Communication in multiplayer gaming: Examining player responses to gender cues

Abstract

The goal of this study is to determine how gamers’ reactions to male voices differ from reactions to female voices. The authors conducted an observational study with an experimental design to play in and record multiplayer matches (N = 245) of a video game. The researchers played against 1,660 unique gamers and broadcasted pre-recorded audio clips of either a man or a woman speaking. Gamers’ reactions were digitally recorded, capturing what was said and heard during the game. Independent coders were used to conduct a quantitative content analysis of game data. Findings indicate that, on average, the female voice received three times as many negative comments as the male voice or no voice. In addition, the female voice received more queries and more messages from other gamers than the male voice or no voice.

Keywords

Computer-mediated communication gender voice multiplayer gaming real-time voice

Over the past several decades, video games have become an increasingly popular form of entertainment for millions of people around the world and online gaming in particular has become very popular. For example, Microsoft’s Xbox LIVE service, which connects gamers with each other across the Internet, had 40 million users as of January, 2012 (PR Newswire, 2012). In addition, specific online multiplayer games (e.g., World of Warcraft) have millions of active users. An interesting feature of many multiplayer video games (e.g., Halo 3, Gears of War) is the ability to communicate with other people, in real-time using one’s voice, from within these artificial environments.

Past research has established that many video games are often violent in nature (Carnagey et al., 2007; Cicchirillo and Chory-Assad, 2005; Smith et al., 2003), that video game play may be linked with aggressive behavior (Anderson and Bushman, 2001; Ballard and West, 1996; Cooper and Mackie, 1986; Eastin, 2006), and that women are portrayed in stereotypical ways in many games (Beasley and Standley, 2002). This has been important research that has examined salient issues occurring within gaming; however, much of the research focuses on the content of video games and has yet to look at the interactions between gamers within multiplayer video games. In contemporary video games, gamers are playing an increasingly influential role in creating content by communicating with each other in real-time. Through real-time voice chat, players are able to ascertain the gender of other gamers and immediately respond to these identity cues (Williams et al., 2007). The goal of the current study is to determine the effect of gender voice on communication directed toward that voice. Specifically, the study sets out to quantify this difference by determining, first, whether male and female voices are responded to differently, and, second, to offer a description of how male and female voice responses diverge. To accomplish this goal, we begin by reviewing relevant literature that examines the intersection of gender and video games, as well as discussing a theoretical model that can provide a framework for understanding interactions taking place in online multiplayer video games.

Literature review

Computer and video games have been a fruitful area of research for computer-mediated communication (CMC) scholars. Part of this body of research has focused on gender in video games, specifically stereotypical portrayals of women and men in games. Some scholars argue that ‘gender differences in participation and character portrayals potentially impact the lives of youth in a variety of ways’ (Ogletree and Drake, 2007: 537). It is likely that these stereotyped representations of gender contribute to misleading gender identities and expectations with regard to beauty, sexual appeal and violence (Beasley and Standley, 2002; Dietz, 1998). Because players are exposed to these misrepresentations, this may affect how players view and communicate with other players, particularly female players. Furthermore, misrepresentations of gender take on particular importance when we consider the sheer popularity of multiplayer video and computer games, which often have millions of players worldwide.

As previously mentioned, many modern video and computer games enable players to interact with others in a virtual environment. This allows for shared virtual environments (SVEs), which let people experience being together in the same computer-generated space (Schroeder, 2006). SVEs are unique spaces as advances in video game and computer technology have shifted gaming from a solitary activity to an inherently social one. For instance, instead of a single gamer playing a game by him or herself, today’s gamers can actively play games with others online and interact with them in SVEs. This shift to social gaming is particularly interesting when we consider that many games allow players to communicate with one another via real-time voice chat. The integration of verbal communication into multiplayer gaming allows for additional context cues to be shared with others. In the context of the current study, we are interested in how gamers respond to verbal gender cues communicated through the real-time voice channel and whether these reactions differ based on the voice of the speaker. Given the emphasis on gender in this research, the following section highlights relevant research that has examined gender and video games.

Gender and video games

Gender stereotypes have changed little over time (Aries, 1996). Men continue to be seen as more dominant, aggressive, forceful, industrious and outspoken, while women tend to be viewed as more emotional, submissive, warm, sympathetic and sensitive (Werner and LaRussa, 1985). It is not surprising, then, that women and men tend to be regarded differently online, with the former more likely to experience unpleasant interactions (Norris, 2004). Ballard and Lineberger (1999) studied gender and video game violence and found that punishment became more stringent for women than for men as the level of violence in the games escalated. Norris (2004) suggests that an overtly hostile environment toward women could be a reason why more women do not play computer games. Other scholars have addressed this overtly hostile environment toward women (Cassell and Jenkins, 1998) and some have interviewed female gamers to gain insight into their experiences playing games (see Jenkins, 1998). In general, research seems to indicate that gender has been, and currently is, an issue with regard to video games. One way of understanding the issue of gender and video games is by examining stereotypical portrayals of gender within these games.

