Gender Aspects of Computer Avatars

(1)

Gender Aspects on

Computer Game Avatars

Anna Larsson1, Carina Nerén1 2005-10-04

SICS Technical Report T2005:06 ISSN 1100-3154

ISRN:SICS-T--2005/06-SE

Abstract.

Women are less dedicated to computer gaming than men. Previous studies show that one reason might be that current games exhibit hypersexualised female avatars: avatars that have exaggerated sexual signals to which female players object. In this thesis the purpose is to find out how female (and male) consumers of computer games really feel about hypersexualized female avatars in computer games and how these feelings differ between the genders. We also explore how the avatars’ appearance being assigned certain attitudes impact the preference of avatar.

The thesis is divided into two areas of study; in the first, the preliminary study, we investigate which attitudes are associated with certain stereotypes; in the second and main study we use the results from the first study when examining how people relate to hypersexual avatars and the reasons for this. The results from the main study of this thesis show that there are no great differences of how the hypersexual avatars are perceived by males and females, a majority in both gender groups do not reflect on the abnormality of the avatars exaggerated body shape; avatars with this appearance are actually preferred as personal representations in a game. It is the avatars clothes that are more in the centre of attention in the study, the avatars that does not show as much skin as the others are the ones preferred by both males and females. Stereotypical attitudes associated with the avatars seem to influence how the avatars are perceived, the avatars that make people think of a negative stereotype are shunned in selections for personal representation in a game, with a positive stereotype the reactions are the opposite.

1

Department of Computer and Systems Sciences, Stockholm University, where the original report was published November 2004. The project was supervised by Annika Waern, SICS. Email: annika@sics.se.

(2)

Keywords:

(3)

1 Introduction

1.1 Background

Women have always to a lesser extent than men had access to and used computers in their homes. Although the spread of computers for home use in, for example, Sweden has more than doubled between 1997 and 2003 there is still a gap between men and women users. (Statistiska C. 2003) This could have a number of reasons. Graner Ray, the author of Gender Inclusive Game Design from 2004, says that women tend to see computers not as a media of entertainment but as merely a useful tool, with which you solve specific tasks. Men, on the other hand, usually have no problems using the computer for all sorts of things; from work-related tasks like making spreadsheets to conquering the world in the virtual worlds of computer games.

The game industry does nothing or little to change the fact that women do not use com-puters for entertainment. In an investigation from 1990 done by a game developer on which game titles were offered in stores, it was found that products without a hypersexual avatar*2 on the cover did not containentertaining software. Instead they were more pro-ductively concerned with how to type better or how to prepare a budget and the like. The computer was not mediated as something fun and entertaining for girls; instead, it was presented only as something useful and practical. (Graner Ray 2004) This has perhaps changed in recent years when quite a number of games have been released for both fe -male and -male players and succeed in appealing to both groups. The Sims from Maxis, for example, was released in 2000 and was aninstant success, especially among girls who make out over 50% of the players (FiringSquad 2002). Some other recent numbers indi-cate that up to 70% of online* players are female, but in traditional* computer games girls are still heavily outnumbered by male players.

(Meretzky 2002, referred to by Graner Ray 2004)

According to Graner Ray (2004), the reason why boys to a larger extent play computer games than girls do is not that girls dislike technology or computers. If it were because of this there would not be as many girl users on the Internet3. The problems lay more likely in the fact that girls have been told all the way through their upbringing that computers and technology are not for girls and the fact that many of the games do not seem to be designed for girls due to their often violent and sexist contents. (Graner Ray 2004) An-other reason for girls not playing traditional computer games as much as boys do could be that there are more male programmers than female ones and therefore the kinds of games designed tend to be those that appeal to a male audience only. (Holland 2002) Despite the fact that there have been success stories such as The Sims, game designers still do not seem to know what female gamers* want. But what they do know is that it is a market with enormous potential. Software targeting only a female audience currently

2 A glossary (Appendix I) contains the most important terms used throughout the thesis. When first occurring, the terms will be marked with an asterisk *.

3

(7)

makes up less than one per cent of the total game industry, which in year 2001 made close to $20 billion, more than Hollywood takes at the box office. To succeed in reaching this new market it will require some deep thought about such fundamental issues as gen-der and stereotyping, and even the definition of games themselves. (Holland 2002)

The avatar is the first thing a computer game consumer gets in contact with when he or she is going to buy a game. This usually happens through the game’s cover. According to Graner Ray (2004), it is highly unlikely that the female gamer buys a game if the female avatar in the game is represented as hypersexual. Although she does not seem to have this statement scientifically established she could have a point.

There are probably a lot of different reasons why women do not want to play certain types of computer games. One reason could be the game developers’ choice of how they design the character* appearance where some appearance could appeal to women and others not. According to T.L Taylor, one of the key ways that games can exclude women is in the "hypersexualized" avatar choices they are offered. She thinks that the proportion of women playing games would likely increase if they were given more freedom to choose the way in which there are represented. (Ratan 2003)

But if we now know some important reasons to why women do not buy or play computer games as much as men do you could ask yourself why the design of computer games is not adjusted to appeal more to female players?

Apparently, most game development companies shy away from developing games for females because no one is really sure how large the market is for that group of people. It seems far safer to create games that arealready well established on the market, and most of these games happen to be designed for the male consumers, like for example the tradi-tional first-person shooter games*. (Holland 2002)

We, the authors of this thesis, are ourselves computer game players and weconstantly react to the fact that there are so many for us non-pleasing avatars in computer games. Instead of being designed for female users they seem to have been produced solely for the purpose of graphically representing some sort of female ideal that (some) men per-haps find attractive. After speaking to other women about this we discovered we were not at all alone in having these opinions. Since we could not find any data about female game players’ views on these kinds of overly exaggerated “sexy” female avatars, we decided to conduct our own study. Since it is the female avatars that concern us, there are only fe -male avatars in the study.

1.2 Problem statement

There is obviously a problem when facing females and games and that problem seems to lie in the fact that there are not that many games developed for female players and that in some of the games that do exist the player’s female representation consists of a hypersex-ualized avatar which women seem to dislike. If that is the case it is perhaps not that strange that there are fewer females than males playing games.

(8)

But why should women play games? Besides the obvious reasons based on the struggle for equality in society (where it seems self-evident that females should have the same opportunities as males to enjoy themselves), it seems as such a waste when (more than) half of the human population is not or scarcely at all considered when designing a commercial product.

But what should we do to make females more interested in playing computer games? In this thesis we are concerned with the design of the computer game avatars, and just by finding out what women (and men) really feel about their appearance can help in finding out what avatars (and perhaps as a consequence; computer games) should be like to ap-peal to all. We will also examine some possible reasons to why there are hypersexualized avatars in computer games and why people like/do not like them.

