Sensory realism and aggression in videogames (Jeong et al., 2012)

This article caught my attention as it examined some of the common concerns about the adverse effects of violent videogames like whether first-person shooters makes people feel immersed because they are looking through the eyes of a soldier or the presence of blood makes people see a hostile world. Eui Jun Jeong (Konkuk University), Frank Biocca (Syracuse University) and Corey Bohil (University of Central Florida) have published their findings in Computers in Human Behavior.

Abstract

This study investigated whether sensory realism cues in violent games – blood color (red vs. blue), screams of pain (on vs. off), and player perspective (first-person vs. third-person) – affect players’ physiological arousal (i.e., skin conductance levels), spatial presence (i.e., sense of being physically ‘‘there’’), and state aggression in a popular violent game (Half-Life 2), controlling for users’ prior game experiences. A path model (N = 160) was examined to see the mediation effects of arousal and presence between realism cues and state aggression. In line with the general aggression model, results showed that realistic blood color and screams increased arousal, but no effect was found for first-person perspective. Presence significantly affected users’ state aggression. However, contrary to our expectation based on the excitation transfer theory, arousal did not show any significant effect on aggression. In addition, presence mediated the influence of realistic blood color on state aggression. In the effects of graphic realism of violence on user aggression, presence did a crucial role. Implications and future studies were discussed.

When Jamie Madigan tweeted my previous post, I realize it was going to be a high impact study. So, I spread a bit around starting with gamasutra.

The authors sought to test two theories of violent video game effects, Bushman & Anderson’s General Aggression Model and Zillmann’s Excitation Transfer Theory. The General Aggression Model posits that aggression are influenced by a person’s internal states: arousal, cognitive and affective state. These internal states are influenced by situational inputs, such as exposure to violent media. The Excitation Transfer Theory posits that aggression is due to being exposed to media that caused excitement and arousal, it would enhance aggressive feelings if evoked. This arousal would take time to dissipate which would explain how people are aggressive after exposure of exciting media of any types.

The authors examined how sensory realism would affect aggression under those two theories and which one or both would explain the data. They defined sensory realism to “formal features of a representation that progressively simulates the same experience in the natural environment.” I don’t recall a paper that comprehensively listed formal features, but the authors examined blood as a visual cue and screams of pains as an auditory cue. The authors also examined point of view, that is whether the play is from a first-person perspective or third-person perspective where the player see the main character’s body, usually from the top-back side.  The authors argued that greater realism would in turn increase presence which is defined as “being there” or being in the video game world.

The authors noted that few studies found the connection between arousal and aggressive feelings. So, they tested this connection in addition to whether sensory realism would increase arousal which in turn leads to increased aggression.

To summarize their theoretical modeling is that videogames cues (i.e., blood, screams of pain and perspective) would have affect aggression (i.e., hostility, anger, physical aggression and verbal aggression) that is mediated by arousal and/or presence.

Method

Participants: 160 undergraduate students somewhere in the Mid-West. Average age is 20.6 (SD = 1.47). There was an equal gender ratio where the sample consisted of 128 men and 32 women. Given this ratio, the authors stratified by making sure that the experimental conditions has an equal ratio of men and women (4 to 1).

Measures

Presence: 20-items from the Independent Television Commission – Sense of Presence Inventory’s spatial subscale.

Physiological arousal: Arousal is measured through participants’ galvanic skin response. So participants wore electrodes on their bodies for about 30 minutes.

Aggression: A modified version of the Buss-Perry Aggression Questionnaire. A 27-item questionnaire that assessed for state aggression, that means how they are feeling right now. The questionnaire has four subscales: anger, hostility, physical aggression and verbal aggression.

Videogame experience: participants were asked number of hours of videogame play.

Videogame used: Half-Life 2 using Garry’s mod. They modded the game so they could create eight experimental conditions: blue vs. red blood, screams of pains vs. no audio, first-person perspective vs. third-person perspective. Each experimental condition has 4 female and 16 male participants.

