF database. RDF database allows incorporation of ontologies as well.

We propose a web-based system called “Bullying Analytics” which support semantic search, indirect efect finding, and aggregated statistics. We present system designs in Figs. 4, 5, 6.

Semantic Search

Semantic Search permits users to formulate user define queries with a GUI that uses ontology-driven search. Fig. 4 shows how Semantic Search retrieves bullying literature that contains “Female” and “Depression” related terms and the total cited-by count is more than or equal 10. By using “Add Statement” button, researchers can incrementally formulate a complex query which helps to find precise information. Query results (literature title, published year, and cited-by) are presented in the results section.

Pathfinder

This service allows users to find indirect efects of bullying and interactive graph-based visualization from published literature. A user can define hope size as well. Fig. 5 presents how bullying indirect efects can be retrieved where hope size is 3 and relation types are COEXISTS_WITH and CAUSES. Users can click an edge of the graph to get detail publication information. Moreover, users can check whether a edges in the path is correct or not and provide “Correct” or “Incorrect” information through “Feedback” section. Path graph updates based on users feedback. Pathfinder service allows users to ifnd a long chain of indirect efect interactively.

Aggregated Statistics

Fig. 6 shows aggregation service that “Bullying Analytics” provides. Users can select a start and end date. Based on users selection the GUI presents the total number of bullyingrelated citations published by year as a bar chart. Moreover, GUI presents top disease, or disorder related terms occurred in bullying literature as a pie chart (a partical example is provided in Fig. 6).

6 CONCLUSION AND FUTURE WORK

In this paper, we describe an LBD approach to study bullying. We show that the LBD approach is capable of performing advanced analytics on bullying. In this paper, we present a small number of direct and indirect consequences of bullying. Our future work include development of a software tool to ifnd the long chain of meaningful indirect consequences automatically. It is challenging to design a software that can understand literature semantics. We also plan to use graph algorithms (e.g., PageRank) for advanced analytics.

7 ACKNOWLEDGMENTS

This manuscript has been authored by UT-Battelle, LLC under Contract No. DE-AC05-00OR22725 with the U.S. Department of Energy. The United States Government retains and the publisher, by accepting the article for publication, acknowledges that the United States Government retains a nonexclusive, paid-up, irrevocable, world-wide license to publish or reproduce the published form of this manuscript, or allow others to do so, for United States Government purposes. The Department of Energy will provide public access to these results of federally sponsored research in accordance with the DOE Public Access Plan (http://energy.gov/downloads/doepublic-access-plan).

We especially thank Dr. Maleq Khan for his useful guidelines and suggestions that greatly improved the paper.

References

[1] 2018 . Facts About Bullying . URL: https://www.stopbullying.gov/ media/facts/index.html.

[2] 2018. Google Scholar . URL: https://scholar.google.com.

[3] 2018. scholarly 0.2 .4. URL: https://pypi.org/project/scholarly/.

[4] 2018 . Semantic MEDLINE . URL: https://skr3.nlm.nih.gov/SemMed/ jsp/SemMedDocumentation2.pdf.

[5] 2018 . Semantic MEDLINE - Database Download . URL: https://skr3. nlm.nih.gov/SemMedDB/download/download.html.

[6] Bushra Akram and Mubeen Ilyas . 2017 . Coping Strategies, Mental Health and HIV Status: Predictors of Suicidal Behaviour among PWIDs . JPMA 67 , 568 ( 2017 ).

[7] Mikael Andersén and Andreas Birgegård . 2017 . Diagnosis-Specific Self-Image Predicts Longitudinal Suicidal Ideation in Adult Eating Disorders . International Journal of Eating Disorders 50 , 8 ( 2017 ), 970 - 978 .

[8] Onesemus Awiria . 1994 . Bullying at School-What We Know and What We Can Do .

[9] Esther Calvete , Liria Fernández-González, Joaquín M GonzálezCabrera, and Manuel Gámez-Guadix. 2018 . Continued Bullying Victimization in Adolescents: Maladaptive Schemas as a Mediational Mechanism . Journal of Youth and Adolescence 47 , 3 ( 2018 ), 650 - 660 .

