Abstract
Cyberbullying and cyberaggression are increasingly worrisome phenomena affecting people across all demographics. More than half of young social media users worldwide have been exposed to such prolonged and/or coordinated digital harassment. Victims can experience a wide range of emotions, with negative consequences such as embarrassment, depression, isolation from other community members, which embed the risk to lead to even more critical consequences, such as suicide attempts.
In this work, we take the first concrete steps to understand the characteristics of abusive behavior in Twitter, one of today’s largest social media platforms. We analyze 1.2 million users and 2.1 million tweets, comparing users participating in discussions around seemingly normal topics like the NBA, to those more likely to be hate-related, such as the Gamergate controversy, or the gender pay inequality at the BBC station. We also explore specific manifestations of abusive behavior, i.e., cyberbullying and cyberaggression, in one of the hate-related communities (Gamergate). We present a robust methodology to distinguish bullies and aggressors from normal Twitter users by considering text, user, and network-based attributes. Using various state-of-the-art machine-learning algorithms, we classify these accounts with over 90% accuracy and AUC. Finally, we discuss the current status of Twitter user accounts marked as abusive by our methodology and study the performance of potential mechanisms that can be used by Twitter to suspend users in the future.
- AllSlang. 2017. List of Swear Words and Curse Words. Retrieved from https://www.noswearing.com/dictionary.Google Scholar
- A. A. Amleshwaram, N. Reddy, S. Yadav, G. Gu, and C. Yang. 2013. CATS: Characterizing automation of twitter spammers. In Proceedings of the 5th International Conference on Communication Systems and Networks (COMSNETS’13). 1--10.Google Scholar
- Anonymous. [n.d.]. “I dated Zoe Quinn.” 4chan. Retrieved from https://archive.is/qrS5Q.Google Scholar
- Anonymous. [n.d.]. Zoe Quinn, prominent SJW and indie developer is a liar and a slut. 4chan. Retrieved from https://archive.is/QIjm3.Google Scholar
- Dzmitry Bahdanau, Kyunghyun Cho, and Yoshua Bengio. 2014. Neural machine translation by jointly learning to align and translate. Arxiv Preprint Arxiv:1409.0473.Google Scholar
- Shane Bergsma, Matt Post, and David Yarowsky. 2012. Stylometric analysis of scientific articles. In Proceedings of the Conference of the North American Chapter of the Association for Computational Linguistics: Human Language Technologies. Association for Computational Linguistics, 327--337.Google Scholar
- Devipsita Bhattacharya and Sudha Ram. 2015. RT News: An analysis of news agency ego networks in a microblogging environment. ACM Trans. Manage. Info. Syst. 6, 3 (2015), 11:1--11:25.Google Scholar
- Jeremy Blackburn, Ramanuja Simha, Nicolas Kourtellis, Xiang Zuo, Matei Ripeanu, John Skvoretz, and Adriana Iamnitchi. 2012. Branded with a scarlet “C”: Cheaters in a gaming social network. In Proceedings of the Conference on the World Wide Web (WWW’12).Google Scholar
Digital Library
- David M. Blei, Andrew Y. Ng, and Michael I. Jordan. 2003. Latent dirichlet allocation. J. Mach. Learn. Res. 3(Jan. 2003), 993--1022.Google Scholar
- V. D. Blondel, J. L. Guillaume, R. Lambiotte, and E. Lefebvre. 2011. The louvain method for community detection in large networks. Stat. Mech.: Theory Exper. 10 (2011).Google Scholar
- Axel Bruns and Stefan Stieglitz. 2013. Towards more systematic twitter analysis: Metrics for tweeting activities. Int. J. Soc. Res. Methodol. 16, 2 (2013), 91--108.Google Scholar
Cross Ref
- Despoina Chatzakou, Nicolas Kourtellis, Jeremy Blackburn, Emiliano De Cristofaro, Gianluca Stringhini, and Athena Vakali. 2017. Mean birds: Detecting aggression and bullying on twitter. In Proceedings of the ACM Web Science Conference (WebSci’17).Google Scholar
