Cogs200/2017W1/group7%

From UBC Wiki

Introduction

Background

Studies find that men and women react differently to competition and stress. Research examining women's performance in competitive chess -- primarily a male domain -- concluded that women underperform and resign quicker compared to their male competitors.[1] Other studies examining specific physiological responses to stress between male and female engineering students reported that female students had a higher heart rate than their male classmates, while male students had an increased level of adrenaline excretion. [2] Studies examining how men and women react to competition in female-dominated tasks found that women are more likely to exhibit signs of competitiveness. Furthermore, women are less likely to exhibit competitiveness against men in situations where they believe they will not win, while men are more likely to compete regardless of their perception of the outcome. [3] While previous research on the topic of gender and competition points to the differences in how men and women react to competition, there could also be other factors that play a role in these differences. Our study measures competitiveness, aggression, and stress between and among men and women in a simulated virtual reality (VR) environment.

Google DeepMind

The research study our group conducted is inspired by a study conducted by Google DeepMind, in which AI neural networks were tasked with picking apples in a game, with the goal of the study being to test whether the neural networks cooperated or competed with each other. The task we will present our participants will be the gender-neutral activity of picking 30 apples in 10 minutes. We will examine the extent to which men and women display signs of aggression and stress when competing against computer-generated players that appear to be of the same gender versus the opposite gender. Moreover, we will introduce specific environmental triggers in our VR environment, such as the sound of a ticking clock, which will gradually increase in speed throughout the game, in order to ascertain whether men and women react differently to these triggers. We will measure our participants' physiological responses to stress (heart rate and salivary cortisol), as well as their behavioural responses. We will measure behavioural responses by arming each participant with a controller, with which they can shoot lasers at their opponents.

Interdisciplinary Research

Our study incorporates the fields of computer science and psychology. We are interested in examining whether differences exist in how men and women compete, and if so, what those differences are. We have chosen VR to be the environment in which our study will be conducted. Recently, VR has gained traction in psychological studies, because it affords researchers the ability to manipulate the environment in order to measure potential responses from participants. [4] Moreover, our background research has pointed to specific differences in the way men and women react when competing against the opposite gender. Our goal is to ascertain whether those same differences hold true when men and women are competing against computer-generated players.

Hypothesis

Our background research has pointed to several differences in the way men and women react to competition, particularly when competing against the opposite gender. We hypothesize that men will exhibit more aggressive tendencies (in the form of shooting lasers at the opponents) compared to women, and that men will be less aggressive when competing against female-appearing computer-generated players. Secondly, we hypothesize that women will show an increase in heart rate and men will show an increase in cortisol levels. Finally, we hypothesize that both men and women will show an increase in heart rate and cortisol levels at the end of the experiment, compared to the beginning.

Implications of This Study

We believe the results of our study can be used in real-world settings to better understand how males and females react to stress and competition in academic settings and in the workplace. Studies have shown that many professional women still face a multitude of barriers when advancing their careers that their male co-workers do not face. [5] By examining our participants’ behavioural and physiological reactions to stress, we hope our research can be used to better improve classroom learning environments, such that they are equally conducive to both genders. Moreover, our results could help us better understand workplace gender dynamics, and the systems in place that trigger stress-induced reactions from men and women. Finally, our results could also be applied to other real-world settings involving competition between and among men and women, such as early education settings and co-ed sports teams.

Methods

Materials

HTC Vive

We will be using the HTC Vive VR system, which is known for its wide usage in the gaming industry, and the sense of realism it creates for players. [6] When a player is immersed in the system, every movement, small or large, is accurately mapped in the virtual environment. The accuracy of the system stems from sensors that seamlessly track players' every movement and position. This method of tracking, called "Lighthouse", uses infrared light to guide the player throughout the game. Much like a radar, LED light is flashed across the room, at a rate of 60 times per second. The sensors collect the LED, while the headset worn by the player records the time at which it is hit by the light, and uses that information to determine the position of the photosensor. Thus, the headset is able to quickly pinpoint the position of the player, relative to the headset's base stations. [7]