This stereotypical portrayal of gender in video games warrants attention, considering the influence of gender stereotypes on shaping behavior (Aries, 1996). Concerned with gender stereotyping in video games, Beasley and Standley (2002) noted that female characters are typically dressed in ways that bring attention to their bodies, particularly their breasts. Other scholars have examined the portrayal of women in video games by focusing on content within the game. In a content analysis of introductory videos for 12 video games, Jansz and Martis (2007) found that women did appear as strong and competent, as well as being leading characters, but they were also portrayed in overly sexualized ways. In the eyes of young men, the female video game characters ‘may come to function as “eye candy” for them’ (Jansz and Martis, 2007: 147a), thus reinforcing gender stereotypes.

Echoing conclusions of research described thus far, Dill and Thill (2007) conducted a two-part study that examined images of video game characters from the top-selling gaming magazines published in the United States (Study 1) and also surveyed college students about their understanding of video game characters (Study 2). The general findings of this study indicated that ‘women should be extreme physical specimens, visions of beauty, objects of men’s heterosexual fantasies, and less important than men’ (Dill and Thill, 2007: 861). With regard to how college students characterized female players, participants of the second study typically characterized female characters in video games as being provocatively dressed, having a curvaceous figure, being thin, being sexual, or being aggressive (Dill and Thill, 2007). Coupled with the stereotypical characterization of female video game characters are the potential ramifications these perceptions have outside virtual environments.

In addition to examining the portrayal of female video game characters in media, research has also examined how exposure to this sexualized content can potentially impact perceptions and behavior. Dill et al. (2008) conducted an experimental study that exposed participants in one condition to images of sexualized video game characters, while the other condition (control group) was exposed to photos of respected professionals. Following this exposure, both groups responded to questions regarding sexual harassment judgments. Dill et al. (2008: 1406) report ‘males who saw the sex-typed images were most tolerant of sexual harassment when judging a real-life case of sexual harassment between a female college student and her male professor’.

In addition to the issues of gender noted above, research has also drawn our attention to the relationship between video games and violence. For example, Eastin and Griffiths (2006) found that gamers who played first-person shooting games exhibited evidence of hostile expectations. Similarly, Chory and Cicchirillo (2007) observed that gamers who played video games more frequently had higher levels of verbal aggression (VA). More importantly, this study also found ‘video game play and player sex interacted to predict trait VA such that men who played video games more frequently reported higher trait VA than did men who played video games less frequently and women’ (Chory and Cicchirillo, 2007: 117). Recently, scholars have indicated that in some online multiplayer games male players were found to be more verbally aggressive than female players (Williams et al., 2009).

In short, past research has found that women are typically portrayed in video games in a stereotypical fashion, that gender influences verbal aggression, and that the portrayal of women in video games may lead to prescriptive norms for how women should behave. Such findings predict that women will be treated differently than men in video games, such that women will experience more negative communication or punishment than men (Ballard and Lineberger, 1999), thus creating an overly hostile environment towards women (Norris, 2004). The following section describes one theoretical model that might be beneficial in examining gender issues in online gaming.

Hyperpersonal model

The hyperpersonal CMC model (Walther, 1996) is perhaps one of the more widely drawn-upon theoretical frameworks in CMC research. The hyperpersonal model posits that CMC users can use advantages presented to them through the characteristics of the CMC channel to exert more control over their self-presentation or impression management, and this takes place through message construction. Essentially, people using CMC leverage the unique characteristics that the CMC channel offers to augment or otherwise enhance relational outcomes in communication with others (Walther, 2007).

One particular concept mentioned as part of the hyperpersonal CMC model, is the concept of hypernegative effects (Walther, 1997; Walther and Parks, 2002). This concept posits that ‘when coupled with time restrictions and no expectation of future interaction, the relatively effortful nature of CMC may trigger overly negative interpretations on the part of receivers, ill regard and hostile message construction by sources’ (Walther and Parks, 2002: 541). It is this hypernegative effect that is of main interest to this study, particularly with regard to how gender might influence this phenomenon.

Taken as a whole, past research demonstrates the groundwork for hostile and negative interactions directed toward a specific gender. A variety of factors have come together to produce these interactions that allow for further hostile treatment of women in online interactions. Thus far, we know that video games portray women in a stereotypical manner (Beasley and Standley, 2002; Dietz, 1998) and that women are treated in negative and hostile ways in CMC interactions (Norris, 2004). Furthermore, the hyperpersonal model posits that interactions via CMC can be just as important and interpersonal as face-to-face interactions (Walther, 1996) and that the affordances generated by the CMC channel (i.e., effortful nature, low expectation of future interactions) can lead to hostile encounters (Walther and Parks, 2002). Given the connections between gender, aggressiveness and hostility in CMC contexts, we can offer some predictions for what might occur in these online interactions. Furthermore, by manipulating gender voice using three experimental conditions (female voice, male voice and no voice), we can compare the types of comments directed toward those conditions and quantify the difference between the conditions:

H1: A female voice will receive more negative comments from gamers than a male voice or no voice.