Digital Library
- Despoina Chatzakou, Nicolas Kourtellis, Jeremy Blackburn, Emiliano De Cristofaro, Gianluca Stringhini, and Athena Vakali. 2017. Measuring #GamerGate: A tale of hate, sexism, and bullying. In Proceedings of the WWW CyberSafety Workshop.Google Scholar
Digital Library
- Despoina Chatzakou, Nicolas Kourtellis, Jeremy Blackburn, Emiliano De Cristofaro, Gianluca Stringhini, and Athena Vakali. 2017. Hate is not binary: Studying abusive behavior of #GamerGate on twitter. In Proceedings of the ACM Conference on Hypertext and Social Media.Google Scholar
Digital Library
- Despoina Chatzakou, Nicolas Kourtellis, Jeremy Blackburn, Emiliano De Cristofaro, Gianluca Stringhini, and Athena Vakali. 2017. Mean birds: Detecting aggression and bullying on twitter. In Proceedings of the ACM Web Science Conference. ACM, 13--22.Google Scholar
Digital Library
- Despoina Chatzakou, Vassiliki Koutsonikola, Athena Vakali, and Konstantinos Kafetsios. 2013. Micro-blogging content analysis via emotionally-driven clustering. In Proceedings of the International Conference on Affective Computing and Intelligent Interaction (ACII’13).Google Scholar
Digital Library
- Despoina Chatzakou, Nikolaos Passalis, and Athena Vakali. 2015. MultiSpot: Spotting sentiments with semantic aware multilevel cascaded analysis. In Proceedings of the International Conference on Big Data Analytics and Knowledge (DaWaK’15), Vol. 9263. Springer, 337--350.Google Scholar
Cross Ref
- D. Chatzakou and A. Vakali. 2015. Harvesting opinions and emotions from social media textual resources. IEEE Internet Comput. 19, 4 (2015), 46--50.Google Scholar
Digital Library
- Despoina Chatzakou, Athena Vakali, and Konstantinos Kafetsios. 2017. Detecting variation of emotions in online activities. Expert Syst. Appl. 89 (2017), 318--332.Google Scholar
Digital Library
- C. Chen, J. Zhang, X. Chen, Y. Xiang, and W. Zhou. 2015. 6 million spam tweets: A large ground truth for timely twitter spam detection. In Proceedings of the IEEE International Conference on Communications (ICC’15).Google Scholar
- Ying Chen, Yilu Zhou, Sencun Zhu, and Heng Xu. 2012. Detecting offensive language in social media to protect adolescent online safety. In Proceedings of the IEEE Joint International Conference on Privacy, Security, Risk and Trust (PASSAT’12) and Social Computing (SocialCom’12).Google Scholar
Digital Library
- Lucie Corcoran, Conor McGuckin, and Garry Prentice. 2015. Cyberbullying or cyber aggression?: A review of existing definitions of cyber-based peer-to-peer aggression. Societies 5, 2 (2015).Google Scholar
- Cyberbullying Research Center. 2016. Retrieved from https://cyberbullying.org/summary-of-our-cyberbullying-research.Google Scholar
- Cyberbullying Research Center. 2017. Retrieved from https://cyberbullying.org/facts.Google Scholar
- Maral Dadvar, Dolf Trieschnigg, and Franciska Jong. 2014. Experts and machines against bullies: A hybrid approach to detect cyberbullies. In Proceedings of the Canadian Conference on Artificial Intelligence.Google Scholar
Cross Ref
- Jesse Davis and Mark Goadrich. 2006. The relationship between precision-recall and ROC curves. In Proceedings of the Conference on Machine Learning.Google Scholar
Digital Library
- Thomas G. Dietterich. 2000. Ensemble methods in machine learning. In Proceedings of the 1st International Workshop on Multiple Classifier Systems.Google Scholar
Cross Ref
- Karthik Dinakar, Roi Reichart, and Henry Lieberman. 2011. Modeling the detection of textual cyberbullying. Soc. Mobile Web 11 (2011).Google Scholar
- Nemanja Djuric, Jing Zhou, Robin Morris, Mihajlo Grbovic, Vladan Radosavljevic, and Narayan Bhamidipati. 2015. Hate speech detection with comment embeddings. In Proceedings of the Conference on the World Wide Web (WWW’15).Google Scholar
Digital Library