1.2.1 Purpose

The purpose of this thesis is to find out how consumers of computer games feel about hy-persexualized female avatars in computer games and how these feelings differ between the sexes. We also want to find out how the avatars’ appearance being assigned certain attitudes impact the preference of avatar.

1.2.2 Research Questions

The questions we seek answer for are thus:

? How do men and women choose if they are given the choice between female hypersexualized avatars and female “normal” avatars? Are there any differences between how male and female subjects make their choice of avatar?

? How do men and women think a member of the opposite sex would choose given a choice between hypersexualized and normal avatars?

? With which avatars do the subjects associate the stereotypical characteristics of a bimbo* and a crammer*4? Are there any differences in how male and female sub -jects associate these stereotypical attitudes with avatars?

? Do stereotypical characteristics associated with the avatars impact the choice of avatar to represent oneself?

1.2.3 Delimitations

The study is conducted using a game prototype we designed for this experiment (see 4.3.2) and the results may not be applicable to other computer games in different genres (further discussed in 6.6). The results of the comparisons between the avatars’ “charac-teristics” to those of stereotypes may only be generalized to a certain extent because of the small number of stereotypes examined in this thesis. Furthermore, the results from

4

(9)

both the preliminary and main studies may not be generalized outside of the span of the two population groups used in the studies, with reference to their limitations in number of subjects, age span, and geographical and cultural context.

1.3 Overview of Studies

The empirical data in this thesis will be gathered in two steps, a preliminary study is first performed to gather information about stereotypical characteristics to be used in the main study; in the main study we examine the computer game avatars. The two studies are de-scribed in more detail in chapter 4.

1.3.1 Preliminary Study – Examining Stereotypes

In the preliminary study we want to find out which attitudes are, of a given set, associated with the stereotypes crammer and bimbo. The results from this study will be used in the main study.

1.3.2 Main Study – Computer Game Avatars

In the main study the subjects play a text-based game where they get to converse with two in- game* characters; Ebba and Lovisa, whose stereotypical dialogues are modelled on the results from the preliminary study.

1.4 Overview of

Methodology

In the studies of this thesis we use different types of methods described below. These methods will be discussed further in chapter 3.

1.4.1 Preliminary Study

The preliminary study is conducted using a quantitative method. Questionnaires contain-ing multiple choice questions are used to gather the research data in the preliminary study.

1.4.2 Main Study

The main study is in its form explorative, since no preconceived perceptions about the results exist. We take both a qualitative and a quantitative approach. In the first part of the main study, where we look at how the players choose avatars, the approach is quanti-tative. To gather the information we want in this part we use questionnaires. In the sec-ond qualitative part we want to find out more about how the player relates to the avatars; and to do this we use an open inquiry and an interview to allow the player to express opinions and attitudes more freely.

(10)

? Chapter 2 consists of a background to the subject “gender and comp uter games”.

? In chapter 3 we discuss the methodology used in the studies as well as the causal models to be examined.

? In chapter 4 we describe the preliminary and main studies in more detail, as well as the results from the preliminary study to be used in the main study.

? In chapter 5 we list the main results of this thesis as well as give critique on our own work.

? Chapter 6 is a discussion of the results and related factors.

(11)

2 Theoretical Background

With this chapter we want to establish a relevant background to the subjects examined in this thesis.

2.1 Gender

In this thesis we will repeatedly refer to the term gender. Gender is a grammatical term that refers to a distinction between different types of nouns (for example une fille, un fils in French). In the 1970s the use of gender entered into a discourse about sociological dif-ferences between the sexes.(Connell 2003)

Sofie Petersson has in her thesis “kön.genus@IT” (Petersson 2002) described gender as being the socially constructed conceptions about the sexes. The differences between sex and gender contain the differences between essentialism and constructivism. Essentialists believe that the sexual identification is essentially bound to the body via a biological de-terminism which is unchangeable over time and space. Constructivists, on the other hand, state that gender is socially and culturally created in a specific time and space context; gender is a construction and the physical body and the biological prerequisites for what sex is are not enough to explain how women and men experience themselves. (Lundal 1998, 92f, Fagrell 2000 referred to in Petersson 2002)

In this thesis we will analyze things from a gender perspective and define differences between people’s identities as being socially and culturally constructed and not bio -logically bound to a specific sex.

2.2 The Female Character’s Development in Computer Games

There are two ways computer game designers have handled female representations in his-tory. The first is to depict woman as a sex ob-ject and the other is to depict her as an obob-ject for the male player to gain as a prize or to save. It is only in recent years female charac-ters have been given a place in the spotlight as capable main characters.

In the 80’s female characters consisted mainly of helpless girls whose main purpose in the game was to be saved by the player.

Donkey Kong (figure 1) and Dragons Lair,

both from 1984, serves as good examples from this time; here the female avatars in the games consisted of helpless girls who desperately

needed help from the male main characters. (Graner Ray 2004) In Donkey Kong, the fe

-Figure 1 – “Donkey Kong”, Nintendo version (Retrogames)

(12)

male avatar did not do anything; she worked only as a go al to be reached by the main character.

Throughout the 90’s female avatars began to appear more frequently and the games’ im-proved graphics made it possible to depict the avatars in a more realistic way, and also to make them “prettier”.

Figure 2 – “Duke Nukem 3D” (Planet Duke)

In the game Duke Nukem 3D (figure 2) from 1996 the female characters are portrayed as a couple of strippers which the player has to save. In Barbie Fashion Designer (figure 3), also from 1996, the female main character is portrayed with a disproportionate body for which the player is supposed to design clothes.

Also in 1996, one game managed to step beyond the “save the prin-cess” scenario. Tomb Raider (fig-ure 4) is a game where the main character is a very capable female avatar that runs, fights, shoots, and climbs. These abilities makes her a well-balanced character, she is no princess who has to be saved; she saves herself. Despite all this she is depicted with an unrealistically large chest, an extremely thin waist and an extremely well- rounded behind. So, although the skills of the avatar are well balanced in the game her appearance still qualifies as hypersexual and is, according to Graner Ray (2004) “[..] essentially ’eye candy’ for male players”.