H2: A female voice will receive fewer positive comments from gamers than a male voice or no voice.

RQ1: Is there any relationship between player skill level and negative comments from gamers toward a female voice, a male voice or no voice?

Method

Online context and participants

While past research has examined aspects of video games in a laboratory setting (Ballard and Lineberger, 1999), this study extends methodological options by situating the experiment in the natural online gaming environment, as suggested by Williams et al. (2009) and Blascovich et al. (2002). By conducting this study in the actual online gaming environment, we were able to maintain control, yet present participants with a more realistic manipulation, as opposed to a fairly mundane one (Blascovich et al., 2002).

For the purposes of this study, the first person shooter (FPS) game Halo 3, for the Xbox 360 video game console, was selected to serve as the online gaming environment from which we would gather data. FPS games allow gamers to interact with the virtual environment by taking control of a character from the first-person point of view. From this viewpoint, gamers move around in and interact with the gaming environment, and this includes interacting with other players.

Halo 3 was a very popular FPS video game released in 2007. Within 24 hours of Halo 3’s release the game sold an estimated $170 million (Reuters, 2007) and within a week this topped $300 million (Microsoft, 2007). Within 18 months over one billion matches had been played online (Bungie, 2009). Besides the popularity of the game, Halo 3 allows players to communicate with others using real-time voice, which allows men and women from around the world to talk with each other in real-time.

In order to collect data for this study, we entered into the Halo 3 matchmaking system and joined multiplayer games to collect data. This matchmaking system automatically pairs up the appropriate number of players based on a selected playlist. The playlists function as a set of predetermined game types, which allow gamers to select the type of game they would like to play, and the matchmaking system automatically places them in a game with players of a similar skill level. A player’s skill level is a numerical indicator of how skilled he or she is in a particular playlist and this number can range from 1 to 50, with 50 being the highest skill level. The player’s skill level changes over time as they become better at the game. In Halo 3, a player has a different skill level for each ranked playlist they have played in. Regardless of the playlist selected, a gamer has no control over who will be assigned to play in that game. Given this restriction, we had no control over which gamers were assigned to our games. Hypothetically, the system would randomly assign players with a similar skill level to the game.

The other gamers we played against were unaware that we were engaged in research or recording game play. Despite this, we did receive Institutional Review Board approval for this research and a waiver of informed consent. Certainly informed consent is a critical component of human subjects research; however, several factors came into play and resulted in a waiver of informed consent. In particular, the fact that participants in this study (i.e. other gamers) used gamertags, essentially self-assigned aliases, instead of their legal names helped to protect participant privacy. Had we contacted participants and asked them to complete a standard informed consent form, we actually would have created a greater threat to privacy than if we did not collect this information. In addition, the terms of use of Xbox LIVE state, ‘you should not expect any level of privacy concerning your use of the live communication features … you understand that these communications can be recorded and used by others, and communications in live-hosted game play sessions may be broadcast to others’ (Microsoft, 2008: 10).

For the purposes of this study, one particular playlist (Team Slayer) was chosen in which all game play took place. Team Slayer is the most popular playlist for Halo 3, with nearly 25 percent of all online games occurring in this playlist (Butcher, 2008). In addition, in order to play against other players with a wide range of skill levels, it is necessary to play in a ranked playlist that allows each condition to earn a numerical skill level and progress from a low skill level to a higher one. From initial pilot testing, it appeared that Team Slayer consistently had a population of over 30,000 gamers on most nights.

Experimental conditions

In order to operationalize gender voice, three Xbox LIVE accounts were acquired, which allowed us to enter into the virtual world of Halo 3 and interact with other gamers. The three accounts each represented a different experimental condition: female voice, male voice and no voice (control). The conditions themselves varied in only two ways: (1) the Gamertag (e.g., self-assigned alias) of each account, and (2) the gender voice each represented. Due to the nature of Xbox LIVE, every Gamertag must be unique and thus three different accounts were required for data collection; however, the Gamertags used in this study varied by only one or two characters.

Prior to data collection, a list of phrases was created and used during the study. The list of phrases was read separately by a female and a male, and recorded onto a computer. These recordings were transferred to an iPod Touch that allowed the researcher to quickly play back a specific phrase for either of the gender voices. The male and female recordings were divided into the three different parts of a typical Halo 3 game: pregame lobby, the game itself, and the postgame lobby. The phrases were, by design, rather innocuous and designed not to illicit a negative response as the typical reaction. In addition, the phrases could conceivably be used in the environment of the game without seeming out of context.