- P. Ekman, W. V. Friesen, and P. Ellsworth. 1982. What emotion categories or dimensions can observers judge from facial behavior? Emot. Hum. Face (1982), 39--55.Google Scholar
- Dennis Fetterly, Mark Manasse, and Marc Najork. 2004. On the evolution of clusters of near-duplicate web pages. J. Web Eng. 2, 228--246.Google Scholar
- Dennis Fetterly, Mark Manasse, and Marc Najork. 2005. Detecting phrase-level duplication on the world wide web. In Proceedings of the 28th Annual International ACM SIGIR Conference on Research and Development in Information Retrieval (SIGIR’05). Association for Computing Machinery.Google Scholar
Digital Library
- Figure Eight. 2019. Retrieved from https://www.figure-eight.com/.Google Scholar
- Jesse Fox and Wai Yen Tang. 2014. Sexism in online video games: The role of conformity to masculine norms and social dominance orientation. Comput. Hum. Behav. 33 (2014).Google Scholar
- Nir Friedman, Dan Geiger, and Moises Goldszmidt. 1997. Bayesian network classifiers. Mach. Learn. 29, 2--3 (1997).Google Scholar
Digital Library
- Maria Giatsoglou, Despoina Chatzakou, Neil Shah, Christos Faloutsos, and Athena Vakali. 2015. Reteeting activity on twitter: Signs of deception. In Proceedings of the Pacific-Asia Conference on Knowledge Discovery and Data Mining (PAKDD’15).Google Scholar
- Dorothy Wunmi Grigg. 2010. Cyber-aggression: Definition and concept of cyberbullying. Austral. J. Guid. Counsel. 20, 02 (2010).Google Scholar
- The Guardian. 2018. Gary Lineker is BBC’s best-paid star and only one not to take pay cut. Retrieved from https://www.theguardian.com/world/2018/jul/11/gary-lineker-bbc-best-paid-star-only-one-not-to-take-pay-cut.Google Scholar
- Joshua Guberman and Libby Hemphill. 2017. Challenges in modifying existing scales for detecting harassment in individual tweets. In Proceedings of the International Conference on System Sciences.Google Scholar
Cross Ref
- Laura D. Hanish, Becky Kochenderfer-Ladd, Richard A. Fabes, Carol Lynn Martin, Donna Denning et al. 2003. Bullying among young children: The influence of peers and teachers. Bully. Amer. Schools 141 (2003), 141--159.Google Scholar
- Trevor Hastie, Saharon Rosset, Ji Zhu, and Hui Zou. 2009. Multi-class adaboost. Stat. Interface 2, 3 (2009), 349--360.Google Scholar
Cross Ref
- Hatebase database. 2017. Retrieved from https://www.hatebase.org/.Google Scholar
- Kaiming He, Xiangyu Zhang, Shaoqing Ren, and Jian Sun. 2016. Deep residual learning for image recognition. In Proceedings of the IEEE Conference on Computer Vision and Pattern Recognition. 770--778.Google Scholar
Cross Ref
- Alex Hern. 2014. Feminist Critics of Video Games Facing Threats in “GamerGate” Campaign. The Guardian. Retrieved from https://www.theguardian.com/technology/2014/oct/23/felicia-days-public-details-online-gamergate.Google Scholar
- Balázs Hidasi, Massimo Quadrana, Alexandros Karatzoglou, and Domonkos Tikk. 2016. Parallel recurrent neural network architectures for feature-rich session-based recommendations. In Proceedings of the 10th ACM Conference on Recommender Systems. ACM, 241--248.Google Scholar
Digital Library
- Gabriel Emile Hine, Jeremiah Onaolapo, Emiliano De Cristofaro, Nicolas Kourtellis, Ilias Leontiadis, Riginos Samaras, Gianluca Stringhini, and Jeremy Blackburn. 2016. A longitudinal measurement study of 4chan’s politically incorrect forum and its effect on the web. Arxiv Preprint Arxiv:1610.03452Google Scholar
- Homa Hosseinmardi, Richard Han, Qin Lv, Shivakant Mishra, and Amir Ghasemianlangroodi. 2014. Towards understanding cyberbullying behavior in a semi-anonymous social network. In Proceedings of the IEEE/ACM International Conference on Advances in Social Networks Analysis and Mining (ASONAM’14).Google Scholar
Cross Ref