Figure 3 – “Barbie Fashion Designer” (PC games)

Figure 4 – “Tomb Raider 1”, Lara Croft (Tomb Raider World)

(13)

The Longest Journey (figure 5) from 1999 is a role-playing game

where the female avatar saves the world. She is attractive with somewhatnormal proportions and clothes. (Graner Ray 2004) In SSX 3 (figure 6); a snowboarding game released for Play-station 2 in October 2003, the female characters are depicted just like ordinary snowboarding girls; with more or less ordinary snowboarding outfits to be selected by the player, with a focus on the style of the clothes being “right” for the snow-boarding arena, instead of the female characters being styled the right way for the male arena. (EA SPORTS BIG 2004)

2.3 The Importance of Avatars in Computer Game Playing

It is important to make the player feel comfortable with the avatar, since it is through the avatar the player interacts with the game. According to Andrew Rollings and Ernest Ad-ams a player will not play a game if the avatar does not interest the player or if the avatar is not believable (Rollings & Adams 2003). The way the avatar is presented, which role it is given to play and how the designers communicate the game to the player through the avatar are, according to Graner Ray (2004), three very important issues to keep in mind when it comes to ensuring the player’s level of satisfaction and joy. A well-constructed avatar encourages the player to identify with the character and this will improve the player’s sense of enjoyment. Even if a well-constructed character that is well-balanced with reference to graphics and capacity is important, it is just as important for a female player to have the possibility to be able to choose a female avatar when she is playing. If that possibility does not exist a female player will feel discontent with the game and, ac-cording to Graner Ray (2004), feel as if the game was not made for her and perhaps not play the game. According to Andrew Rollings and Ernest Adams (2003) the player will

Figure 5 – April, the main character in “The Longest Journey” (Longest Journey)

(14)

not develop an affinity for the avatar and will not be particularly sympathetic for the avatars plight if the avatar, among other things, does not intrigue the player or if the player does not like the avatar.

Computer games quite often offer some selected characteristics of specific sexes. In a role-playing game the player may, for example, choose from playing a male warrior, a male magician or a female thief. In this case where there is only one female character to choose from, who then is a thief, a female player who wanted to play a warrior would first be forced to overcome a possible feeling of discontentment with playing a male character. Incorporating less female characters than male characters in a game makes it, according to Graner Ray; seem as if the female character is nothing more than a symbolic character, especially if the female character is portrayed as a weaker kind of stereotype. (Graner Ray 2004)

2.4 Images of Women in Mass Media and Computer Games

In “Provocateur – Images of Women and Minorities in Advertising” (Cortese 1999), Cortese takes a look at how women are portrayed in western-culture (in this case US) ad-vertising. He draws a straight line between what men see as attractive in women and how women are made to appear in ads. According to Cortese men see a narrow waist and a high-pitched voice as a sign of vulnerability that attracts a man’s self- image as the Pro-tector. Large pupils are perceived as sexually attractive and dilation occurs unconsciously during arousal. Youth is also a sign of good health and sex appeal. An exaggerated leg length is perceived as more adult and therefore more sexual. A smile symbolises approval or attraction. Unaware blushing is considered very sexual because it signals innocence. The blushing starts at the cheeks and spreads to the rest of the face and often to other parts of the head and the body. The cleavage between the breasts and buttocks are also viewed as highly sexual. Women colour their lips so they become a symbol for sexual arousal. Big, red female lips resemble the female labia. (Cortese 1999)

In computer games, many of the above stated traits very often apply to how the female avatars are designed. According to Graner Ray, a female avatar can be represented in dif-ferent ways in a computer game, but the typical representation is one with unnaturally large breasts, unnaturally narrow waist and with a very shapely behind. All this plus her hypersexual way of dressing makes her send signals like: I am young, fertile and ready for sex. These signals are sent from the female conception where breasts placed high up on the chest, a narrow waist and a firm backside, indicate youth. Large breasts and round breasts indicate fertility. In the same way there are physical signals, which indicate that a woman is sexually susceptible, these are blushing, pouting lips, “bedroom eyes” and hardening nipples. These signals are also usually very exaggerated. (Graner Ray 2004)

2.5 Stereotypes

People sometimes are unable or unwilling to obtain all of the information they would need to make fair judgments about other people or situations. When people turn out to be in such situations they develop stereotypes. A stereotype is a generalization about a

(15)

per-son or group of perper-sons. Stereotypes allow people in many cases to “fill in the blanks” when they are in absence of the “total picture”. Without knowledge of all facts, people in virtually every case, resorting to prejudice by ascribing characteristics about a person based on a stereotype. By stereotyping, people assume that a person or a group has cer-tain characteristics and quite often people have stereotypes about persons who are mem-bers of groups with which they have not had firsthand contact. (Grobman 2004)

The journalist and commentator Walter Lippmann identifies four aspects of stereotypes. The first one is that a stereotype is simple; its characteristics may be summarized in two to three sentences. The second is that it is acquired second hand; people tend to acquire stereotypes from cultural mediators rather than from their own direct experience with the groups being stereotyped. The third aspect of stereotype is that all stereotypes are false. The fourth and last aspect is that stereotypes are resistant to change. (Lippmann 1922 re-ferred to in Nachbar et al 1992)

People stereotype around characteristics of age ("All teenagers love rock and roll and have no respect for their elders."), sex ("men want just one thing from a woman."), race ("All Japanese look and think alike."), religion ("All Catholics love the Pope more than their country."), vocation ("All lawyers are greedy weasels.") and nationality ("All Germans are Nazi warmongers."). According to John Nachbar and Kevin Lausé (1992) stereotyping is such a natural human function and it is so common that there are a number of useful functions of stereotyping. For one thing it is sometimes valuable to create clas-sifications of individuals. For example there is the stereotype freshman, which is a naive newcomer who is familiar with neither the social nor intellectual life of a campus. This stereotype serves the purpose of encouraging the constructing of introductory courses and it also encourages campus social organizations to sponsor group activities. A second useful function of stereotypes is in the use of what can be termed "countertypes." This is a positive stereotype (one which arouses "good" emotions and associates a group of people with socially approved characteristics). A “countertype” evolves as an attempt to replace or “counter” a negative stereotype that has been applied previously to a specific group of people. For example the negative stereotype of "Women as Helpless Vic tims" has been challenged as well with countertypes on US television ranging from the tough cops of Cagney and Lacy to the headstrong, independent Murphy Brown. Countertypes are important reflections (and shapers) of popular beliefs and values. (Nachbar et al 1992) In this thesis the stereotypes bimbo and crammer are used in our effort to explore to which extent certain physical traits can be associated to a set of attitudes. Bimbo is gen-erally believed to be extremely self-centred and extremely concerned with her appear-ance, the crammer, on the other hand, is extremely focused on her studies and dresses more casually.

(16)

3 Methodology

In this chapter we will declare and discuss the methods used in this thesis as well as give an overview of the aspects examined in a causal model.