Uncanny Valley

In our experiment, we will be giving each participant two Vive controllers; one will be used to collect the apples and the other will be used to shoot lasers at the computer-generated players. In deciding how to create our computer-generated players, we looked at The Polar Express [8] as a model. The computer-generated players will be made to look realistic, yet inhuman at the same time. While the scope of our study, limited by financial constraints, may not allow for as impressive a production pipeline as the one uses by The Polar Express, [9], we will be incorporating the concept of 'uncanny valley' in our experiment. The computer-generated players will be made to look like 'humanoid robots' that appear both realistic and exaggerated at the same time. The goal is to create players that appear somewhat grotesque, [10] and ascertain whether this plays any role in the way the male and female participants compete. We plan to use MotionBuilder, which is a tool used for 3D modelling and animation, to create the computer-generated players. [11]

'Accounting' and 'Keep Talking and Nobody Explodes'

We found two VR games that incorporate some of the ideas that we will be using in our game: 'Accounting'.[12]

The former places participants in an environment where they are tasked with completing as many tasks as they can, as quickly as they can. The game introduces environmental triggers that cause participants to feel as if they are in an endurance test. Some of the environmental triggers include ominous music and aggressive behaviour from the computer-generated players. Furthermore, the game lulls players into a false sense of security, by not introducing environmental triggers until the player has had a chance to situate themselves within the game. Similarly, we plan to introduce environmental triggers after participants have had a chance to orient themselves in the environment. Our study differs from 'Accounting' in that, while we are introducing environmental triggers to elicit responses from our participants, we are not testing how much stress our participants can handle.

Similarly, 'Keep Talking and Nobody Explodes' traps participants in a room with a ticking time bomb, that he or she must diffuse before it blows up. involves a similar concept where the game is about a player trapped in a room in the VR game with a ticking time bomb that he or she must diffuse before it blows up. This game features environmental triggers, such as a ticking clock, which incites specific behaviours from participants.

Physiological Measures of Stress

We will be measuring the physiological signs of stress exhibited by participants by equipping them with a heart-rate monitoring chest strap, which they will wear throughout the game. The straps are made using an elongated elastic strap, which resembles a belt that is wrapped around the chest. Additionally, a small electrode pad is placed against the participant's’ skin; a snap-on transmitter will transmit the data to a machine. These heart-rate monitors use electrocardiography to record the electrical activity of the participants’ heart. [13] We will be measuring the heart rates at intervals of 2 minutes to test how stressed the players get throughout the game. Aggression will be measured to examining the frequency at which participants use the controller to shoot lasers at the computer-generated players.

Furthermore, we will be measuring the salivary cortisol levels of our participants, by taking a saliva swab before and after the experiment. Since saliva sampling is a relatively easy and non-invasive method for examining stress levels, we will be able to use an additional method for measuring stress levels in our participants. [14]

Procedure

Participants will be screened to ensure that they do not have any medical conditions that might hinder their ability to participate in a VR game. We will use random assignment to assign our participants into one of two groups: an experimental group and a control group.

The Task

The goal of our game is to collect 30 apples in 10 minutes, while competing against computer-generated players.

The Environment

The VR environment will be simulated to look like an orchard, with apples hidden behind branches of apple trees. Using the HTC Vive controller, participants will be tasked with picking the apples. At the beginning of the game, the sun will be shining, and the participants will be able to hear the sounds of birds chirping, which will combine to create a soothing environment. The computer-generated players will be picking apples at a rate of 3 apples per minute. Three minutes into the game, the participants will be able to hear the sound of a ticking clock. Five minutes into the game, the ticking clock will start ticking at twice the speed it was ticking at initially. Six minutes into the game, the computer-generated players will start picking apples at twice the rate they were picking at initially, and ominous music will start to play. Seven minutes into the game, the room will start to get darker, as the sun starts to set, to symbolize the passing of time. Eight minutes into the game, the computer-generated players will be picking apples at four times their initial picking speed. The sun will have completely set, and the participants will be able to see stars in the sky. The clock will be ticking at a speed that is four times its initial speed. During the last sixty seconds of the game, a voice will count down the number of minutes until the end of the game, in increments of 20 seconds. At any point in the game, if a participant comes close to a computer-generated player, the player will respond with statements such as, “Watch out!” or “This apple is mine!" When the game is finished, the room will be completely dark. By introducing the aforementioned environmental triggers, including the sound of the ticking clock gradually increasing in speed, the environment becoming darker to signify the passing of time, and the computer-generated players increasing the speed at which they pick apples, we will ascertain whether any of these stressors yield physiological or behavioural responses from the participants.