- Homa Hosseinmardi, Sabrina Arredondo Mattson, Rahat Ibn Rafiq, Richard Han, Qin Lv, and Shivakant Mishra. 2015. Analyzing labeled cyberbullying incidents on the instagram social network. In Proceedings of the International Conference on Social Informatics (SocInfo’15).Google Scholar
Cross Ref
- Fang Jin, Edward Dougherty, Parang Saraf, Yang Cao, and Naren Ramakrishnan. 2013. Epidemiological modeling of news and rumors on twitter. In Proceedings of the Social Network Analysis Workshop (SNAKDD’13).Google Scholar
Digital Library
- Ji-Hyun K. 2009. Estimating classification error rate: Repeated cross-validation, repeated hold-out and bootstrap. Comput. Stat. Data Anal. 53, 11 (2009).Google Scholar
- Imrul Kayes, Nicolas Kourtellis, Daniele Quercia, Adriana Iamnitchi, and Francesco Bonchi. 2015. The social world of content abusers in community question answering. In Proceedings of the Conference on the World Wide Web (WWW’15).Google Scholar
Digital Library
- Keras. 2017. Retrieved from https://keras.io/.Google Scholar
- Aamera Z. H. Khan, Mohammad Atique, and V. M. Thakare. 2015. Combining lexicon-based and learning-based methods for twitter sentiment analysis. Int. J. Electron. Commun. Soft Comput. Sci. Eng. (2015), 89.Google Scholar
- Kenji Kira and Larry A. Rendell. 1992. A practical approach to feature selection. In Proceedings of the 9th International Workshop on Machine Learning.Google Scholar
- Jon M. Kleinberg. 1999. Hubs, authorities, and communities. Comput. Surveys 31, 4es, Article 5 (1999).Google Scholar
- Kurt Wagner. 2017. Twitter says it’s punishing 10 times more users for being abusive than it was a year ago. Retrieved from https://www.vox.com/2017/7/20/15999636/twitter-safety-abuse-update-suspensions-increase.Google Scholar
- Haewoon Kwak, Jeremy Blackburn, and Seungyeop Han. 2015. Exploring cyberbullying and other toxic behavior in team competition online games. In Proceedings of the SIGCHI Conference on Human Factors in Computing Systems.Google Scholar
- Kyumin Lee, James Caverlee, and Steve Webb. 2010. Uncovering social spammers: Social honeypots + machine learning. In Proceedings of the 33rd International ACM SIGIR Conference on Research and Development in Information Retrieval (SIGIR’10). ACM, New York, NY, 435--442. DOI:https://doi.org/10.1145/1835449.1835522Google Scholar
Digital Library
- A. Massanari. 2015. #Gamergate and the fappening: How Reddit’s algorithm, governance, and culture support toxic technocultures. New Media Soc. 19, 3 (2015), 329--346.Google Scholar
Cross Ref
- M. McCord and M. Chuah. 2011. Spam detection on twitter using traditional classifiers. In Autonomic and Trusted Computing. Springer, Berlin, 175--186.Google Scholar
- Tomas Mikolov, Kai Chen, Greg Corrado, and Jeffrey Dean. 2013. Efficient estimation of word representations in vector space. CoRR abs/1301.3781 (2013).Google Scholar
- Meagan Miller. 2016. Retrieved from goo.gl/n1W6nt.Google Scholar
- Torill Elvira Mortensen. 2016. Anger, fear, and games: The long event of #GamerGate. Games Culture 13, 8 (2016), 787--806.Google Scholar
Cross Ref
- Vinita Nahar, Sayan Unankard, Xue Li, and Chaoyi Pang. 2012. Sentiment analysis for effective detection of cyber bullying. In Proceedings of the Asian Pacific Web Conference (APWeb’12).Google Scholar
Digital Library
- Gonzalo Navarro. 2001. A guided tour to approximate string matching. Comput. Surveys 33, 1 (2001).Google Scholar
- Shirin Nilizadeh, François Labrèche, Alireza Sedighian, Ali Zand, José Fernandez, Christopher Kruegel, Gianluca Stringhini, and Giovanni Vigna. 2017. Poised: Spotting twitter spam off the beaten paths. In Proceedings of the ACM SIGSAC Conference on Computer and Communications Security (CCS’17).Google Scholar
Digital Library
- Chikashi Nobata, Joel Tetreault, Achint Thomas, Yashar Mehdad, and Yi Chang. 2016. Abusive language detection in online user content. In Proceedings of the Conference on the World Wide Web (WWW’16).Google Scholar
Digital Library