When one looks at the research questions in this thesis one will see that it consist of three concepts. One of them is the participants in the study - the units, which in our study con-sist of yo ung people in the age between 17 and 22. One other concept represents what we have to say about the units, what we want to examine, that is the variables. There are in principle two different types of variables. The first variable is the one being explained (the dependent variable), and the other contains the elements that explain the quality amined (the independent variable). In this study the dependent variables include, for ex-ample, the subject’s choice of avatar and the character description of the avatar made by the subject. The independent variables, on the other hand, include for example the sub -ject’s gender and the avatar’s graphical appearance (see 3.2 Causal Models). (Holme&Solvang 1996)

3.1 Quantitative and Qualitative Methods

There are two main methodological approaches in research; quantitative and qualitative methods. Both qualitative and quantitative methods make out a classification system of the data collected in a research study.Quantitative methods are used to generalize the re-sults of the study and make statistics. The information is converted into numbers and quantities, which are used in statistical analyses. One reason and benefit of using a quan-titative method in the preliminary study and the first part of the main study is that we want the disposal of the study to be standardized. In both the preliminary and the main studies we use questionnaires to gather information and giving all subjects the same type of questions and answer alternatives standardizes that information. The standardized formation can then be generalized. The standardization makes the gathering of the in-formation completely controlled by the researchers and this is important when we, for example, want to get the answer to an “either-or question” with multiple choices such as when we gather information about how the players choose their avatar. Another benefit of using a quantitative method is that the gathered information is exact, but this also means that the information can be superficial; it disregards more subtle individual dif-ferences. (Holme&Solvang 1996) In the case of the subject choosing an avatar, however, this is not a problem since it is only the choice itself we are interested in at this stage. Qualitative methods, on the other hand, are used to get a deeper understanding of the ob-ject that we are studying. These methods combine five research techniques: direct obser-vation, participatory obserobser-vation, informant and respondent interviews and secondary data analysis. One benefit of using a qualitative method is that you get a comprehensive picture of the aspects examined. The use of qualitative methods in our study makes it possible for us to get a more complete picture of the players’ opinions examined in the study. Furthermore, the use of these methods may lead to a prevention of misunder-standings between the researchers and the subject by the researchers using usual and or-dinary conversations adjustments. By performing interviews with the subjects we can,

(17)

through this direct dialogue hopefully get as reliable information as possible; information that is a result of the players’ own opinions. The players can express themselves rela-tively freely and the study is not that much controlled by the researchers. Another benefit of using qualitative methods is that it makes the study more flexible. In this study we can use this to, for example, reformulate some questions from the interview if we find it nec-essary. (Holme&Solvang 1996) However, too extensive changes could make it hard to compare information from the different subjects.

3.2 Causal models

By thoroughly examining the research questions of this thesis we have extracted a few variables and relationships between variables to be the main issues of investigation. The primary variable we want to investigate is the subject’s choice of avatar to represent her/him in the game. This variable is a dependent variable, depending on the independent variables of the subject’s gender, the stereotypical attitudes the avatar is associated with (if any), and the appearance of the avatars. Other dependent variables are the perceptions the subjects have about the choices a member of the other gender would or would not do when selecting avatars. Finally we want to examine the way the avatars’ characters are perceived by the subjects as stated in the character descriptions5.

It is the relationship between the independent and the dependent variables we are con-cerned with here, thus; if and how the subject’s gender influences the choice of avatar and how it relates to how a member of the other gender is believed to choose; and also how the avatars’ appearance influence the choice of the subject’s own avatar. Lastly we want to find out if and how stereotypical attitudes associated with the avatars influence the choice of avatar6 and if they influence how the subjects perceive the avatars’ charac-ters.

The variables and interrelationships are shown in the causal model, figure 7.

5_{The character description is part of the main study, the subject describes how he/she perceives the}

character or personality of an avatar (see 4.2.2)

(18)

Figure 7 - Causal Model – The variables and relationships examined in this thesis

The independent variables are: a the gender of the subject

b the avatar’s graphical appearance, with a hypersexual or normal body and less or fully dressed

c the stereotypical attitudes of crammer and bimbo The dependent variables are:

d the choice of avatar that the subject makes

e the choice of own avatar made by a member of the other gender, as perceived by the subjects

f the description of an avatar’s character/personality made by the subject The relationships are:

a-d the relationship between the subject’s gender and the choice of avatar

a-e the influence of a subject’s gender when making preconceptions of the avatar choices of the other gender

b-d the relationship between the appearance of the avatar and the choice of avatar c-d the influence of stereotypical characteristics on the choice of avatars

c-f the influence of stereotypical attitudes on the way the subject perceive the character of an avatar

Subject’s gender - Female

- Male

Avatar Appearance - Hypersexual body, fully clothed - Hypersexual body, less clothed - Normal body, fully clothed - Normal body, less clothed

Stereotypical Attitudes

Choice of Avatar

Preconceptions about the other Gender Character Description b a c d e f

(19)

4 Empirical Study

This chapter describes the empirical elements of this thesis.

The empirical study consisted of two parts; the preliminary and the main studies. In the main study, where the subjects wereto play a game, one of the in- game characters was given attitudes that can be associated with the stereotype “bimbo” while the other char-acter has the attitudes of a stereotypical “crammer”.

These stereotypes were chosen because we think that they are frequently used among young people when referring to girls or women, and because they stand in opposition to one another. Another reason to choose the stereotypes bimbo and crammer was a precon-ception among the authors that the body types hypersexual and normal could be asso-ciated with them. The benefit of using well-known stereotypes, as we see it, is that they are well-established in people’s conceptions and people generally already have strong attitudes associated with them. Which attitudes that actually are associated with bimbo and crammer are investigated in the preliminary study.

Both the preliminary and the main studies was conducted in Swedish, thus all material from the studies shown in English in this thesis is translated by the authors.

4.1 Preliminary Study – Stereotypes

Since we did not find a relevant source to describe which attitudes are related to bimbo and crammer, we conducted a small survey among young people with a median age of 25 years to clarify which attitudes are associated to a certain stereotype. In this survey we asked the participants with which stereotype, given bimbo and crammer, they associated a given number of attitudes. Lastly the participants were given the opportunity to add atti-tudes to the stereotypes that they thought were missing in the questionnaire. All answers from the survey, as well as the questionnaire used, can bee seen in appendix II.

4.1.1 Study participants

The 21 participants were students belonging to the department of Computer and Systems Sciences at Stockholm University.

4.1.2 Material

The study material in the preliminary study consisted of a questionnaire with 51 ques-tions about attitudes and if and how the participant associated them with the stereotypes bimbo, crammer, both or none (see appendix II). The topics in the questionnaire were chosen because we think they play an important role in the thoughts and concerns of young people and therefore perhaps would make the dialogue of the game prototype seem somewhat natural.

(20)

4.1.3 Study procedure

The questionnaire was sent out by e-mail to a number of students who were enrolled at the department of computer and system sciences (at Stockholm university) in 2003. By choosing these students we were hoping to reach as many young people as possible. 21 students chose to fill out and return the questionnaire and those was then used in the analysis.

