• Nie Znaleziono Wyników

Pilot experiment for the research game and first evidence for the effects of intergroup

The experiment was a pilot study for the research game. Its objective was to assess whether a short session of playing a very simple game would impact players’ attitudes. More specifically, whether a game with a diverse social world, that includes intergroup contact with NPCs, would impact players’

attitudes towards out-groups in a different way than a session with a game that depicts a homogenous social world and has no intergroup contact with NPCs.

Participants

Participants were volunteers, recruited with posts on Polish and European Facebook groups for gamers. Each post informed that the study was conducted for a PhD research project and participation meant playing a game and filling out a short survey, along with information about the experiment’s length (about an hour) and technical requirements (keyboard, mouse and a browser with an internet connection).

Prospective subjects were informed that participation was voluntary and they could opt out of it at any time during the study. They were assured no private data would be collected. No incentive was offered for participation in the study.

The experiment was designed to examine intergroup contact. Therefore, only subjects belonging to one racial group were included in the analysis (white, the majority in Europe) to make certain that in one of the two experimental conditions there was no intergroup contact with NPCs and players saw a homogenous in-game world populated by their own racial group. To ensure each person would play Figure 4. A graphic attached to posts on Facebook groups to recruit participants.

48

only once, a unique ID was randomly assigned to each player by the game, stored in the browser and sent to the server along with information about gameplay time. For those that played twice, only the first playthrough was taken into account in the analysis. The fastest respondent finished the game in 25 minutes, the slowest in 409 minutes (M = 67.5 minutes, median = 57, SD = 47.6). Respondents who took longer than three hours to beat the game (4 people) were excluded from the analysis as the long gameplay time indicated there might have been something unusual in their session, e.g. breaks or that the participant might have found it very difficult to understand the game. The final sample (N

= 152) was mostly male (92.1%). The rest self-identified as female (11 respondents, 7.2%) or “other”

gender (1 person). Respondents’ age ranged from 13 to 66 years with the average of 24 years (SD = 6.9). Most of the respondents were residents of largest cities (43.4% lived in a city with over 500 thousand inhabitants). In the final sample, participants took, on average, 62 minutes to beat the game (SD = 22.5 mins). The Polish version of the game was played by 62.5% (95 people) and the English version by 37% (57 people).

Experimental manipulation

Experimental manipulation consisted of playing the research game. Players were randomly assigned to one of two conditions: 1) the All White condition and 2) the Diversity condition.

1. In the Diversity condition each building was randomly assigned a Black or a White manager so that players saw both Black and White characters (in-groups and out-groups) in their own team as well as in the competitors’ teams. Therefore, the Diversity condition featured both positive and negative intergroup contact (cooperation and common goals as well as conflicting goals and fighting with Black NPCs)

2. In the All White condition all the characters in the game were White (that is, of the same racial group as the player) so that the game featured no intergroup contact. This condition created a control group.

49

Measures

After the game participants filled out a short survey. The survey contained questions about demographical data and a measure of attitudes toward Black people.

Attitudes towards Black people were measured using the short version of the social distance scale (Bogardus, 1933). The social distance scale was often used by intergroup contact hypothesis researchers. In their meta-analysis, Tropp and Pettigrew (2004) recorded its use in close to 19% of the studies and confirmed it has shown strong results for the effects of contact on attitudes as well as for generalisation effects. Participants who played the game in Polish received a translated version of the scale, validated in Poland in three representative, nationwide prejudice surveys (Bilewicz, 2009, Winiewski, 2017). The scale consists of three questions about acceptance of a minority representative as a family member (e.g. “I would be willing to accept a Black person as a close relative by marriage.”), a co-worker (e.g. “I would be willing to accept a Black person as a co-worker”), and a neighbour (e.g.

“I would be willing to accept a Black person as a neighbour on the same street.”). The answers were given on a 5-point Likert scale. The scale was coded so that higher scores represented a more positive attitude, that is a lower score signified a higher social distance (i.e. lower minorities acceptance). Both English and Polish versions of the scale were reliable with Cronbach’s α at .936 and .928 respectively (M = 13.56, SD =2.61 for the English version and M = 15.97, SD = 2.91 for the Polish version).