- Donie O’Sullivan. 2018. Bomb suspect threatened people on Twitter, and Twitter didn’t act. Retrieved fromxbrk https://edition.cnn.com/2018/10/26/tech/cesar-sayoc-twitter-response/index.html.Google Scholar
- Jeffrey Pennington, Richard Socher, and Christopher D. Manning. 2014. GloVe: Global vectors for word representation. In Proceedings of the Conference on Empirical Methods in Natural Language Processing (EMNLP’14). 1532--1543. Retrieved from http://www.aclweb.org/anthology/D14-1162.Google Scholar
- Pew Research Center. 2014. Retrieved from http://www.pewinternet.org/2017/07/11/online-harassment-2017/.Google Scholar
- J. Pfeffer, T. Zorbach, and K. M. Carley. 2014. Understanding online firestorms: Negative word-of-mouth dynamics in social media networks. J. Market. Commun. 20, 1--2 (2014).Google Scholar
Cross Ref
- Pham, Sherisse. 2017. Twitter tries new measures in crackdown on harassment. CNNtech. Retrieved from https://money.cnn.com/2017/02/07/technology/twitter-combat-harassment-features/.Google Scholar
- Stephanie Pieschl, Torsten Porsch, Tobias Kahl, and Rahel Klockenbusch. 2013. Relevant dimensions of cyberbullying—Results from two experimental studies. J. Appl. Dev. Psychol. 34, 5 (2013).Google Scholar
Cross Ref
- Robert Plutchik. 1980. A general psychoevolutionary theory of emotion. Theor. Emot. 1 (1980), 3--31.Google Scholar
Cross Ref
- Associated Press. 2016. Retrieved from https://www.dailymail.co.uk/wires/ap/article-3419263/Venezuela-doctors-worried-official-silence-Zika.htm.Google Scholar
- J. R. Quinlan. 1986. Induction of decision trees. Mach. Learn. 1, 1 (1986).Google Scholar
- Edward Raff. 2017. JSAT: Java statistical analysis tool, a library for machine learning. J. Mach. Learn. Res. 18, 23 (2017), 1--5. Retrieved from http://jmlr.org/papers/v18/16-131.html.Google Scholar
- Twitter Safety. 2017. Enforcing new rules to reduce hateful conduct and abusive behavior. Retrieved from https://blog.twitter.com/official/en_us/topics/company/2017/safetypoliciesdec2017.html.Google Scholar
- Huascar Sanchez and Shreyas Kumar. 2011. Twitter bullying detection. In Proceedings of the USENIX Symposium on Networked Systems Design and Implementation (NSDI’11).Google Scholar
- A. Saravanaraj, J. I. Sheeba, and S. Pradeep Devaneyan. 2016. Automatic detection of cyberbullying from twitter. Int. J. Comput. Sci. Info. Technol. Secur. 6 (2016).Google Scholar
- SentiStrength. 2017. Retrieved from http://sentistrength.wlv.ac.uk/.Google Scholar
- P. K. Smith, J. Mahdavi, M. Carvalho, S. Fisher, S. Russell, and N. Tippett. 2008. Cyberbullying: Its nature and impact in secondary school pupils. In Child Psychology and Psychiatry. Cambridge University Press.Google Scholar
- Richard Socher, Alex Perelygin, Jean Y. Wu, Jason Chuang, Christopher D. Manning, Andrew Y. Ng, and Christopher Potts. 2013. Recursive deep models for semantic compositionality over a sentiment treebank. In Proceedings of the Conference on Empirical Methods in Natural Language Processing (EMNLP’13), Vol. 1631, page 1642.Google Scholar
- Nitish Srivastava, Geoffrey Hinton, Alex Krizhevsky, Ilya Sutskever, and Ruslan Salakhutdinov. 2014. Dropout: A simple way to prevent neural networks from overfitting. J. Mach. Learn. Res. 15, 1 (2014), 1929--1958.Google Scholar
Digital Library
- Statista. 2019. Number of monthly active Twitter users worldwide from 1st quarter 2010 to 1st quarter 2019. Retrieved from https://www.statista.com/statistics/282087/number-of-monthly-active-twitter-users/.Google Scholar
- stopbullying.gov. 2014. Facts About Bullying. Retrieved from https://www.stopbullying.gov/media/facts/index.html.Google Scholar
- Gianluca Stringhini, Christopher Kruegel, and Giovanni Vigna. 2010. Detecting spammers on social networks. In Proceedings of the Annual Computer Security Applications Conference (ACSAC’10).Google Scholar
Digital Library