4.1.4 Data Processing

To be able to analyze the results from the survey, we compiled all the answers and made graphs to compare the results. Some of the questions with the most predominant results either towards bimbo or crammer were used when modelling the dialogue (the dialogue is in Appendix III) of the characters Ebba and Lovisa in the game.

4.1.5 Results from Preliminary Study

Some of the questions where a strong majority of the participants were unanimous were used in Ebba and Lovisa’s dialogue in the game prototype, the results that were not used there worked as a comparison for the answers given by the participants in the character descriptions in the main study, to establish if the description could be referring to a bimbo or crammer.

The results from the questions about what kind of courses a bimbo and a crammer takes in upper secondary school clearly showed that a crammer is associated with Natural sci-ence (Swe. Naturvetenskapligt program) and a bimbo with Hotel and Restaurant (Swe. Hotell och Restaurang) (see Figure 8). This was used in the dialogue.

Upper Secondary School

0 5 10 15 20

Bimbo Crammer Both None

Number of participants

The Hotel and Restaurant Program The Natural Science Program

(21)

Furthermore, a bimbo was considered thinking that her looks are very important, and a crammer was more associated with the attitudes “looks are somewhat important” or “looks are not important” (see Figure 9). This was also used in the dialogue.

The Importance of Looks

0 5 10 15 20

Bimbo Crammer Both None

Number of participants

Looks are very important Looks are somewhat important Looks are not important

Figures 9 – Looks are very important to bimbos but somewhat or not important to crammers

A bimbo was considered more likely wanting to start working right after school; a cram-mer was believed to wanting to continue studying at a university. A bimbo was more as-sociated with partying, salsa dancing and dressing provocatively, and a crammer with more intellectual things like playing an instrument.

4.2 Definitions of Hypersexual and Normal

The term hypersexual is in this thesis referring to a way of designing computer game avatars’ appearances. The focus of the depiction of hypersexual avatars seems to be on mak ing the players perceive the avatars as (extremely?) sexy though the context of the game does not call for it. Here, we are concerned with female avatars and the traits that make us define them as hypersexual are, as stated by Graner Ray (2004), unnaturally large breasts, a shapely behind, a narrow waist (often to the extreme), blushing face, big red pouting lips, “bedroom eyes” and hardening nipples. These traits in combination make up the term hypersexual.

When designing the normal avatars we used normal female anatomical measurements. By using these measurements we designed a body with normal “weight” and height and with normal proportions.

4.3 Main Study

In the main study we wanted to examine how young men and women relate to hyper-sexual avatars. By making the subjects choose from normal- looking and hyperhyper-sexualized avatars and then answer a set of questions regarding these avatars, we tried to learn how the choices were made and why. By introducing the subjects to the stereotypical attitudes of Ebba and Lovisa through the game prototype, and then asking the subjects to reselect Ebba’s and Lovisa’s avatars, we wanted to find out to which avatars these characters

(22)

were associated. Furthermore, by examining the characteristics associated with the ava -tars by the subjects and comparing these to the attitudes associated with the stereotypes bimbo and crammer in the preliminary study, we wanted to investigate if and how there were any connections between a certain appearance in the avatars with said stereotyp ical attitudes.

4.3.1 Study participants

The participants in the main study were 11 females and 11 males ages 17-22 from differ-ent parts of Stockholm. The participants were found through posters, e- mail send-outs to friends, schools, and online communities.

4.3.2 The Measurement Instrument

The study is conducted using a simple game prototype (see Appendix IV). The game is text-based and consists of two in-game female characters, Ebba and Lovisa, and the player’s character, who is also a girl. The three girls are just finishing upper secondary school and are in the middle of a conversation regarding themselves, their opinions in different matters and their future. The dialog of Ebba and Lovisa is based upon the atti-tudes which, by a majority of the preliminary study participants, are associated with the stereotypes bimbo and crammer. Ebba is the bimbo and Lovisa is the crammer.

Five questionnaires were used in the main study; the first mainly about the participants’ personal information such as age and game experience, and the second about the partici-pants’ views of the game7.

To record the participants’ opinions about the three avatars’ (the subject’s own avatar, Ebba, and Lovisa) “personalities”, we used three questionnaires with some compulsory ques tions about the avatar’s interests and hobbies, and an optional section where the par-ticipant could add additional information about the avatar.

4.3.2.1 Designing the Avatars

The appearances of the avatars are grouped into two variation groups; the body and the clothes. The body type consisted of two settings; normal and hypersexual. Since the ava -tar with the hypersexualized body was so exaggerated it appeared too different from the normal body. To reduce the differences between the bodies of the two types of avatars we slightly changed the proportions of the normal body by giving it a bit longer legs and making it somewhat slimmer. By doing so the normal avatars resembled more the exam-ple of a normal avatar given by (Graner Ray 2004).8 The appearance of the avatars clothes was divided into two levels of sexy. The levels where characterized by the amount of skin shown by the avatars’ clothes (the more the sexier). The first level con-sists of avatars with less clothing, and the second level avatars were fully dressed (see

7_{English versions of all questionnaires from the main study can bee seen in Appendix V(the questionnaires}

were in Swedish when used in the study).

8_{The avatar from the game “The Longest Journey” was given as an example of a normal avatar in (Graner}

(23)

Figure 10).

The four variables hypersexual body, normal body, less clothed, and fully clothed were distributed among six avatars, avatars A-F. The reason to why the number of eligible avatars amounted to six was that we wanted to prevent giving away the purpose of the study which could happen if we, for example, would use too few avatars (which could make their differences stand out too much). The solution to this problem, as we saw it, was to have enough avatars to give the participants several variants to choose from given a scale from normal to hypersexualized in body type and from normal to sexy in clothing; and we thought six avatars, three normal and three hypersexualized, would be sufficient for this task. The avatars were designed by compiling combinations of the avatars body type and type of clothes. By doing this we brought forward six different avatars of two combinations of body and clothing types. Three avatars with a normal body of which one was fully dressed, and two with less clothing; and three avatars with a hypersexualized body of which one had less clothing and two were fully dressed. By designing the avatars (with regard to the two types of appearance) in this manner, the two types of clothing became equally distributed among the eligible avatars. By giving the majority of the hypersexualized avatars more “normal” clothes and the majority of the normal avatars more “sexy-looking” clothes, we tried to further eliminate the too obvious differences between the two body types.

Figure 10 - The Avatars. A is hypersexual, fully clothed; B is normal, less clothed; C is normal, fully

clothed; D is hypersexual, fully clothed; E is hypersexual, less clothed; F (Ebba) is normal, less clothed; G (Lovisa) is normal, less clothed.