The Study

The goal of our experiment is to ascertain whether or not the participant will respond differently to either a male-appearing avatar or a female-appearing avatar. We will measure whether participants exhibit physiological and/or behavioural responses when competing against the same gendered avatars versus the opposite gendered avatars.

Our control group participants will be told the scope of the game, and that there are enough apples to in the orchard for each of them to collect 30 apples. Our experiment group participants will not be told the scope of the game, or the fact that there are enough apples in the orchard.

Once the participants have filled out the pre-study survey, they will be given a consent form to sign, which will assure them of the confidentiality of their data during the study. Additionally, as mentioned before, the control groups will be briefed on the scope of the study, and what they can expect throughout the experiment, including the environmental triggers we will introduce and the fact that they will be playing against computer-generated players. They will be clearly explained their task in the game, as well as the controllers they will be given during the experiment.

Independent Variables:

  • 'Gender' of the computer-generated players
  • Environmental triggers: speed of clock ticking, speed of computer-generated players

Dependent Variable:

  • Behavioural and physiological responses from participants

Participants

Participants will be screened to ensure that they are all native English speakers, grew up in North America, between the ages of 18-25, and are students of one of the following universities: University of British Columbia, Simon Fraser University, Capilano University, and Kwantlen Polytechnic University College. Participants will be compensated for their participation with either course credit for a psychology course or $5. Our sample size will consist of 240 participants: 120 male and 120 female. We have chosen this number to be our sample size as it would allow us to obtain more accurate results under budget constraints than a smaller sample size would.

Participants will be randomly assigned into one of the following groups:

Experimental Groups:

  • Group 1F: Female participants competing against female-appearing computer-generated players
  • Group 2F: Female participants competing against male-appearing computer-generated players
  • Group 1M: Male participants competing against male-appearing computer-generated players
  • Group 2M: Male participants competing against female-appearing computer-generated players


Control Groups:

  • Control Group A: Female participants competing against female-appearing computer-generated players; participants will be told the scope of the study
  • Control Group B: Female participants competing against male-appearing computer-generated players; participants will be told the scope of the study
  • Control Group C: Male participants competing against male-appearing computer-generated players; participants will be told the scope of the study
  • Control Group D: Male participants competing against female-appearing computer-generated players; participants will be told the scope of the study


The participants will not be informed of the category that they will be placed in; they will only be told that they will be competing against compute-generated players. We are using random assignment to place the participants in their respective groups, as it will increase the internal validity of our study, reduce bias, and decrease the chances of confounding variables affecting our data and results.

Prior to participating in the experiment, participants will be given a survey that will ascertain their cultural background, marital status, sexual orientation, familial environment (such as whether or not they grew up with siblings), and computer and/or video game habits (such as whether they have played a VR game before). These questions will allow us to assess how the participants' habits may influence the data and results of the experiment.

After the game is over, the participants in the experimental group will be asked qualitative questions about their experiences in the game, including which environmental trigger affected them the most, whether they felt like quitting the game at any point, how they felt when competing with computer-generated players of the same gender versus the opposite gender, whether they felt like they were competing against real people, and whether they would ever play the game again. The participants in the control group will be asked if, despite knowing the scope of the game, they were still affected by the environmental triggers, and how they felt when they were competing against computer-generated players of the same gender versus the opposite gender. The purpose of this debrief will be to hear firsthand from the participants the factors that played in any increase in aggression, competitiveness, or stress. The qualitative data we collect from our participants will be coded by assigning words and phrases into a category, in order to find themes in our data.

Expected Results

As part of our preliminary research, we did a mini meta-analysis of similar studies that had similar hypotheses, questions, and experimental purposes. We examined the results from four separate research studies in order to predict the results we might receive upon initiation of our proposed experiment. In the first study, researchers measured physiological changes between male and female students when they were introduced to a stressful task. They found that women's heart rates increased more than men's when faced with the task, while men showed an increase in adrenaline and cortisol excretion levels. [15]. Using this information, we predict that the female participants in our study will react more physiologically when confronted with competition. However, we also predict that male participants' cortisol levels will increase more noticeably when playing our competitive game. This study also helped us determine the physiological responses we can measure with our experiment when participants are confronted with competition.