The authors created a maze of rooms and corridors which made all participants go through the same path set by the researchers. They also included 20 enemies in that maze at pre-determined location. This is high experimental control and it took much effort.

Participants were given 10 minutes practice with no enemies, afterwards they started playing through the maze until they reached the end of the maze. The average playtime is 12 minutes (SD = .25). I am a bit concerned of the relatively short playtime as other studies’ playtime were around 15-20 minutes.

Procedure

Participants completed some questionnaires. They practiced the videogame. Then, they put on the electrodes to measure physiological arousal. They waited 5 minutes to get some baseline physiological arousal by having them sit quietly and relax. After, they play and complete the maze. Again, average completion time is 12 minutes. When they completed the maze, they completed questionnaires assessing for presence and aggression.

Results

The authors conducted a series of ANCOVAs to test sensory realism cues on physiological arousal, spatial presence and aggression with videogame experience as the covariate.

The ANCOVA on physiological arousal revealed that, separately, red blood, and screams of pain led to significantly higher skin conductance level (i.e. higher arousal). An interaction effect has been found between first-person perspective and red blood on arousal. Red blood was more arousing (M = .45, SD = .15) than blue blood (M = .28, SD = .11).

The ANCOVA on spatial presence revealed one main effect where exposure to red blood (M = 3.13, SD = .48) lead to greater scores on spatial presence than blue blood (M = 2.93, SD = .44). Videogame experience had a significant effect.

The ANCOVAs on aggression revealed main effects for blood conditions. Red blood lead to greater hostility score (M = 2.31, SD = .53) and anger score (M = 2.51, SD = .65) than blue blood’s hostility score (M = 2.09, SD = .56) and anger (M = 2.20, SD = .59). No other significant effects were found.

These analyses showed what prior studies have found, the authors analysed the data through structural equation modeling to examine whether physiological and/or spatial presence mediates the relationship between sensory realism and aggression. Their results are shown in the graph below.

jeong

In general, arousal did not affect aggression even if it’s affected by colour of blood and screams of pain. What they found is that the colour of blood and screams of pain had an positive effect on spatial presence which in turn lead to greater  hostility, anger and physical aggression. Given their ANCOVA results, the authors argued that colour of blood and screams of pain has an effect on hostility and anger as mediated by spatial presence.

Discussion

The take home message is that sensory realism cues, such as colour of blood and screams of pain, can indirectly affect aggression. This effect is mediated by our feelings of spatial presence. On the other hand, physiological arousal is affected by sensory realism cues, but it has inconclusive effects on aggression. Finally, the perspective of playing a videogame has no discernible effect on spatial presence, arousal or aggression. The authors argued the effect of blood realism having a role in public policy on game rating systems. Although, I must ask where are the limits of violence between enjoyment and disgust, action versus horror… I should ask around.

The authors earlier referred to sensory realism to “formal features of a representation that progressively simulates the same experience in the natural environment”. There are some issues I’d like to point out.

My eyes! They hurt so much from this resolution! (Half Life 2 mod)

My eyes! They hurt so much from this resolution! (Half Life 2 mod)

The first issue is the technological advancements of three decades since Wolfenstein 3D was released, few studies I identified examined this issue [see link] using multiple iterations of the same game. This study took a dichotomous point, the presence or absence of a realistic feature which seem oddly related to sensory realism. Is it the presence, quantity or quality of a realism cue that affect our feelings of presence? Does pixelated splatters of blood, pre-rendered animation of an expanding pool of blood or the smooth motion of blood streaming down a dead body makes us feel inside the world? Does hearing to a MIDI versus MP3 sound of a dying soldier have an effect?  For future studies, we should tweak the graphics and audio settings of a game, for example playing Half-Life 2 to its lowest settings, medium settings and ultra-high settings. On a tangent, maybe we should do the same with movies, starting with the Rambo series.