[10] CDC . 2014 . The Relationship between Bullying and Suicide: What We Know and What it Means for Schools . Available on the Internet: http://www. cdc. gov/violenceprevention/pdf/bullying-suicidetranslation -final-a . pdf ( 2014 ).

[11] Chiara Conti , Chiara Mennitto, Giulia Di Francesco, Federica Fraticelli, Ester Vitacolonna, and Mario Fulcheri . 2017 . Clinical Characteristics of Diabetes Mellitus and Suicide Risk. Frontiers in Psychiatry 8 ( 2017 ), 40 .

[12] D. R. Swanson . 1988 . Migraine and Magnesium: Eleven Neglected Connections . Perspectives in Biology and Medicine 31 , 4 ( 1988 ), 526 - 557 .

[13] Michael D Gordon and Robert K Lindsay. 1996 . Toward Discovery Support Systems: A Replication, Re-Examination, and Extension of Swanson's Work on Literature-Based Discovery of a Connection Between Raynaud's and Fish Oil . JAIST 47 , 2 ( 1996 ), 116 - 128 .

[14] Amy L Gower , Molly Cousin , and Iris W Borowsky. 2017 . A Multilevel, Statewide Investigation of School District Anti-Bullying Policy Quality and Student Bullying Involvement . Journal of School Health 87 , 3 ( 2017 ), 174 - 181 .

[15] Melissa K Holt , Jennifer Greif Green, Mina Tsay-Vogel, Joanna Davidson , and Claire Brown . 2017 . Multidisciplinary Approaches to Research on Bullying in Adolescence. Adolescent research review 2 , 1 ( 2017 ), 1 - 10 .

[16] Jun Sung Hong , Dexter R Voisin, Sujung Cho, Douglas C Smith , and Stella M Resko . 2017 . Peer Victimization and Substance Use Among African American Adolescents and Emerging Adults on ChicagoâĂŹs Southside . ( 2017 ).

[17] Dimitar Hristovski , Carol Friedman, Thomas C Rindflesch, and Borut Peterlin . 2006 . Exploiting Semantic Relations for Literature-Based Discovery . In AMIA symposium , Vol. 2006 . AMIA, 349 .

[18] Jaana Juvonen and Sandra Graham . 2014 . Bullying in Schools: The Power of Bullies and the Plight of Victims . Annual Review of Psychology 65 ( 2014 ), 159 - 185 .

[19] Halil Kilicoglu , Marcelo Fiszman, Alejandro Rodriguez, Dongwook Shin , A Ripple , and Thomas C Rindflesch. 2008 . Semantic MEDLINE: A Web Application for Managing the Results of PubMed Searches . 2008 ( 2008 ), 69 - 76 .

[20] Iliana Kotwas , Aileen McGonigal , Mireille Bastien-Toniazzo, Fabrice Bartolomei , and Jean-Arthur Micoulaud-Franchi. 2017 . Stress Regulation in Drug-Resistant Epilepsy . Epilepsy & Behavior 71 ( 2017 ), 39 - 50 .

[21] Nicole Krämer , Sabrina Sobieraj, Dan Feng, Elisabeth Trubina, and Stacy Marsella . 2018 . Being Bullied in Virtual Environments: Experiences and Reactions of Male and Female Students to a Male or Female Oppressor. Frontiers in Psychology 9 ( 2018 ), 253 .

[22] Judith MacIntosh. 2005 . Experiences of workplace bullying in a rural area . Issues in Mental Health Nursing 26 , 9 ( 2005 ), 893 - 910 .

[23] Karen A Matthews , J Richard Jennings , Laisze Lee , and Dustin A Pardini. 2017 . Bullying and Being Bullied in Childhood are Associated with Diferent Psychosocial Risk Factors for Poor Physical Health in Men . Psychological Science 28 , 6 ( 2017 ), 808 - 821 .

