Date of Graduation


Document Type


Degree Name

Master of Arts in Psychology (MA)

Degree Level



Psychological Science


Denise Beike

Committee Member

Scott Eidelman

Second Committee Member

William Levine


Computer-mediated communication, Counter-attitudinal exposure, Online incivility, Political communication, Social media, Trolling


Although the term “troll” has existed since the 1980s, its meaning has shifted in recent years as social media use has increased. People provide contrasting and imprecise definitions for what constitutes “trolling,” and often apply the term subjectively to describe online discussants who are uncivil, who are deviant, and who and present counter-attitudinal opinions. Exposure to deviance, counter-attitudinal information, and incivility often leads to unwanted psychological effects. In theory, labeling an uncivil, counter-attitudinal deviant as a “troll” proposes that their intention is to disrupt the conversation and upset other discussants, which provides a reason for why incivility is used, and diminishes the threat of counter-attitudinal exposure. Participants were placed into an online discussion about transgender identities; while participants believed they were discussing with real people, they were actually interacting with a scripted computer program. All discussion comments were pro-attitudinal, save one. A 2 X 3 between-subjects design was employed to examine the effects of the civility of the counter-attitudinal comment (civil vs. uncivil) and the label used against this counter-attitudinal discussant (no label vs. rude label vs. troll label). Incivility exposure overall produced higher ratings of anger, attitude certainty, intentions to participate, and identification with the discussion group. The rude label overall decreased attitude certainty, while the troll label overall increased identification with being a person with their pre-existing attitude. In the uncivil condition, participants were marginally more willing to participate again when the troll label was applied, when compared to the two other label conditions. The intersecting influences of gender, pre-existing attitudes, and suspicions about the deception used are discussed.