The second issue will be the messy job of examining other aspects of realism, such as narrative realism. The authors used Half-Life 2 which is pretty close in depicting a world close to ours, other video games depict fantastical worlds, such as Fallout 3 series, Skyrim, Halo, Gears of War, etc. How does the role of sensory realism in fantasy settings affect presence, arousal and in turn aggression? This becomes more complex when we integrate transportation. The concept of transportation is similar to presence in that we are transported into the narrative world that we are immersed in. For example, how we are so engaged in Game of Thrones and started caring for our favourite characters. As far as I know, there are several studies on the role of narrative in  videogames, although not sure they examined transportation as well.

The rules of the game world is another consideration. I have seen some online matches from various videogames where the physical rules were changed, such as low gravity, party hats, melee-only matches, etc. People would find it jarring when their characters can perform feats beyond a normal human could do, excluding technological or narrative devices. Such violations of physical realism might affect players’ spatial presence. Although, this is contingent upon the narrative realism, so this is quite a messy affair. Nevertheless, this ties into the mental models research where people’s mental models of a weapon, battle, or anything is dependent on their knowledge (if little). I doubt many have fired a shotgun or an uzi and realize that the videogame version is far different from the real life version.

The third  issue is related the perspectives of seeing and experiencing the game world. The argument is that adopting the first-person perspective would lead to greater identification with the virtual characters and it is more realistic which in turn to greater presence and aggression. In relation to other studies, the results are quite mixed. It seems an intuitive argument was met by a mixture of supporting and counter-intuitive evidence. The authors argued that identification may play a role, having an avatar on the screen can increase the player’s sense of presence. This is echoed from a kotaku piece from Stephen Totilo.

First-person gaming, however, often fails to be what first-person is in prose. It is sometimes a vehicle for an individual’s self-reflexive expression, but that is not guaranteed. […]Rarely have I felt, while playing a first-person game, that The Guy In The TV Whose Hand I See Holding The Gun is me.

This suggest that there are other formal features we have yet to identify. Totilo, however, indicated that game developers are improving towards players’ spatial presence.

Today, the first-person advantage is most prominently in the visuals, in the ability to make a video game player feel like I Am In There.

Although, many of the videogames he cited allowed players to self-reflexively express in an open-world. Perhaps, spatial presence is an interaction between narrative and sensory realism.

The authors discussed about their results with arousal and they argued that the non-significant findings is probably attributed to videogames’ interactivity. They critiqued the earlier paradigm and found two important points. The first point is that arousal effects occurs if there was a provocation at an earlier event. The second point is that participants could not retaliate at that earlier event. In a videogame environment, the provocation is usually an enemy character and players can retaliate by whatever  means provided in the videogame resulting in different responses. The authors suggested examining arousal effects at different levels of interactivity. I would go by whether participants could not get retaliate against a confederate in a videogame, or participants missed a chance to retaliate.

Related to arousal, I have some concerns with their aggression measure as it is a self-report measure which is different from a behavioural measure as the former is about behavioral intent. Furthermore, I am concerned about social desirability effects. I would like to see this replicated with a behavioural aggression measure.

Jeong, E. J., Biocca, F. A., & Bohil, C. J. (2012). Sensory realism and mediated aggression in video games. Computers in Human Behavior, 28 (5), 1840-1848. DOI: 10.1016/j.chb.2012.05.002

One thought on “Sensory realism and aggression in videogames (Jeong et al., 2012)

  1. Pingback: Sensory realism and aggression in videogames (Jeong et al., 2012) | Society for Media Psychology and Technology (APA Division 46)

Leave a Reply

Fill in your details below or click an icon to log in:

WordPress.com Logo

You are commenting using your WordPress.com account. Log Out / Change )

Twitter picture

You are commenting using your Twitter account. Log Out / Change )

Facebook photo

You are commenting using your Facebook account. Log Out / Change )

Google+ photo

You are commenting using your Google+ account. Log Out / Change )

Connecting to %s