The seventh avatar, G, was designed for one of the opponent characters in the game,

(24)

Lovisa, by making a copy of F (Ebba) and altering its colouring slightly. G could not be chosen to represent the player. When “casting” the avatars for the characters Ebba and Lovisa we wanted them to resemble each other so that we perhaps could avoid influenc -ing the test participants by mediat-ing our opinions about how a bimbo and a crammer should look like.

4.3.2.2 Testing the Instruments

Before the main study we performed a test run that contained the same procedure that we intended to do in the main study. The test was performed on five test persons. This was done to find out ifthe system and the study material worked out well. The outcome of the tests showed that the study procedure and its material worked out properly.

4.3.3 Study Procedure

The study started off by the participant answering the questions in the first questionnaire. Then he/she played the game, and after that he/she answered the questions in the second questionnaire. After the participant answered the questions in the second questionnaire he/she was interviewed. The interview was recorded on tape.One question in the interview was if the participant would choose a different avatar than the one he/she had cho -sen in the first place if he/she were to play the game once more. A similar question ad-dressed if the participant would choose a different avatar for Ebba and Lovisa if he/she were given the opportunity to play the game once more. Following the interview the par-ticipant got three questionnaires were he/she was to write down a character description on his/her own re-chosen avatar, re-chosen Ebba and re-chosen Lovisa. If the participant did not re-choose any of the avatars, the character description was to be written about the “old” avatar/s.

4.3.4 Data Processing

To be able to analyze the results from the main study, we compiled all the answers from the questionnaires and the interviews were transcribed.

The quantitative data was entered into graphs to make comparisons between the sexes easier. The quantitative results were backed up with extracts from the subjects’ own writings and the interviews.

The results from the character descriptions of the re-chosen Ebba and Lovisa were compared with the results from the pre liminary study. Doing this helped decide however the characteristics of Ebba and Lovisa were, ac cording to the subjects, consistent with the, in the preliminary study, defined characteristics of a bimbo and a crammer. To be labelled a bimbo or a crammer, the description must contain at least 3 out of 4 stereotypical traits9.

9

(25)

5 Results

In this thesis we wanted to find out the answers for these questions:

? How do men and women choose if they are given the choice between female hypersexualized avatars and female “normal” avatars? Are there any differences between how male and female subjects make their choice of avatar?

? How do men and women think a member of the opposite sex would choose given a choice between hype rsexualized and normal avatars?

? With which avatars do the subjects associate the stereotypical characteristics of a bimbo and a crammer? Are there any differences in how male and female subjects associate these stereotypical attitudes with avatars?

? Do stereotypical characteristics associated with the avatars impact the choice of avatar to represent oneself?

In the following chapter we will try to answer these questions.10

5.1 Choice of Own Avatar

We first attend to the main scopeof this study. Is there any significant difference between the avatars chosen by young men and women? Figure 11 shows how the total number of people selected each avatar, together with the number of men and women that selected the same avatar. As we can see from this figure, men and women had somewhat similar preferences. The interesting and most prominent differences occur for avatar C (selected by more men than women) and avatar D (selected by more women than men): Both these avatars are fully clothed, but avatar D has a hypersexualized body.

Choices of avatars 0 1 2 3 4 5 6 7 8 A B C D E F avatars number of participants All participants Female participants Male participants

10_{All material collected in the main study is compiled in Appendix VII.} Figure 11 – The choices made by the subjects for own avatar

(26)

Body types, own avatar 0 2 4 6 8 10 12 14 16 Hypersexual Normal body types number of participants All participants Female participants Male participants

Figure 12 – The body types of the chosen own avatars

The similarities between the male and female subjects in choice of own avatars become even greater when you look at the body types and clothing types of the avatars. The hy-persexualized body is preferred by both women and men (figure 12). A possible expla-nation is that this body was interpreted as more distinctly female, as opposed to the ‘nor-mal’ body, which might have been considered more gender-neutral. The preference of avatar clothing level points in the opposite direction, most women and men prefer fully clothed avatars (Figure 13).

Clothing, own avatars

0 2 4 6 8 10 12 14 16 18 20

Less clothed Fully clothed

clothing

number of participants

All chosen avatars Women's avatars Men's avatars

Figure 13 – The clothing of the chosen own avatars

To find an explanation for this result, we examined the reasons that subjects voiced for their avatar selection. In general, the main reason to why both men and women selected avatars that they felt resembled them:

- “[…] I chose it mainly because of it’s clothes, It resembled me the most some-how…”(male, 20 years)

- “[…] it was the only one looking like me, the only one you could relate to…” (fe-male, 17 years)

(27)

In total, 9 (out of 22) subjects mention personal similarity as the main reason for selecting a particular avatar. The distinct similarities mentioned concern style of clothing (5 sub -jects).

When a man chose avatar C the motivation was 2 out of 4 times personal similarity. The subjects mentioned the clothes as the main significance. Another reason mentioned was that the avatar appeared the most normal.

- “[…] The most reasonable clothing and appearance I think…” (male, 17 years) - “ […] she didn’t stand out as much as the others, she seemed the most normal.”

(male, 18 years)

When a female subject chose avatar D the motivations were basically the same: 4 out of 4 subjects mentioned the clothes being similar to their own, but nobody mentioned the body shape.

5.2 Cross-Gender Stereotypes

In this thesis we also want to explore to what extent men and women could predict how the other gender would select their avatars. However, not all of the subjects were able to answer the question concerning which avatar they thought a member of the opposite gen-der would choose, so they were asked to name the avatar they thought a man/woman would not choose. In some cases the subject answered both these questions, and in others the subject did not answer any of them. This was a compromise made to let us have some sort of an insight in how men and women perceive the preferences of the other gender. This also makes the total number of participants to seem incorrect.

Figure 14shows a comparison between which avatars women thought that men would select or definitely not select, and what the men actually selected.

Women's conceptions of men's choice of avatar

0 1 2 3 4 5 6 7 8 9 A B C D E F avatars number of participants

Men's choice of avatar

What women think men will choose

What women think men will not choose

(28)

Figure 15 shows the same comparison for women. As we can see from these figures, the guesses made by the male subjects were roughly correct, the majority of the male sub -jects thought that the female sub-jects would not choose E (which no one did); whereas women guessed completely wrong. According to a majority of the female subjects, the men would choose E (only one did) and not C (one of the two most popular avatars among the male subjects).

Men's conceptions about women's choice of avatar

0 1 2 3 4 5 6 7 A B C D E F avatars number of participants

Women's choice of avatar

What men think women will choose

What men think women will not choose

Figure 15 - What avatars the male subjects think the female subjects would choose

5.3 Choice of Avatars for Stereotypic Characters

We now turn to the differences in how avatars were attributed to Ebba and Lovisa, the two stereotypic characters expressed in the dialogue. Remember that “Ebba” is here the ‘party girl’ (bimbo) stereotype, whereas Lovisa is more study and career-oriented (cram-mer).