Example phrases used during the pregame portion of multiplayer games include ‘hi everybody’ and ‘I like this map’. Some of the phrases used during the actual games included ‘that was a great kill you just had’, ‘nice job so far’, and ‘I think I just saw a couple of them heading this way’. For the postgame portion of the matches, the following phrases were used: ‘that was a good game everyone’, ‘I had fun playing that game’, and ‘thanks for the game, bye’. Again, the phrases were intended to be innocuous and non-controversial, yet represent potential phrases that might be used by gamers in actual games of Halo 3. The phrases were developed by the lead author, who has experience in FPS games and substantial experience playing Halo 3. Phrases were also reviewed by another Halo 3 player. Pilot testing of the phrases, in games not recorded for this study, indicated that other gamers did not have an adverse reaction to the phrases.

Procedures

We partnered with the Game Research and Immersive Design (GRID) Lab at Ohio University to develop a method of recording game play. This method involved connecting an Xbox 360 console to a device that recorded the audio/video output from the console. In addition to the game audio and video, the audio that the avatar transmitted during the matches was simultaneously recorded into the video file. Thus, after each game was finished, a video file was created which featured not only what the avatar saw and heard, but also what was ‘said’ by the avatar during the game. The second author, who prior to this study had no experience playing FPS games, typically took the controls when playing in games in which the accounts had low skills levels. The lead author, who has experience playing FPS games, took the controls when playing in games with a moderate to high skill level. This division of labor allowed us to interact with other gamers of similar skill levels. The risk to validity by having the researcher play the game was deemed to be less than by having an unskilled ‘other’ play the games. Other researchers have also directly played in video games, interacting with others, as part of scholarly research. For example, Williams et al. (2006) played World of Warcraft (WoW) in their study of social dynamics of guilds in WoW. Thus, researchers collecting data by actually interacting with others through game play is not without precedent.

Coding and coders

Independent coders were recruited for this study. The coders were trained in what to look for in the recordings and how to code comments witnessed in the game. The coders were instructed to look for directed comments aimed at the avatar/experimental condition and to mark them on a coding sheet. A directed comment is one in which the coder is reasonably certain the comment is directed at the avatar. For example, if the avatar just finished speaking and another player reacted to the comment, the other player is clearly directing her or his comment at the avatar. Another directed comment is when another player uses the avatar’s unique Gamertag to address his or her comment to the avatar.

Coders were trained on how to determine if the comment was directed and how to code it. Non-directed comments were not coded onto the coding sheet. Comments were coded into three main categories: directed negative (DN), directed positive (DP) and queries. If the coder felt a comment directed toward our avatar was negative in nature, that comment was coded in the negative category while positive comments were coded in the positive category. The queries category existed for when other gamers asked the avatar questions. Ultimately, the coders used the coding sheet to mark down how many occurrences of each type of comment occurred, as well as when it occurred (pregame, in game, postgame). The three independent coders were paid for their services with funding obtained from a university research grant.

Reliability

A subset of 10 percent of the total games recorded (N = 245) were randomly chosen and used to estimate intercoder reliability. All three coders coded these games separately and their coding sheets were used for calculating intercoder reliability. After addressing minor differences between the coders, intercoder reliability statistics were calculated. Cohen’s kappa was calculated for individual pairings of coders in the three possible configurations: coders 1, 2 (.79), coders 1, 3 (.64), and coders 2, 3 (.77). The overall Cohen’s kappa was found to be .74. Past research (Landis and Koch, 1977) indicates that kappa values of .61−.80 can be considered to have substantial agreement. In addition, percentage agreement statistics indicated that coders agreed 88 percent of the time on how a in game should be coded.

Conditions

A total of 245 games were recorded, with 82 games in the control condition, 82 in the female condition, and 81 in the male condition. A total of 1,711 gamers participated in this study, 1,660 of whom were unique gamers. Typically, each game had seven other players in addition to the experimental condition. Across all three conditions, 58 percent of the games resulted in a win for the team the experimental condition was on and a loss for the remaining 42 percent. Broken down by condition, the win percentage for control was 59 percent, for female it was 56 percent and for male it was 61 percent. With regard to verbal communication, nearly half (44 percent) of games in the control condition did not have any speaking taking place; however, 77 percent of the male condition games and 79 percent of the female condition games did have verbal communication taking place. Ultimately, the three conditions appear to be relatively homogeneous with regard to win percentage, while the percentage of games with verbal communication for the female and male condition appears to be nearly equal.

For each game, an average skill level was calculated by adding together the highest skill level achieved for each player in that particular game, including the respective experimental condition, and dividing this number by the total number of players in that game. This calculation created a numerical value that represented the mean skill level of all players who played in that particular game. We used the highest skill level achieved, for each player in each game, in order to more precisely calculate a meaningful average. For example, one gamer may have a skill level of 5 in the Team Slayer playlist used for this research; however, that same player may have a skill level of 50 in another playlist. If the smaller value for skill level was used in the analysis, the results would likely be skewed and the average skill level of the game would artificially be decreased. Thus, the highest skill level for each player was used to calculate the mean. This numerical score offers more precision when measuring the skill and experience of those gamers we played against and is a more precise indicator of the average skill level of all players in a specific game. The average skill level for all of the games was 18 (N = 245, SD = 8.33). For the control condition games the average skill level was 17 (N = 82, SD = 7.23) for the female condition games it was 19 (N = 82, SD = 8.96), and for the male condition games it was 19 (N = 81, SD = 8.63). The differences in average skill levels between the games used in the different conditions were not found to be statistically significant, F(2, 242) = 1.61, p > .05.