Figure 1 Anderson & Carnagey study.png


We examined 3 other studies to help us predict the results and data we might obtain from our experiment. The first study involved two experimental groups, consisting of male and female participants; one group was tasked with playing a violent sports video game, and the other group with playing a non-violent sports video game. The results of the study indicated in Figure 1, show that those who were high on trait aggression (a participant variable; the level of aggression the participant had before coming into the study), expressed exponentially higher levels of aggressive cognition when playing a violent sports game, than those low on trait aggression. Interestingly, participants low on trait aggression expressed an increase in aggressive cognition in non-violent sports games, than those who were high on trait aggression. [16]. Figure 3 shows the results of the sex effect as one of the variables they were looking at, and according to the graph and the results men were more aggressive than women in a high-intensity setting, just like in the video game effect where violent video games showed more participant aggression in high-intensity settings [17].

qualitative results on video game characters predictions


The next study we examined incorporated a qualitative approach to video game aggression. Each participant was given an ambiguous story stem from a video game. They were then asked to predict what the video game character would do, say, think, and feel. As in Figure 2, the results indicated that participants predicted an aggressive response in violent video games across all 3 dependent measures, compared to nonviolent video games. [18] Once again, the video games that were used in the experiments were sports games involving the provocation of competition.

Sex effect and video game variable figure


Finally, we examined the third study on aggression in violent video games. In the first condition, players played against a computer (PVC), and in the second, they played against other players (PVP). According to Figure 4, the results exhibited that players experienced higher levels of aggressive feelings after playing against the computer, than after playing against another player. The interpersonal dynamic of face-to-face gameplay was more positive and had lower levels of aggression, possibly due to social factors, such as social facilitation theory. [19]


Aggression levels between PVC and PVP video games

Discussion

The purpose of this study is to determine whether environmental triggers will result in behavioural or physiological responses from participants. Furthermore, our goal is also to investigate whether gender plays a role in aggressive and competitive tendencies. As mentioned before, we hypothesize that our male participants will behave aggressively, compared to our female participants. In order to test our hypothesis, we will measure physiological changes in the participant (heart rate monitor, testing cortisol levels, etc).

We turned to our background research in order to determine how our research expands on existing studies. By examining the first study by Collins & Marianne (2010), we were able to better understand how our participants might respond when confronted with competition [20] By measuring physiological responses in our study, we will be able to better understand our competition and stress affects our participants, and whether gender plays any role in that. In the second study by Anderson & Carnagey (2009), we noticed that the participants who exhibited higher aggressive cognition were those who tested high on trait aggression [21] Another study done by Anderson & Bushman had a more qualitative approach to their testing. They found that participants subjectively explained reasons why and when characters in violent and non-violent video games acted aggresively [22] Furthermore, by administering a qualitative interview to each participant after they participate in the experiment, we will be able to rule out social factors that might have played a role in our results, such as participant fatigue.

One aspect of our study that may increase our knowledge of this area of study is the fact that we not only study sex effect and differences, but we are also testing the effects of interaction between the two sexes. A female vs. female test and a female vs. male test may show some interesting results as to how and why they may act aggressively. However, some ethical issues we will have to consider include the implications of administering a system where computer-generated entities are able to think, move, and behave on their own. As AI increasingly becomes an emerging field, one in which technologies are consistently advancing, we will have to be cognizant of the ethical implications of exploring this field. Potential gaps in our study that could skew the results include not accounting for the sexual orientation of our participants, the cultural background of our participants, as well as participants' online gaming habits.

While the scope of our study does not cover this, future research into the topic of gender, competition, and AI could examine the implications of competing against non-human AI entities. One of the studies we examined by Williams & Clippinger (2002) found that participants exhibited higher levels of aggression when playing against a computer, as opposed to playing against another player. These results might be indicative of social facilitation that we have towards each other, wherein the presence of other humans leads us to believe that we are being evaluated, and that we must conform to a certain public standard or norm [23]. This study is interesting because it highlights the complex social factors involved in competition and aggression, both in video games and in real life. Future research could take our ideas and expand on them, by testing the extent to which social facilitation holds true in a competitive VR setting.