(29)

Avatars to the character Ebba 0 1 2 3 4 5 6 7 8 9 A B C D E F G avatars number of participants All participants Female participants Male participants

Figure 16 – The avatars chosen to represent Ebba

Figure 16 shows avatar selections for Ebba. Note that avatar F was selected from start. However, there could be several non-relevant reasons to why a subject decided to keep avatar F for Ebba (such as the subject just being uncomfortable with the interview situa -tion) and therefore we decided not to include the avatar F in the analysis. As we see from this figure, 14 persons in total choose a new avatar for Ebba after playing the game. The avatar was changed by 7 of the male subjects and 7 of the female subjects. When rese-lecting Ebba, women choose between avatars B and E, whereas some men also selected the avatars A and G. One of the male subjects chose a fully clothed avatar (Figure 17) but in general, the sexes agreed that Ebba should wear less clothing. Furthermore, both male and female subjects selected the hypersexualized body more often than the normal body (Figure 18). However, it should be noted that the hypersexualized body was also pre-ferred for choice of own avatar by a large majority of the participants.

Clothing, Ebba (without F)

0 2 4 6 8 10 12 14

Less clothed Fully clothed clothing

All participants Female participants Male participants

(30)

Body type, Ebba (without F) 0 1 2 3 4 5 6 7 8 9 Hypersexual Normal body type number of participants All participants Female participants Male participants

Figure 18 - The body types of the avatars chosen for Ebba (except the original avatar F)

A similar pattern occurs for Lovisa (see figure 19). For Lovisa, avatar G was selected from start. Again, women that changed the avatar for Lovisa used only two avatars, A and C, whereas men also used avatar D.

Avatars to the character Lovisa

0 1 2 3 4 5 6 7 8 9 10 A B C D E F G avatars number of participants All participants Female participants Male participants

(31)

In reselections, the male subjects selected avatars with hypersexualized body type (Figure 20) as often as avatars with a normal body, where as a majority of the female subjects selected avatars with a normal body. In total, the normal body was preferred for Lovisa.

Body type, Lovisa (without G)

0 1 2 3 4 5 6 7 8 9 10

Hypersexual body type normal

Figure 20 – The body types of the avatars chosen for Lovisa (except the original avatar G)

All subjects re-selected fully clothed avatars for Lovisa (Figure 21).

Clothing, Lovisa (without G)

0 2 4 6 8 10 12 14 16

Less clothed _clothing Fully clothed

Figure 21 – The clothing of the avatars chosen for Lovisa (except the original avatar G)

5.4 Stereotypes and Avatars

The last research question to be answered in this thesis was:

Do stereotypical characteristics associated with the avatars impact the choice of avatar to represent oneself?

To be able to more exactly attach appearance to the stereotypes bimbo and crammer we had to dig deeper in how the subjects really perceived the opponent characters in the game. When looking more closely at the character descriptions for Ebba and Lovisa, we

(32)

see that 15 of 22 subjects recognize Lovisa as a crammer (8 women and 7 men) and 15 of 22 subjects recognize Ebba as a bimbo (10 women and 5 men). The avatars chosen by these subjects to represent Ebba and Lovisa can be seen in figure 22.

"Ebba the bimbo" avatars

0 2 4 6 8 avatars number of subjects Avatar A Avatar B Avatar E Avatar F

"Lovisa the crammer" avatars

0 2 4 6 8 avatars number of subjects Avatar A Avatar C Avatar D Avatar G

Figure 22 – Avatars chosen to the established stereotypical Ebba and Lovisa

When examining the results from the preliminary study we see that the stereotype bimbo is very much regarded as being a negative stereotype with which one would not like to be associated.

- “The bimbo is defined by putting herself in the centre, no empathy towards other people, she is the one that comes first. Whether she is fat, thin, or trim she thinks she’s the most good-looking” (male, 27 years)

- “I don’t need your attitude, coz I have my own... With that you see that it’s a hell of a lot missing in a Bimbo’s head. Ego-centred human being that makes you puke. Her headache is reoccurring since her only brain cell is trying to understand how STUPID it is.” (male, 24 years)

When we take a look at the results from the selections of avatars for Ebba (Figure 16) made by the subjects that found Ebba to be a bimbo we find that re-choosing the avatar for Ebba the most frequently chosen avatars were F, the original Ebba avatar (6 out of the 15 relevant subjects), and E (5 of the subjects). When comparing these results to the re-sults from the choice of own avatars (Figure 11) we see that E and F are merely chosen by one subject each to represent the subject in the game. Should we choose, like earlier in the analyses, to exclude the given avatar F because of the chance that there could be non-related reasons to why the subject did not want to reselect the avatar (see 5.3) the second most popular avatar to represent Ebba is B, also only chosen by 1 subject to represent the subject in the game.

From this we can conclude that the stereotype bimbo has had quite an impact on the choice of avatar and that this influence could be a reason to why the avatars B, E, and F, were so unpopular in the choice of own avatar.

(33)

up so many negative feelings as bimbo does. The general opinion seems more neutral. - “…the crammer is mostly shy, perhaps a bit less confident.” (male, 27 years) - “Pretty passive… afraid of failing. So she needs some courage. She needs to

toughen up” (male, 24 years)

Looking at the 15 subjects deciding Lovisa is a crammer we see that the most popular avatars to represent her was C (6 subjects), and G (the given avatar) and A (both chosen by 4 subjects). When comparing this to the results from the choice of own avatars we see that A is chosen by 7 subjects, making it the mos t popular of all avatars, closely followed by C (6 subjects). The problem of why G was left were it was, is in this case avoided since the avatar G was not made available for choice of own avatar and the results for G therefore could not be compared.

Even though the popularity of C (especially among men (4 of 6)) could have to do with C’s neutrality towards “masculine” appearance being appreciated by the male subjects the stereotype crammers impact on the choice of own avatar seems to have been positive, making the “crammer-avatars” be the most popular of all.

5.5 Criticism

The results from the preliminary study could have been influenced by the fact that a ma -jority of the population used was male (15 of 21), and perhaps made the results give a more of a male oriented perspective of which attitudes are associated with the stereotypes bimbo and crammer. However, since our analysis of the main study shows that the female subjects to a greater extent recognised the attitudes to being those of bimbo and crammer than the males do (will be discussed in 6.3), the unevenness between the sexes in the population of the preliminary study does not seem to have made too much of an impact on the final results.

Since the men participating in our main study invariably had higher game experience than the women, it was impossible to distinguish between the selections of men and those of experienced gamers.