Analysis

In order to examine the communication occurring in the games played, the analysis focused on the 163 games in which verbal communication occurred. Coders viewed the games and noted instances of communication occurring in three main categories: directed negative (DN), directed positive (DP) and queries. Hypothesis 1 predicted that the female condition would receive more negative comments than the other two conditions. An ANOVA revealed differences in mean values for DN, DP and queries in each of the three conditions (see Table 1). The ANOVA test indicated that the difference in mean DN between the conditions was found to be statistically significant, F(2,160) = 5.75, p < .05, η² = .067. In addition, mean DP was also statistically significant, F(2,160) = 12.09, p < .05, η² = .131, as well as mean queries, F(2,160) = 6.09, p < .05, η² = .071. Post hoc comparisons using the Tukey HSD were performed for each variable. In the DN variable, the control condition (M = .61, SD = 1.79) was significantly different than the female condition (M = 2.82, SD = 4.79) but not significantly different than the male condition (M = 1.24, SD = 2.29). In addition, the mean score for the female condition was found to be significantly different than the male condition and the control condition. The male condition was only significantly different from the female condition. For directed negative comments overall, this study found that the female condition received roughly three times as many DN comments as the male condition and the control condition. Results from this study lend support to Hypothesis 1.

Table 1.

Means comparison by experimental condition

	Values
Variable	Control	Female	Male
Directed	M = 0.61_a	M = 2.82_ab	M = 1.24_b
Negative	SD = 1.79	SD = 4.79	SD = 2.29
	n = 36	n = 65	n = 62
Directed	M = 0.94_cd	M = 3.03_c	M = 4.23_d
Positive	SD = 1.62	SD = 3.46	SD = 3.5
	n = 36	n = 65	n = 62
Queries	M = 0.08_e	M = 1.92_e	M = 1.02
	SD = 0.37	SD = 3.47	SD = 2.16
	n = 36	n = 65	n = 62

Note. Common subscripts indicate a statistically significant difference between groups.

Hypothesis 2 predicted that the female condition would receive fewer positive comments than the other two conditions. In the DP variable, the control condition (M = 0.94, SD = 1.62) was found to be significantly different than the female condition (M = 3.03, SD = 3.46) and the male condition (M = 4.23, SD = 3.54). No significant difference between the female and male condition was found. In the remaining variable, queries, the only significant difference in mean score was between the control condition (M = 0.08, SD = 0.37) and the female condition (M = 1.92, SD = 3.47). Although the male condition received, on average, more positive comments than the female condition, this difference could likely be due to chance and does not fully support Hypothesis 2.

RQ1 asked if there was any relationship between player skill level and negative comments. No correlation was found between the number of DN comments and the adjusted skill level (r = .02, n = 163, p >.05) among all of the games. No correlation was found in the control condition games (r = .22, n = 36, p >.05), female condition games (r = .02, n = 65, p >.05), or male condition games (r = −.12, n = 62, p >.05) between adjusted skill level and DN comments. However, a correlation was found between the number of DP comments and adjusted skill level for the female condition only (r = .25, n = 65, p <.05). In addition, another correlation was found between the number of DP comments and the number of queries for all of the games (r = .30, n = 163, p <.05); however, when this correlation was run in the individual conditions, only the female condition provided evidence of the correlation existing (r = .37, n = 65, p < .05).

Discussion

This study sought to examine the effect of gender voice on communication occurring within an online multiplayer game. One goal of the study was to quantify the difference in how gamers reacted to different gender voices (i.e. male and female). Past research (Beasley and Standley, 2002; Burgess et al., 2007; Dietz, 1998; Norris, 2004; Ogletree and Drake, 2007) has been unable to do this since the focus of those studies was on the content of games and not what actual gamers bring to the game. While the goal of this study was more quantitative in nature, and therefore a formal analysis of the actual verbal exchanges was not conducted, it is worth noting that we observed a clear pattern of negative comments associated with the female condition. On several occasions the female condition was exposed to derogatory gendered language. For example, in one particular game nearly every utterance made by the female condition was met with a negative response by a particular gamer. When the female condition said ‘hi everybody’, the other gamer responded with ‘shut up you whore’ followed a few seconds later with ‘she is a nigger lover’. When the female condition said, ‘alright team let’s do this’, the other gamer replied, ‘fuck you, you stupid slut’. Later on in the game the female condition was again referred to as a ‘slut’. Similarly, in another game, the female condition was asked questions like ‘are you good?’ and then told, ‘your voice is beautiful’. Further into that game, that gamer asked the female avatar to be his girlfriend.