[24] Christopher M Miller , Thomas C Rindflesch , Marcelo Fiszman, Dimitar Hristovski, Dongwook Shin, Graciela Rosemblat, Han Zhang, and Kingman P Strohl. 2012 . A Closed Literature-Based Discovery Technique Finds a Mechanistic Link Between Hypogonadism and Diminished Sleep Quality in Aging Men . Sleep 35 , 2 ( 2012 ), 279 - 285 .

[25] Tomoyo Mitsui , Toshiyuki Yoshida, and Gen Komaki . 2017 . Psychometric Properties of the Eating Disorder Examination-Questionnaire in Japanese Adolescents . BioPsychoSocial Medicine 11 , 1 ( 2017 ), 9 .

[26] Dan Olweus . 1973 . Hackkycklingar och översittare: Forskning om skolmobbning .

[27] Dan Olweus . 1994 . Bullying at School . In Aggressive Behavior. Springer, 97 - 130 .

[28] Naomi Priest , Tania King, Laia Bécares , and Anne M Kavanagh . 2016 . Bullying Victimization and Racial Discrimination among Australian Children . American Journal of Public Health 106 , 10 ( 2016 ), 1882 - 1884 .

[29] Charlotte Rayner . 1997 . The Incidence of Workplace Bullying . Journal of Community & Applied Social Psychology 7 , 3 ( 1997 ), 199 - 208 .

[30] Thomas C Rindflesch , Halil Kilicoglu, Marcelo Fiszman, Graciela Rosemblat, and Dongwook Shin . 2011 . Semantic MEDLINE : An Advanced Information Management Application for Biomedicine . Information Services & Use 31 , 1 - 2 ( 2011 ), 15 - 21 .

[31] Tanveer Singh , Neetu Bagga , Anureet Kaur, Navjot Kaur, Dinesh Yugraj Gawande, and Rajesh Kumar Goel. 2017 . Agmatine for Combined Treatment of Epilepsy, Depression and Cognitive Impairment in Chronic Epileptic Animals . Biomedicine & Pharmacotherapy 92 ( 2017 ), 720 - 725 .

[32] Jeremy G Stewart , Linda Valeri, Erika C Esposito, and Randy P Auerbach. 2018 . Peer Victimization and Suicidal Thoughts and Behaviors in Depressed Adolescents . Journal of Abnormal Child Psychology 46 , 3 ( 2018 ), 581 - 596 .

[33] Don R Swanson . 1986 . Fish Oil, Raynaud's Syndrome, and Undiscovered Public Knowledge . Perspectives in biology and medicine 30 , 1 ( 1986 ), 7 - 18 .

[34] Swanson , Don R. 1978 . Aggression in the Schools: Bullies and Whipping Boys . Hemisphere.

[35] Rebekah L Thomas , Adrian B Kelly, Gary CK Chan, and Joanne W Williams. 2017 . Linking Eating and Weight Control Attitudes to Relationship Experiences: A Large-Scale Survey of Adolescent Females . Journal of Adolescence 58 ( 2017 ), 67 - 73 .

[36] Anthony A Volk , René Veenstra , and Dorothy L Espelage. 2017 . So You Want to Study Bullying? Recommendations to Enhance the Validity , Transparency, and Compatibility of Bullying Research. Aggression and Violent Behavior 36 ( 2017 ), 34 - 43 .

[37] US WBI . 2014 . Workplace Bullying Survey . Workplace Bullying Institute 8 , 2014 ( 2014 ), 6 .

[38] Han Zhang , Marcelo Fiszman, Dongwook Shin, Bartlomiej Wilkowski, and Thomas C Rindflesch. 2013 . Clustering Cliques for Graph-Based Summarization of the Biomedical Research Literature . BMC Bioinformatics 14 , 1 ( 2013 ), 182 .

[39] Xiaoyan Zhang , Chaelin Karen Ra, Donglan Zhang, Yunting Zhang, and Kara E MacLeod. 2016 . The Impact of School Social Support and Bullying Victimization on Psychological Distress among California Adolescents . Californian Journal of Health Promotion 14 , 2 ( 2016 ), 56 .