Results

To assess the effects of intergroup contact with NPCs on players’ attitudes, a factorial analysis of variance was conducted on the social distance toward Black in the two experimental conditions.

Demographic data (size of the city of residence, gender and age) and the language respondents used (English for people recruited in the European gamer groups or Polish for players from Polish gamer groups) were entered as covariates into the model.

The analysis confirmed that a diverse virtual world impacts players’ attitudes differently than a homogenous virtual world F(1,146) = 4.68, p = .032, η2 = .030. Respondents who experienced contact with Black NPCs had significantly more positive attitudes towards Black people (M = 5.18, SD = 0.80)

50

than respondents who played in a homogenous virtual world (M = 4.87, SD = 1.1). Additionally, two of the covariates proved significant. Female respondents had a significantly higher acceptance of Black people (M = 5.12, SD = 0.93) than male respondents (M = 5.04, SD = 0.96, with gender significantly differentiating players’ scores at F(1,146) = 6.54, p = .01). Polish respondents had a higher acceptance of Black people (M = 5.32, SD = 0.97) than English speaking respondents (M = 4.51, SD = 0.87) at F(1,146) = 21.63, p < .00.

Discussion

The results show that even brief, simplified intergroup contact with NPCs in a game can influence attitudes towards real-life out-groups. The crudeness of the contact in the research game provided players with very little knowledge about the out-group and allowed for very little empathy, both of which are known mediators of the relationship between contact and prejudice reduction (Dovidio et al., 2003). The characters in the game were simple, they had no story or characteristics of their own (aside from looks

visible in the graphic representations) and contact was restricted to task management (players ordered NPCs to manage buildings, collected money from them or attacked them) or exchanging task-related information (all statements made by NPCs informed players about the situation on the map).

Moreover, due to the presence of numerous NPCs on the map, contact with each one of them was extremely brief and not conducive to creating emotional connections or to individuating any of them from the observed group in general, which typically enhances the effects of contact on attitude (Brewer

& Miller, 1984), with stronger affective ties (Pettigrew, 1997, Hamberger & Hewstone, 1997) and greater feelings of intimacy (Tropp, 2008) resulting in stronger positive effects. This in turn suggests Figure 5. Acceptance of Black people measured with a social distance scale for participants in the two experimental conditions. The error bars represent the confidence intervals.

51

that intergroup contact with NPCs may constitute a valid path for shaping intergroup attitudes, even in simple mobile games. It stands to reason, that the effects would be stronger for robust video games, which often contain narrative elements, last for hours and include prolonged contact with multiple individuated NPCs, many of which with their own background and story. However, the simplicity of contact could also have had positive aspects for intergroup outcomes. In the absence of any other defining characteristics aside from facial features (unequivocally racially categorising) and team affinity, the salience of racial categories was high. The salience of out-group defining categories, in turn, is an important factor contributing to positive attitude change in intergroup contact (Brown &

Hewstone, 2005). Additionally, the simplicity of contact, that is conducting it in a repeating manner at specific times in a game made the contact predictable. As predictable elements may be less threatening and produce less anxiety (Miller, 1981), this might have contributed to the reduction of out-group threat and anxiety, associated with positive effects (Riek et al., 2006). Therefore, the simplicity of the game might have, in some ways, been inducive for positive intergroup outcomes.

The Diverse experimental condition, which exposed White players to contact with Black NPCs resulted in a more positive attitude towards Black people (measured with a social distance scale).

However, this outcome is not entirely intuitive. Diversity itself is known to cause both positive as well as negative outcomes for intergroup attitudes (Koopmans & Veit, 2014, Laurence et al., 2018). More specifically, diversity establishes opportunities for intergroup contact but that contact might be positive as well as negative. It could be said that in the research game players experienced both. Their intergroup interactions were positive with Black NPCs that belonged to their team and managed their restaurants. Contact with these NPCs included of cooperation and common goals which are important situational elements for attitude improvement and were part of optimal conditions defined by Allport (1954). Additionally, because those NPCs were part of the players’ team, a single superordinate identity was established that encompassed the players and Black NPCs, which could further contribute to positive effects (Gaertner & Dovidio, 2000). However, players also had conflicting goals with Black NPCs on other teams, their interactions were based on rivalry and included fighting, all of which define negative contact (Brewer & Miller, 1984). Although the literature mostly focuses on the positive effects of contact, researchers recognise the possibility of negative outcomes (Barlow et al.,