In another game, after we played back the female voice, another player on our team said ‘hold on, who the fuck are you? Shut the fuck up, o my God I don’t want hear your bullshit’. That same gamer repeatedly told us to ‘shut the fuck up’ and also repeatedly called us a ‘bitch’.

In yet another game, a gamer responded to our opening comment by saying ‘so whatever that voice was, are you a hooker or are you a dude’? At the end of the game another player mocked the female voice by sarcastically repeating the phrase ‘good game everyone’ followed by ‘shut the fuck up’.

Even though the female and male conditions had nearly the same win percentage (56 percent and 61 percent respectively), nearly the same number of games with verbal communication taking place (79 percent and 77 percent respectively), the same average adjusted skill level (19 out of 50), and nearly the same number of total games played (82 for female and 81 for male), gamers reacted differently to the female condition than to the male or control condition. In general, the female voice received three times as many directed negative comments than the male voice or no voice conditions, as well as receiving more messages and friend requests from other players than the other two conditions.

With regard to the connection between skill level and comments, this study did not find an overall correlation between skill level and negative comments in the three voice conditions. That being said, we did find evidence of two statistically significant correlations that occurred in the games with the female voice. Specifically, we found low to moderate correlations in the female condition for skill level and positive comments as well as positive comments and queries. Given that the games with the female voice received three times as many negative comments as the other conditions, these correlational findings seem to indicate that another relationship may be present. Although the female condition received more negative comments, it does appear that when gamers reacted positively to a female voice they also seemed to ask more questions. In addition, this research suggests that as skill level increased so did directed positive comments. However, correlation is not causation and exploring this relationship further is clearly an area for continued research.

The hyperpersonal CMC model (Walther, 1996) posits that in CMC interactions users can use the advantages of the medium to enhance communication with others through message creation. Users can apply more control over their self-presentation and impression management through additional control over message construction than is available in face-to-face interactions. By leveraging the unique characteristics that CMC channels offer, users can augment or enhance relational outcomes when communicating with others (Walther, 2007). This can lead to hyperpersonal communication, or relations that are just as salient or more so than face-to-face relationships; however, hypernegative interactions may occur. Hypernegative effects (Walther, 1997; Walther and Parks, 2002) are those interactions in which users engage in hostile or negative communication due to the characteristics of CMC.

In the case of the current study, we do see evidence of hypernegative effects at play. Findings indicate that the female condition received more negative comments than the male condition, despite controlling for message content. Many of the messages directed to the female condition were indeed hostile. We would argue that the effortless nature of the communication channel (i.e., real-time voice), the low expectation of future interaction with the experimental condition, and past research dealing with attitudes and portrayals of women in video games (Aries, 1996; Eastin, 2006; Norris, 2004) create an environment conducive to generating hypernegative effects, particularly toward women. Despite the male and female conditions performing nearly equally and broadcasting messages with identical content, hypernegative effects were present and targeted disproportionally to the female condition.

While this study provides evidence supporting the claim that the gender of the player does impact communication occurring in a multiplayer game, it is not free of caveats. As with any study, the sample used may have had an impact on results. While we feel the game used in this study is appropriate, had a different video game, or video game genre, been used for this research the results may well have been different. Caution should be shown when generalizing the findings of this study to other games or genres. Beyond the game used, the number of games played may also have been a factor as well as the skill level of other players.

The most significant suggestion for future research would be to change the unit of analysis from the individual games to the individual gamers. Instead of looking at the 245 games played in for this project, future researchers may want to look instead at the over 1,700 gamers who participated in this research. While the goal of the current study is preliminary in nature and meant to quantify the difference in communication directed towards a female voice and a male voice, future studies would do well to expand coding by transcribing what each gamer said. In addition, with the aid of computer programs it should be possible to analyze the text of what gamers said and perhaps even conduct a textual analysis comparing the language used toward or about the female and male condition.

Finally, the way in which we collected data is something other researchers should, at the very least, consider. By playing in and recording actual multiplayer games, researchers are left with an accurate representation of exactly what occurred. Furthermore, this dataset represents real-world data, not data collected from participants given an arbitrary task in a laboratory-like setting. We would argue this method allows researchers to capture more accurate and representative data that has stronger ties to the way people use this technology to communicate. This allows for a very rich data set that can be analyzed from a multitude of perspectives.

Another implication of this study concerns our focus on the gamer’s ability to influence the game content. Future gaming and gender research should continue to examine the significance of the gamer and his or her impact on the game content. Game developers may go to great lengths to avoid stereotypical portrayals of women in their games, yet it may be possible for gamers to continue perpetuating sexualized and stereotypical portrayals of women in their actual gameplay. Coming to a more thorough understanding of this phenomenon and providing answers to various different stakeholders is an important next step for game and gender researchers, especially as our society continues to develop and adopt new communication technologies.