However, within the group of male participants we can distinguish between those with high and those with medium game experience. Out of the men with medium game ex-perience a majority chose an avatar with a hypersexualized body, and the majority of the men belonging to the high experience category chose an avatar with normal body. Since the number of study participants is rather small we cannot draw any definite conclusions from this. However, we can note that young men with high game experience do not nec-essarily become accustomed to the female body stereotype prevalent in games, or at least this does not show in their avatar preferences. Nevertheless, the optimum population would have been one with a game experience equally distributed between the sexes.

(34)

6 Discussion

In this chapter we discuss the results from chapter 5 and state different factors that could have had an influence on their outcome.

6.1 The Hypersexualized Body is the Female Gender Norm

Both women and men selected the hypersexualized body to represent a woman. This body type was preferred for selections for Ebba, and for both the male and female sub -jects own avatars. Furthermore, very few participants ever reflected on the fact that this body was exaggerated. Only one subject commented on this body shape, when motivat-ing the reason why she chose C and not any of the other avatars:

- “[…] she looked the most normal. I don’t like… there are always [avatars] with really big breasts and blond hair in computer games…” (female, age 18)

The result is not surprising seeing the overrepresentation of the overly sexualized body in public space, as for example through advertisement (Cortese 1999). The hypersexualized body has been adopted by our society as being the female gender norm in female repre-sentations, to the extent that it is not even noted as being exaggerated.

6.2 Clothes Matter

Whereas the body shape is firmly established, the level of clothing for an avatar is critical in avatar selections. Both men and women prefer the fully clothed avatars, and attribute a consistent clothing level for Ebba and Lovisa, where Ebba wears less clothing than Lovisa. Furthermore, the level of clothing and the type of clothes that the avatars wear figure prominently in the motivations for selections (27 out of 66 comments).

6.3 Female Stereotypes are More Established Among Women than

Among Men

From the choices of avatars for Ebba and Lovisa, we see that the women chose a smaller range of avatars than men did. The reason could be either that the women found it easier to recognize the stereotype attitudes expressed in the dialogue, or that they were more similar in the way they attributed “looks” to these attitudes. To examine this, we studied the character descriptions for Ebba and Lovisa. In these we find that 10 of 11 of the women subjects give Ebba a personality that is associated with the stereotype bimbo, and 8 of 11 associate Lovisa with the attitudes of a crammer.

- About Ebba: “Looks are everything. Won’t commit to studies. A typical anti-feminist. Easily influenced by others. Not particularly intelligent.” (female, age 18)

(35)

- About Lovisa: “A goal-oriented girl. Thinks about the future. Doesn’t attach great importance to her appearance. Rather practical. By Ebba surely conceived as a crammer.” (female, age 21)

The male subjects were not as skilled with recognizing said stereotypes; 7 of 11 gave Lovisa the attributes of a crammer and only 5 of 11 gave Ebba bimbo-traits.

Another result that also indicates that women have a stronger sense of stereotypes was that women conformed more to the expectations concerning which avatar they selected for themselves. While the me n guessed rightly at women’s avatar preferences, the women’s guesses were false to a larger extent. The women conformed to what was ex-pected of them by the men (they did not select the avatar that was both hypersexualized and had little clothes), whereas the avatar that the women definitely did not expect men to choose was, together with another avatar, the most popular one.

6.4 Preconceptions

The women’s misconceptions about how men choose their avatars can be used to exem-plify how people’s preferences can be misjudged. Such a falsely preconceived notion could possibly explain why hypersexualized avatars figure so prominently in computer games of today: the games are directed towards young men (Rende&Steen) and the de-velopers seem to share the erroneous belief found in our study: that the hypersexualized avatars are preferred by such players. However, even though our study indicates that this stereotypic belief is wrong, there are three possible error sources that prevent us from drawing that conclusion from this study alone. Firstly, since this study used only female avatars, the men that never would have played a female avatar had no choice but to choose an avatar that was as gender- neutral as possible. Secondly, the fact that the experiment leaders were (young) women may have inhibited the men from choosing freely. Thirdly, the majority of game developers are men, and women’s conceptions about men alone may perhaps not be applicable to them, further information about how

men perceive other men’s preferences could be needed.

6.5 Stereotypes’ Effect on Avatar Preference

The reason to why we used stereotypes in this study was because we thought that appear-ance always, in some way, is associated with a certain behaviour or set of attitudes. The question to find answer to in this case was how these attitudes could have an impact on the choice of the avatar to represent you in the game of the main study. Our results show that there is a connection between perceived stereotypical attitudes with an avatar and whether that avatar is chosen to represent a person playing a game or not.

Avatars associated with bimbo attitudes are those with less clothing and to a certain ex-tent also with a hypersexualized body type. Avatars with these physical traits were the least chosen to represent the subjects in the game. On the other hand, the avatars associ-ated with the crammer attitudes, are fully dressed and the most popular ones in the avatar selections.

(36)

The results clearly indicate that attitudes associated with a certain appearance very well could influence the preferences of avatars in computer games. The question is whether the game designers’ intentions really are that the bimbo- like avatars present in games should be appreciated by those who identifies with a bimbo stereotype, we hope and think not. Perhaps it is the misconception found earlier in this thesis that is prevalent; that avatars with a “bimbo” appearance in some way are pleasing to the male audience the game is intended for, but our research show that the males like these avatars no more than the fe males do.

6.6 The Importance of Game Genre

One factor that will impact the choice of avatar is the genre of the game. In our study, the participants chose their avatar before they knew how the game worked. As the game sce-nario was rather realistic and dialogue-oriented, we could expect that gamers would pre-fer a more realistic avatar for this type of game than for a more typical computer game, where the subjects perhaps to a larger extent would have adjusted differently to the un-realistic context and assumed a “role”. After playing this game, 15 out of 22 claimed to have played the game as being themselves rather than the characteristics of an imaginary character.

Of all subjects, some preferred to keep the same avatar, but 11 actually did select another avatar after playing the game. Figure 23 shows the reselections

More sexualized body

Same body Less sexualized body Less clothed 1 3 1 Same clothing level 1 3 More clothed 1 1

Figure 23 – Re-chosen own avatars, variables

From this table we can see that there is a slight tendency to reselect the avatar to become more sexualized, either in body or in clothing. However, the majority of the subjects se-lected an avatar with the same body type.

This goes against our original thought, that the realistic stance of the dialogue would in-spire participants to select a more natural-looking avatar. 2 out of 4 of the subjects that, the second time, chose an avatar with less clothing did so because they thought that they would be treated differently by the opponents in the game.

(37)

- “I would perhaps change [avatar] just to see if there is any difference. I would perhaps choose someone tougher, maybe E” (female, 18)