52

2012). Moreover, negative valence asymmetry research shows negative contact to have more prominent effects (Graf & Paolini, 2017) that is, a quantity of negative contact is a stronger and more consistent predictor of increased prejudice than a similar quantity of positive contact for decreased prejudice. This holds especially for contexts when category salience is high. In the research game, the amount of both types of contact was similar and category salience was considerable because categories of race and team were the only defining things for NPCs. Therefore the improvement of intergroup attitudes in the first experiment requires consideration. One of the reasons for the asymmetric outcomes for positive and negative contact is the fact that negative contact increases group salience more than positive contact and thus produces a bigger change in attitudes (Graf & Paolini, 2017). In the research game, where both types of contact occurred simultaneously, the group salience, regardless of its source, would change for both negative as well as positive interactions simultaneously, rendering this ground for negative asymmetry invalid. Additionally, some researchers found that intergroup contact conducted in more structured situations does not produce negative asymmetry (Graf & Paolini, 2017). The research game was a highly structured situation, in which both sides had clearly defined, narrow options for actions and allotted time for action. This is true for many video games, as by definition, a game restricts players’ options by defining rules and possible moves (Egenfeldt-Nielsen et al., 2008).

Therefore, negative asymmetry might be less pronounced in video games and including positive intergroup contact in them could offset violent interactions easier than in real life. These situational aspects do explain the lack of negative asymmetry. They do not, however, provide a reason for positive contact offsetting the negative, as was the case in the experiment. One reason for this might be that conflict experienced in the research game was relatively mild compared to many games on the market as it did not include depictions of violence (it featured explosions and destruction of buildings only).

However, so was positive contact. Both were seen as part of building management, observed on a simplified map and experienced through simplified interactions.

A different explanation might arise from the findings of Graf and colleagues (Graf et al., 2014). In their research, they established that the prominence of negative contact holds only when the

53

valuations of contact valence were framed around the behaviours of contact partners (e.g. “she was unfriendly”), but not when they were framed around the contact situation (e.g. “I did not enjoy the meeting”). Their studies highlight two aspects of the situation that might mitigate the impact of negative contact on players’ attitudes. Firstly, as previously mentioned framing of the negative contact might have been partially mitigating. Several players sent unprompted feedback to the researcher after the game and whereas some described the competition framing contact around the NPCs (e.g.

“The red managers were very tough”), others framed it around the game (e.g. “I had to work really hard to beat the game”) and not around the opponent, which might have made the negative contact less impactful. Secondly, the important thing to consider is the valence of contact itself. Whereas a considerable proportion of contact in the research game was negative and included fighting (as players had to destroy every rival building and remove all opposing managers), playing a game is a fun activity and the situation could have been at least partially pleasant for the players. Although participants in the experiment may have had other motivations aside from fun (e.g. helping a researcher or contributing to knowledge about gaming), the experiment was conducted on volunteer gamers so they were more likely to enjoy the game itself (they liked playing games in general and they chose to participate in the experience). Additionally, the study was conducted over the internet and offered no rewards so the cost of quitting was very low and some people who did not enjoy themselves could have easily opted out. Therefore it stands to reason that for most participants the situation was at least pleasant. The negative valence of the contact could have been mitigated by the positive valence of the situation.

Such an argument would hold for all games in general and, to a lesser extent, for all entertainment media e.g. television (actions in games are depicted as players’ own experience and have the potential to be more engaging as opposed to passively observing or reading about someone else’s actions on TV or in a text). Both games and television were shown to produce negative effects on attitudes if they depict highly negative contact (e.g. Sherry, 2001), so the pleasure of playing would not reverse the valence of contact completely. It could, however, mitigate it to some extent and thus make it easier to offset negative experiences with the positive. This hypothesis requires testing in future research. If true, this would make games a convenient vehicle for reducing prejudice as less positive contact would

54

be needed alongside negative contact to achieve prejudice reduction as compared to other forms of contact.

Powiązane dokumenty