Conclusion

Overall, this study sought to examine how manipulating gender voice impacts how others communicate in a virtual environment. The current study builds upon the work of past research and helps to bring our understanding of gender and gaming a step further by providing three major contributions. The first is the findings from this study. In particular, this study quantified a communicative phenomenon occurring among gamers in an online virtual environment. Past research has demonstrated that, in the content of many video games, women are portrayed in stereotypical and sexualized forms. However, until this point research has typically focused on the content of video games, developed by the game developers, and not the interactions between gamers. Given this limitation, the current study provides a more thorough understanding of how gamers treat women in a multiplayer game by providing evidence to help answer the question of how women are treated differently than men.

The second main contribution focuses on the way in which data was collected. Broadly, collecting data by capturing the audio and video feed from a video game console as a video file for later analysis is a unique way of conducting communication research. However, the more worthwhile contribution this study provides is situating the research in the actual gaming environment as opposed to a more artificial one. While past research has typically examined gamers either in isolation and/or in a laboratory-like setting, disconnected from the inherent social aspect of contemporary gaming, this study avoids those pitfalls by engaging with gamers in the environment of a multiplayer game. Because the participants of this study were actual gamers, unaware of the study, we argue that this study is better able to capture the way in which gamers communicate with others.

The third main contribution of this study is a shift in the focus of what is actually being studied. Past research has often focused on the sexualized and stereotypical portrayal of women in video games, and this is important, worthwhile research. Yet such research ignores the massive amount of communication occurring through multiplayer games. Put another way, past research has not fully examined the content generated by gamers and instead has focused on that content created by game developers. While we can still analyze those aspects of games created by game developers, we propose that a fruitful area of future research, one that game studies, genders studies, and other research areas should explore, is the content generated by the interaction of gamers within the virtual space of multiplayer games. This is the area in which people are spending their time and the area that researchers should focus on.

In summary, this study has established that gamers react differently to a female voice than to a male voice or no voice. Specifically, gamers reacted more negatively to a female voice than a male voice, even when both voices were communicating the same message, and the female voice received three times as many negative comments as the male voice. The next logical step is to determine exactly what causes this reaction and why gamers feel inclined to communicate in this way. It could be, as others have hypothesized, that characteristics of the communication channel may allow this negative behavior to flourish. It may also be that, within video games, people feel more freedom to express themselves without fearing the consequences of their actions. Future studies may also find a combination of factors that interact together to influence communication and behavior in this area. However, it is clear that additional research is necessary in order to more fully understand communicative and behavioral phenomena occurring within online virtual spaces.

Footnotes

Funding

This research was supported by an original research grant from the Ohio University Graduate Student Senate.

Author biographies

Jeffrey H. Kuznekoff (PhD, Ohio University, 2012) is a graduate of the School of Communication Studies at Ohio University. His research examines the influence of new technology on communication.

Lindsey M. Rose (PhD, Ohio University, 2012) is a graduate of the School of Communication Studies at Ohio University. Her research interests focus on health communication and community organizing.

References

Anderson

Bushman

(2001) Effects of violent video games on aggressive behavior, aggressive cognition, aggressive affect, physiological arousal, and prosocial behavior: a meta-analytic review of the scientific literature. Psychological Science 12(5): 353–359.

Aries

(1996) Men and Women in Interaction: Reconsidering the Differences. New York: Oxford University Press.

Ballard

Lineberger

(1999) Video game violence and confederate gender: effects on reward and punishment given by college males. Sex Roles 41(7): 541–558.

Ballard

West

(1996) Mortal Kombat (TM): the effects of violent video game play on males’ hostility and cardiovascular responding. Journal of Applied Social Psychology 26(8): 717–730.

Beasley

Standley

(2002) Shirts vs. skins: clothing as an indicator of gender role stereotyping in video games. Mass Communication and Society 5(3): 279–293.

Blascovich

Loomis

Beall

. (2002) Immersive virtual environment technology as a methodological tool for social psychology. Psychological Inquiry 13(2): 103–124.

Bungie (2009) One billion served and counting. Available at: www.bungie.net/News/content.aspx?type=topnews&;link=CGHV33KGD2R9D3W9V4QY7MF9Q.

Burgess

MCR

Stermer

Burgess

(2007) Sex, lies, and video games: the portrayal of male and female characters on video game covers. Sex Roles 57(5): 419–433.

Butcher

(2008) E pluribus unum matchmaking in Halo 3. Available at: www.bungie.net/images/Inside/publications/presentations/gdc2008_butcher_chris_matchmaking.ppt.

10.

Carnagey

Anderson

Bushman

(2007) The effect of video game violence on physiological desensitization to real-life violence. Journal of Experimental Social Psychology 43(3): 489–496.

11.

Cassell

Jenkins

(eds) (1998) From Barbie to Mortal Kombat. Cambridge, MA: MIT Press.

12.