- Aatif Sulleyman. 2017. Twitter temporarily limiting users for abusive behaviour. Independent. Retrieved from goo.gl/yfJrZn.Google Scholar
- Theano. 2017. Retrieved from http://deeplearning.net/software/theano/.Google Scholar
- Aditya Timmaraju and Vikesh Khanna. 2015. Sentiment analysis on movie reviews using recursive and recurrent neural network architectures. Semantic Scholar (2015).Google Scholar
- Naftali Tishby and Noga Zaslavsky. 2015. Deep learning and the information bottleneck principle. In Proceedings of the Information Theory Workshop (ITW’15). IEEE, 1--5.Google Scholar
Cross Ref
- Robert S. Tokunaga. 2010. Review: Following you home from school: A critical review and synthesis of research on cyberbullying victimization. Comput. Hum. Behav. 26, 3 (2010).Google Scholar
- Tweet NLP. 2018. Retrieved from http://www.cs.cmu.edu/∼ark/TweetNLP/.Google Scholar
- Twitter. 2017. About suspended accounts in Twitter. Retrieved from https://help.twitter.com/en/managing-your-account/suspended-twitter-accounts?lang=browser.Google Scholar
- Twitter. 2018. Twitter API. Retrieved from https://developer.twitter.com/en/docs/accounts-and-users/follow-search-get-users/api-reference/get-followers-ids.Google Scholar
- UMICH SI650—Sentiment Classification. 2011. Retrieved from https://inclass.kaggle.com/c/si650winter11.Google Scholar
- Cynthia Van Hee, Els Lefever, Ben Verhoeven, Julie Mennes, Bart Desmet, Guy De Pauw, Walter Daelemans, and Véronique Hoste. 2015. Automatic detection and prevention of cyberbullying. In Proceedings of the Conference on Human and Social Analytics.Google Scholar
- A. H. Wang. 2010. Don’t follow me: Spam detection in twitter. In Proceedings of the International Conference on Security and Cryptography (SECRYPT’10).Google Scholar
- J. B. Watson. 1930. Behaviorism. University of Chicago Press, Chicago.Google Scholar
- Matthew Weaver. 2017. BBC accused of discrimination as salaries reveal gender pay gap—as it happened. Retrieved from https://www.theguardian.com/media/live/2017/jul/19/bbc-publishes-salaries-of-highest-earning-stars-live-updates.Google Scholar
- Nick Wingfield. 2014. Feminist Critics of Video Games Facing Threats in “GamerGate” Campaign. New York Times. Retrieved from https://www.nytimes.com/2014/10/16/technology/gamergate-women-video-game-threats-anita-sarkeesian.html.Google Scholar
- Qing Zou and Eun G. Park. 2015. Trust and trust building of virtual communities in the networked age. In Handbook of Research on Emerging Developments in Data Privacy. IGI Global, 300--328.Google Scholar
Index Terms
Detecting Cyberbullying and Cyberaggression in Social Media
Recommendations
Cyberbullying and traditional bullying involvement among heterosexual and non-heterosexual adolescents, and their associations with age and gender
IntroductionTraditional (offline) bullying and cyberbullying involvement are associated with severe psychosocial problems. Non-heterosexual (LGBQ) youth are more often victimized by traditional bullying than heterosexual (non-LGBQ) youth, but little ...
Why did you do that? Differential types of aggression in offline and in cyberbullying
AbstractTraditional conceptualizations of aggression distinguish between reactive (e.g., rage) and proactive (e.g., reward) functions of aggression. However, critiques of this dichotomy have pointed out these models conflate motivational ...
Highlights- A fine-grained typology of aggression was used to compare the underlying motives in offline and cyberbullying.
Cyberbullying and Traditional Bullying: The Experiences of Poly-Victimization Among Diverse Youth
This study sought to advance the authors' understanding of the relationship between traditional bullying i.e., verbal and physical and cyberbullying. Data were collected from 1,182 participants, ages 13 to 25 M = 19.66; SD = 3.03 from 75 different ...






Comments