Future research could examine how participants feel about AI humanoids that are made to look human. In-depth studies with advanced AIs might be able to gain some insight on how we as humans feel about artificial intelligence, and how we socially and biologically recognize, categorize, and behave around them. It would be interesting to further investigate human behavioural responses to AI behaviour, as it is still a relatively emerging field. Specifically, future studies could examine whether humans feel guilt when exhibiting aggressive behaviour against an AI competitor, as they might when competing against a human.

Conclusion

Our background research has shown us that men and women react differently to competition and stress. For instance, one study found that while women have better time-management skills than men, they experience higher levels of stress and anxiety than their male classmates. [24] Other differences between the genders include the fact that men experience higher levels of cortisol excretion when under stress or in competitive environments. [25]. While these studies point to gender-specific differences in responses to competition, we want to explore whether differences exist when participants are competing against the same gender versus the opposite gender.

In our study, our goal is to determine if there is a distinct relationship between gender and aggression. We plan to gather 240 participants -- 120 male and 120 female -- and randomly assign them into one of eight groups, consisting of 4 experimental groups and 4 control groups. We plan to create a simulated orchard, where participants will be tasked with picking 30 apples in 20 minutes, while playing against 3 computer-generated players, who will either be the same gender as them or the opposite gender. In total, our groups consist of males competing against male-appearing computer-generated players, females competing against female-appearing computer-generated players, males competing against female-appearing computer-generated players, and lastly, females competing against male-appearing computer-generated players. The goal of our experiment is to understand how each gender reacts to competition, specifically when they compete against the opposite gender.

To reiterate our hypothesis, we hypothesize three results from our experiment. Firstly, male participants would be more inclined to exhibit competitiveness and aggression, through terms of hindering the opponents by shooting lasers at them. Secondly, we hypothesize that male participants will be less aggressive when competing against female-appearing computer-generated players. Lastly, we hypothesize that female participants will have higher levels of physiological responses to stress, specifically an increase in heart rate, compared to male participants; male participants will have an increase in cortisol levels. In our study, the independent variable is the environmental triggers in the VR game and the "gender" of the computer-generated players, and the dependent variable is the behavioural and physiological response from our participants.

We will use random assignment to place participants in respective groups, and lessen the changes of biases clouding our results. The control groups will be told the scope of the game, as well as the fact every player will successfully be able to collect 30 apples in the given time frame. The experimental groups will not be told any of this.

We will introduce environmental triggers in order to elicit potential behavioural and physiological responses from our participants. These triggers will include the sound of a ticking clock increasing in speed throughout the game, the environment becoming darker to show the passage of time, and the computer-generated players increasing the speed at which they pick apples.

As technology advances, future research could consider exploring whether computer-generated players might perform differently if they know or recognize that the player is female or male.

