The study uses a Between-Subject design. The independent variable is comic book type, either Extremely-Violent Comic Books (EVCB) or Non-Violent Comic Books (NVCB). The Dependant Variable is the participant responses to the ambiguous provocation stories. Subjects are coded on the amount of negative emotion, the amount of retaliation and on the subjects’ appraisal of the intent of the provocateur in the stories. High scores on these variables is taken as evidence of a Hostile Attribution Bias or faulty Social Information Processing. There are a number of within group differences such as gender and trait hostility. The NEO PI questionnaire also provides details of the subjects’ personality variables. These five personality variables are Neuroticism, Extraversion, Openness, Agreeableness and Conscientiousness. Subjects’ experience of comic books is reflected in the variable of years of comic reading and the subjects perceptions of comic books are measured by the variables of like, humour interest and perceived aggression.

There were 127 subjects in total. 75 of these were male (mean age 29.7, age range from 18-42). 52 of the subjects were female (mean age 23.8, age range from 18-48). The subjects were recruited in one of 2 ways. One sample was a convenience sample recruited at the University of Glasgow. These subjects completed a self administered paper copy of the test. 35 subjects completed this version. The remaining subjects were recruited by adverts placed on the Glasgow University HATII site, the 2000AD newsgroup and the Marvel and DC comics’ message boards. 92 of the subjects recruited completed this web site based e-version of the test. Because of the violent content of the EVCB only subjects over 18 participated in the experiment.

The recruited subjects are presented with either the paper or the e-version of the test. Subjects who take the paper version of the test are randomly assigned to either EVCB or NVCB by the experimenter who was blind to the stimulus condition to which subjects were assigned. Subjects who completed the e-version were randomly assigned EVCB and NVCB on the basis of their subject number. Odd numbered subjects were given an EVCB to read and even numbered subjects were given a NVCB to read.
The layout for both tests is identical. Subjects complete a combined NEO-PI and Aggression Questionnaire. This gives a 5-factor personality profile and a measure of Trait Hostility. Subjects then read either an EVCB or an NVCB. This primes them to one of the studies two experimental conditions. After reading the stimulus subjects respond to the ambiguous provocation scenario questions. These responses give a measure of aggression post-stimulus. Finally subjects answer some questions about comic books using the Comic Book Rating and History forms. The data from these gives a measure of the subjects’ years of involvement of comic books and their perceptions of the material they have just read. The complete test takes between 10-15 minutes.

To ensure that subjects do not link their responses on the ambiguous provocation scenarios to the comic books they have just read deception is necessary. Subjects are told that they’re participating in a study that concerns personality development over time. The method they’re told is to complete a personality index, read some comics and then complete the projective test. Subjects are told that the projective test gives a measure of personality at age 10. By correlating the two variables pre- and post-test a measure of personality change can be determined. Subjects are told that the comic books in this study act as distractor task to prevent reflection on the results of the personality questionnaire. To ensure subjects do not ignore the comics they’re told that they will be quizzed on what they have read at the end of the test.

            The Extremely-Violent Comic Book used for this test was single issue of Rose Black.  This comic tells the tale of a vampire secret agent who has become a nun to escape a lifetime of horror. Approximately 90% of the panels in this comic contain aggression or violent acts. The Non-Violent Comic Book used is Archie Comics. Archie describes the lives of teenagers living in the fictional American suburban every-land of Riverdale. Traditional storylines address themes such as school, friendship, dating and family life (Norton, 2003). Archie is never violent and the stories are gently humorous. Subjects read a complete issue of Rose Black or ¾ of an issue of Archie. A good rule of thumb when assessing comic books is the more action there is the fewer the number of frames (Lee & Buscema, 1986). This means that ¾ of an issue of Archie is comparable to a single issue of Rose Black.

The following measures were used to gather the data that will be used for the statistical analysis. Subjects completed the NEO-PI and combined Aggression Questionnaire before reading the stimulus. Post-stimulus, subjects answered questions on the Ambiguous Provocation Scenarios and the Comic Book History and Comic Book Ratings questionnaires.

