穿越虛擬與現實之間：大學院校學生現實生活及網路世界的攻擊行為與相關因素之研究

Abstract

本研究旨在探討大學院校學生在虛擬網路世界的論戰經驗和現實生活的攻擊行為，並分析引發網路攻擊的各種可能成因。將引發網路攻擊行為的因素區分為遠端－必要因素、及近端－充分因素，其中上網時數、上網年資、電腦中介溝通環境特性、性別、自尊、內外控人格、依附、社經地位等八個變項，是遠端必要因素，如果個體四周出現遠端變項，則其出現網路攻擊行為的機會大增，但若僅僅出現此八個變項，並不一定導致網路攻擊行為。其次以網路論戰經驗頻率、現實生活攻擊行為和網路認知扭曲作為近端充分因素，若個人具備此三個近端充分因素，則直接導致網路攻擊行為的出現。 研究者自編「電腦中介溝通環境特性」量表，也參酌其他相關研究編製的工具，修改成適合本研究目的的量表，以網路施測的方式，收集了三所國立、五所私立大學的學生樣本共1124份。 研究結果發現，第一，男生在整體網路論戰經驗頻率、目擊網路論戰和親身經歷網路論戰方面的頻率都比女生來得高；同時男生的整體現實生活攻擊行為以及身體攻擊、言語攻擊和減敏感性方面比女生高；網路認知扭曲以及負向思考、自我中心、責怪他人方面，也都是男生高於女生。第二，個人的上網時數多寡會影響其論戰經驗的高低，掛在網路上的時間愈多，自然經歷到的論戰頻率愈多；愈認同於網路環境特性，以及在現實生活中常有較多攻擊行為者，有較多的網路論戰經歷。第三，分析論戰的可能成因，發現現實生活攻擊行為及減敏感性分量表最能預測經驗網路論戰。第四，分析攻擊行為的背後可能成因，發現人格特質中愈傾向於外控人格者、自尊愈低者、在人際相處中愈焦慮不安者，愈容易在現實生活中出現攻擊行為；所有預測現實生活攻擊行為的變項中，以性別最有預測力。第五、網路論戰頻率與現實生活攻擊行為都和網路認知扭曲有顯著相關，並且以現實生活攻擊和分量表「減敏感性」最能預測網路認知。第六、就網路攻擊的遠端因素方面，大學院校男生的網路攻擊傾向遠高於女生，並且，對網路環境的愈認同、對人際相處關係愈是緊張焦慮、上網資歷愈少的新手、自尊愈低、愈傾向外控的人，就愈易在網路上出現攻擊行為；而對網路攻擊預測力最強者為認同網路環境是無規範的人。第七、就網路攻擊的近端因素來看，整體網路論戰經驗頻率、現實生活攻擊行為以及網路認知扭曲和目擊、經驗網路論戰，身體、言語攻擊、減敏感性，負向思考、錯誤標籤、自我中心、責怪他人所有分量表都是顯著的正相關，其中預測力最強的是網路認知扭曲的分量表「自我中心」。以上的研究結果經總結，以路徑分析建立出社會認知論的網路攻擊模式。 總之，人類的攻擊行為似乎從現實穿越到虛擬世界，採取網路攻擊手段的人是個人的人格特質中具有攻擊傾向，同時也受到外在網路環境的影響，個人對其他網路使用者採取負面或不理性的認知評價，更是促成網路攻擊的最重大因素。網際網路已盛行多年，各項不適當或充滿敵意的網路攻擊行為盛行於網路虛擬世界中，加上科技的日新月異，攻擊的手段愈來愈五花八門，本研究以探究網路攻擊者背後的「心理特質」，期能開啟國人對於「攻擊」變項從「現實生活世界」延伸至「虛擬網路世界」的初步認識，更希望研究的發現能對網管人員提供更深入的心理層因了解，以尋求適當的網路管理規約或教育上的輔導策略。

Abstract

This study explores relations of aggressive behavior in real-life situation, sending flaming message in virtual world, cognitive distortion, and Internet aggressive behavior for college students. A social cognitive model is offered to explain why college students engage in Internet aggressive behaviors. In this model, the explanatory factors are further categorized as distal factors – necessary causes and proximal ones – sufficient causes. Individuals having or exposed to distal factors are more likely, but not definitely, to show Internet aggressive behavior; while those who having proximal factors will definitely show Internet behavior. The 8 distal factors are accessibility to the Internet (usage and experience), features of Computer-Mediated Communication environment, gender, self-esteem, locus of control, adult attachment, and social economic status. The 3 proximal factors are flaming experience, real-life aggression, and cognitive distortion about the Internet. Most scales were adopted from previous studies and the reliability and validity were reexamined in a pilot study. Besides, the author designed a new scale “Features of Computer-Mediated Communication environment” (FCMC). These scales were distributed through a website to 1124 Internet users who responded to a recruitment advertisement. Seven results were summarized in below. First of all, Males were more aggressive no matter in real-life or on the Internet. Male subjects had more Internet flaming experiences (observing and participating), higher real-life aggression (including physical aggressive behaviors, verbal aggressive behaviors, and desensitization) and higher cognitive distortion for the Internet (including negative thinking, self-centered, and blaming others). Second, personal Internet usage affects Internet flaming because spending more time in the Net increases the opportunity to see and receive flaming messages. Besides, higher FCMC (perceiving the Internet as an environment of social disinhibitory, deinvidualizied, and with reducing social cues) as well as higher real-life aggression are positively correlated with Internet flaming. Third, according to the results of path analysis, real-life aggression, and its disensitization subscale were the most capable predictors of Internet flaming. Fourth, those who are more external control, lower self esteem, more anxious displayed greater aggressive in real-life but the strongest predictor of real-life aggression was gender. Fifth, real-life aggression, Internet flaming, and cognitive distortion were correlated. The strongest predictor of cognitive distortion of Internet was real-life aggression and its subscale, desensitization. Sixth, the distal factors of gender, FCMC, anxiety, Internet usage, self esteem, and locus of control could predict Internet aggression. Among them, FCMC was the strongest predictor. Finally, among the proximal factors, cognitive distortion of Internet was the strongest predictor of Internet aggression. In general, human aggression displayed in real-life seems penetrates into virtual world. Facing people in the virtual world, individual can react with different strategies. Those who attack other Internet users are of higher aggressive personality during their history of development while Internet environment does play an affecting role also. However, cognitive distortion is a critical mediator in the social cognitive model. In the final section, implications of the results, suggestions for future study, and practical suggestions for Internet management were offered.