Chory

Cicchirillo

(2007) The relationship between video game play and trait verbal aggressiveness: an application of the general aggression model. Communication Research Reports 24(2): 113–119.

13.

Cicchirillo

Chory-Assad

(2005) Effects of affective orientation and video game play on aggressive thoughts and behaviors. Journal of Broadcasting & Electronic Media 49(4): 435–449.

14.

Cooper

Mackie

(1986) Video games and aggression in children. Journal of Applied Social Psychology 16(8): 726–744.

15.

Dietz

(1998) An examination of violence and gender role portrayals in video games: implications for gender socialization and aggressive behavior. Sex Roles 38(5): 425–442.

16.

Dill

Thill

(2007) Video game characters and the socialization of gender roles: young people’s perceptions mirror sexist media depictions. Sex Roles 57(12): 851–864.

17.

Dill

Brown

Collins

(2008) Effects of exposure to sex-stereotyped video game characters on tolerance of sexual harassment. Journal of Experimental Social Psychology 44(5): 1402–1408.

18.

Eastin

(2006) Video game violence and the female game player: self and opponent gender effects on presence and aggressive thoughts. Human Communication Research 32(3): 351–372.

19.

Eastin

Griffiths

(2006) Beyond the shooter game: examining presence and hostile outcomes among male game players. Communication Research 33(6): 448–466.

20.

Jansz

Martis

(2007) The Lara phenomenon: powerful female characters in video games. Sex Roles 56(3): 141–148.

21.

Jenkins

(1998) Voices from the combat zone: game Grrlz talk back. In: Cassell

Jenkins

(eds) From Barbie to Mortal Kombat: Gender and Computer Games. Cambridge, MA: MIT Press, pp. 328–341.

22.

Landis

Koch

(1977) The measurement of observer agreement for categorical data. Biometrics 33(1): 159–174.

23.

Microsoft (2007) Halo 3 records more than $300 million in first week sales. Available at: www.microsoft.com/en-us/news/press/2007/oct07/10-04Halo3FirstWeekPR.aspx.

24.

Microsoft (2008) Xbox LIVE games for Windows LIVE terms of use. Available at: www.xbox.com/en-US/live/legal/legal-tou.htm.

25.

Norris

(2004) Gender stereotypes, aggression, and computer games: an online survey of women. Cyber Psychology & Behavior 7(6): 714–727.

26.

Ogletree

Drake

(2007) College students video game participation and perceptions: gender differences and implications. Sex Roles 56(7): 537–542.

27.

Newswire

(2012) Xbox LIVE partners with ESPN, GameSpot, Last.fm, Manga Entertainment and MUZU. TV to expand TV-like advertising on Xbox LIVE. Available at: www.prnewswire.com/news-releases/xbox-live-partners-with-espn-gamespot-lastfm-manga-entertainment-and-muzutv-to-expand-tv-like-advertising-on-xbox-live-148703415.html.

28.

Reuters (2007). Microsoft’s Halo 3 registers biggest day in US entertainment history with $170 mln in sales. Available at www.reuters.com/article/2007/09/27/idUSIN20070927063131MSFT20070927.

29.

Reuters (2007) Microsoft’s Halo 3 registers biggest day in US entertainment history with $170 mln in sales. Available at: www.reuters.com/article/2007/09/27/idUSIN20070927063131MSFT20070927

30.

Schroeder

(2006) Being there together and the future of connected presence. Presence-Teleoperators and Virtual Environments 15(4): 438–454.

31.

Smith

Lachlan

Tamborini

(2003) Popular video games: quantifying the presentation of violence and its context. Journal of Broadcasting & Electronic Media 47(1): 58–76.

32.

Walther

(1996) Computer-mediated communication: impersonal, interpersonal, and hyperpersonal interaction. Communication Research 23(1): 3–43.

33.

Walther

(1997) Group and interpersonal effects in international computer-mediated collaboration. Human Communication Research 23(3): 342–369.

34.

Walther

(2007) Selective self-presentation in computer-mediated communication: hyperpersonal dimensions of technology, language, and cognition. Computers in Human Behavior 23(5): 2538–2557.

35.

Walther

Parks

(2002) Cues filtered out, cues filtered in: computer-mediated communication and relationships. In: Knapp

Daly

(eds) Handbook of Interpersonal Communication, 3rd edn. Thousand Oaks, CA: SAGE, pp. 529–563.

36.

Werner

LaRussa

(1985) Persistence and change in sex role stereotypes. Sex Roles 12(9): 1089–1100.

37.

Williams

Caplan

Xiong

(2007) Can you hear me now? The impact of voice in an online gaming community. Human Communication Research 33(4): 427–449.

38.

Williams

Consalvo

Caplan

. (2009) Looking for gender: gender roles and behaviors among online gamers. Journal of Communication 59(4): 700–725.

39.

Williams

Ducheneaut

Ziong

. (2006) From tree house to barracks: the social life of guilds in World of Warcraft. Games and Culture 1(4): 338–361.