In previous studies the Buss Durkee Hostility Index (BDHI) was used to measure a participants’ propensity to respond aggressively (Kirsch & Olczak, 2002). This test has been replaced by the Aggression Questionnaire (Buss & Arnold, 1996). This test uses 29 items which are marked on a 7 point Likert scale. Subjects’ answers to the AQ give a 4 factor measure of an individuals’ aggression. These mean of these 4 factors; physical aggression, verbal aggression, anger and hostility is taken as a measure of the respondents aggression pre-stimulus. A median split is used to divides subjects into high- and low-aggression groups. This variable, called Trait Hostility is a two level factor that acts as a measure of individual innate aggressive tendencies.

The Big Five organisation of personality traits is common currency in the field of personality research (Funder, 2001). This test uses a NEO Personality Inventory based on research by Costa & McCrae (Costa & McCrae, 1992). The 5 basic dimensions of this test are Neuroticism (sometimes called Emotional Stability), Extraversion, Openness to experience (sometimes known as Intellect), Agreeableness (which has Antagonism at its opposite pole) and Conscientiousness (or Will to Achieve). The full version of this particular version of the NEO-PI uses 140 items. The test for this study was simplified to 25 items. This cut down version of the test uses 5 items for each personality factor. Responses were marked on a 7-point Likert scale. Low scores on particular items are correlated with subjects who are low in a specific personality factors. High scores correlate with subjects high in a particular personality factors.

The ambiguous provocation scenarios detail an incident that has occurred where the motivation of the provocateur is unclear. How subjects react to these kinds of scenarios is an excellent measure of their underlying aggressive tendencies. Subjects who are low in aggression infer benign intent to the provocateur in the ambiguous scenarios. Those with high aggression infer hostile intent (Dodge 1980). The ambiguous provocation task is a type of projective test. Subjects are asked to respond to the scenarios, which involve incidents that occur in a primary school setting, as if they were “10 years old”. It’s thought that although the incidents are out-with the subjects’ day-to-day experiences that the subjects’ project onto these hypothetical scenarios responses that are a reflection of their current cognitive condition. These “childlike” responses are an ideal way to gain true, raw feelings from subjects.

The method is simple. Subjects read 3 hypothetical stories in which an aggressive incident occurs to a child. In the stories a provocateur, whose intentions are ambiguous, acts relationally or overtly aggressive to another child (Dodge, 1980). Participants then answer questions about the scene. Responses are coded for the amount of negative and positive content expressed. Composite variables are then created for intent, emotion and retaliation. The overt scenarios involve an incident where the protagonist is either hit by a ball or has their new shoes dirtied. In the relational aggression scenario the protagonist is not invited to a birthday party. In each scenario the gender of the victim and of the provocateur is not stated. The scenarios used in this study come from the Kirsch & Olczak 2002b study. These scenarios were modified from provocation stories that originate in Dodge (1980) and Crick’s (1995) work on media aggression.

After reading each scenario subjects are asked 6 questions. 2 of these questions involved the intent of the provocateur (e.g. “Why did the kid throw the ball/muddy the shoes/not invite you to the party?” or “Do you think the kid was trying to be mean or not trying to be mean?”).  Two of the questions involved potential retaliation or punishment of the provocateur (e.g. “How would you get back at the provocateur?” or “What would you do after this happened?”). The final two questions concerned the emotion state of both the victim and the provocateur (e.g. “How does the kid feel about you now?” or “How would you feel after the incident?”). These responses were coded in terms of the amount of negative and aggressive content. A response like “I would hit them” or “I’d make them pay!” was coded as a 2. High scores were also given for quantity of the responses. Subjects who detailed a long involved answer about extreme emotional response (e.g. “they hate me” or “I’m not popular because I’m fat!”) were also coded as a 2. A moderate response such as, “I’d start a rumour” or “they didn’t like me” were coded as a 1. Neutral or positive responses such as “I don’t care” or “they didn’t see me” were coded as a 0. These responses were coded by the experimenter. A second coder blind to the studies’ true intent coded a randomly selected 10% of the tests. The interrater reliability was a Kappa of .89.

Subjects detailed their experience of comic books. They provided a list of the titles they read and how long they have been reading them. Subjects were classified as non-readers (0 years experience), Mild readers (1-5 years experience), Moderate readers (5-10 years experience) or long-term readers (10+ years experience).

Subjects were also asked to rate the comic book they had just read. They subjects rated the comics on how much they liked them (like), how funny they found them (humour), how interested they were in them (interest) and how much violence they perceived (aggression). All of these variables were scored on a 7-point Likert scale.