References

  1. ↑ Lavy, Victor. 2013. Gender differences in market competitiveness in a real workplace: Evidence from Performance‐based pay tournaments among teachers. The Economic Journal 123 (569): 540-73.
  2. ↑ Collins, A., and M. Frankenhaeuser. 1978. Stress responses in male and female engineering students. Journal of Human Stress 4 (2): 43.
  3. ↑ Gunther, C., NA Ekinci, C. Schwieren, and M. Strobel. 2010. Women can't jump?-an experiment on competitive attitudes and stereotype threat. Journal of Economic Behavior & Organization 75 (3): 395-401.
  4. ↑ Wilson, CJ, and A. Soranzo. 2015. The use of virtual reality in psychology: A case study in visual perception. Computational and Mathematical Methods in Medicine 2015 : 1-7.
  5. ↑ Ragins, Belle Rose, Bickley Townsend, and Mary Mattis. 1998. Gender gap in the executive suite: CEOs and female executives report on breaking the glass ceiling. The Academy of Management Executive 12 (1): 28.
  6. ↑ "VIVE | Discover Virtual Reality Beyond Imagination." VIVE™. Accessed November 27, 2017. https://www.vive.com/eu/.
  7. ↑ "Raspberry Pi 3 is on sale now." SciFiNow. January 01, 1968. Accessed November 27, 2017. https://www.gadgetdaily.xyz/raspberry-pi-3-is-on-sale-now/.
  8. ↑ www.imdb.com/title/tt0338348/
  9. ↑ "Creative Human Character Animation: 'The Incredibles' vs. 'The Polar Express'." Animation World Network. Accessed November 27, 2017. https://www.awn.com/vfxworld/creative-human-character-animation-incredibles-vs-polar-express.
  10. ↑ "Creative Human Character Animation: 'The Incredibles' vs. 'The Polar Express'." Animation World Network. Accessed November 27, 2017. https://www.awn.com/vfxworld/creative-human-character-animation-incredibles-vs-polar-express.
  11. ↑ "MotionBuilder 2018 | Free Trial Autodesk." Autodesk 2D and 3D Design and Engineering Software. Accessed November 27, 2017. https://www.autodesk.com/products/motionbuilder/free-trial.
  12. ↑ Valentina Palladino - Apr 2, 2017 1:00 pm UTC. "How wearable heart-rate monitors work, and which is best for you." Ars Technica. April 02, 2017. Accessed November 27, 2017. https://arstechnica.com/gadgets/2017/04/how-wearable-heart-rate-monitors-work-and-which-is-best-for-you/.
  13. ↑ Valentina Palladino - Apr 2, 2017 1:00 pm UTC. "How wearable heart-rate monitors work, and which is best for you." Ars Technica. April 02, 2017. Accessed November 27, 2017. https://arstechnica.com/gadgets/2017/04/how-wearable-heart-rate-monitors-work-and-which-is-best-for-you/.
  14. ↑ Takai, Noriyasu, Masaki Yamaguchi, Toshiaki Aragaki, Kenji Eto, Kenji Uchihashi, and Yasuo Nishikawa. 2004. Effect of psychological stress on the salivary cortisol and amylase levels in healthy young adults. Archives of Oral Biology 49 (12): 963-8. Cancel
  15. ↑ Collins, A., and M. Frankenhaeuser. 1978. Stress responses in male and female engineering students. Journal of Human Stress 4 (2): 43.
  16. ↑ Anderson, Craig A., and Nicholas L. Carnagey. "Causal effects of violent sports video games on aggression: Is it competitiveness or violent content?" Journal of Experimental Social Psychology 45, no. 4 (2009): 731-39. doi:10.1016/j.jesp.2009.04.019.
  17. ↑ Anderson, Craig A., and Nicholas L. Carnagey. "Causal effects of violent sports video games on aggression: Is it competitiveness or violent content?" Journal of Experimental Social Psychology 45, no. 4 (2009): 731-39. doi:10.1016/j.jesp.2009.04.019.
  18. ↑ Bushman, Brad J., and Craig A. Anderson. "Violent Video Games and Hostile Expectations: A Test of the General Aggression Model." Personality and Social Psychology Bulletin 28, no. 12 (2002): 1679-686. doi:10.1177/014616702237649.
  19. ↑ Williams, Russell B., and Caryl A. Clippinger. 2002. Aggression, competition and computer games: Computer and human opponents. Computers in Human Behavior 18 (5): 495-506.
  20. ↑ Collins, A., and M. Frankenhaeuser. 1978. Stress responses in male and female engineering students. Journal of Human Stress 4 (2): 43.
  21. ↑ Anderson, Craig A., and Nicholas L. Carnagey. "Causal effects of violent sports video games on aggression: Is it competitiveness or violent content?" Journal of Experimental Social Psychology 45, no. 4 (2009): 731-39. doi:10.1016/j.jesp.2009.04.019.
  22. ↑ Bushman, Brad J., and Craig A. Anderson. "Violent Video Games and Hostile Expectations: A Test of the General Aggression Model." Personality and Social Psychology Bulletin 28, no. 12 (2002): 1679-686. doi:10.1177/014616702237649.
  23. ↑ Williams, Russell B., and Caryl A. Clippinger. 2002. Aggression, competition and computer games: Computer and human opponents. Computers in Human Behavior 18 (5): 495-506. Cancel
  24. ↑ Gunther, C., NA Ekinci, C. Schwieren, and M. Strobel. 2010. Women can't jump?-an experiment on competitive attitudes and stereotype threat. Journal of Economic Behavior & Organization 75 (3): 395-401.
  25. ↑ Collins, A., and M. Frankenhaeuser. 1978. Stress responses in male and female engineering students. Journal of Human Stress 4 (2): 43.
Retrieved from "https://wiki.ubc.ca/index.php?title=Cogs200/2017W1/group7%25&oldid=483557"